From 6eb4da8cf54537992fc9843be8b2af4f83f717e0 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:39 -0300 Subject: iommu: Have __iommu_probe_device() check for already probed devices This is a step toward making __iommu_probe_device() self contained. It should, under proper locking, check if the device is already associated with an iommu driver and resolve parallel probes. All but one of the callers open code this test using two different means, but they all rely on dev->iommu_group. Currently the bus_iommu_probe()/probe_iommu_group() and probe_acpi_namespace_devices() rejects already probed devices with an unlocked read of dev->iommu_group. The OF and ACPI "replay" functions use device_iommu_mapped() which is the same read without the pointless refcount. Move this test into __iommu_probe_device() and put it under the iommu_probe_device_lock. The store to dev->iommu_group is in iommu_group_add_device() which is also called under this lock for iommu driver devices, making it properly locked. The only path that didn't have this check is the hotplug path triggered by BUS_NOTIFY_ADD_DEVICE. The only way to get dev->iommu_group assigned outside the probe path is via iommu_group_add_device(). Today the only caller is VFIO no-iommu which never associates with an iommu driver. Thus adding this additional check is safe. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Acked-by: Rafael J. Wysocki Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/1-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/acpi/scan.c | 2 +- drivers/iommu/intel/iommu.c | 7 ------- drivers/iommu/iommu.c | 19 +++++++++---------- drivers/iommu/of_iommu.c | 2 +- 4 files changed, 11 insertions(+), 19 deletions(-) (limited to 'drivers') diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c index 5b145f1aaa1b..daa64dd68752 100644 --- a/drivers/acpi/scan.c +++ b/drivers/acpi/scan.c @@ -1581,7 +1581,7 @@ static const struct iommu_ops *acpi_iommu_configure_id(struct device *dev, * If we have reason to believe the IOMMU driver missed the initial * iommu_probe_device() call for dev, replay it to get things in order. */ - if (!err && dev->bus && !device_iommu_mapped(dev)) + if (!err && dev->bus) err = iommu_probe_device(dev); /* Ignore all other errors apart from EPROBE_DEFER */ diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 5c8c5cdc36cf..d5ca2387e65c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -3756,7 +3756,6 @@ static int __init probe_acpi_namespace_devices(void) for_each_active_dev_scope(drhd->devices, drhd->devices_cnt, i, dev) { struct acpi_device_physical_node *pn; - struct iommu_group *group; struct acpi_device *adev; if (dev->bus != &acpi_bus_type) @@ -3766,12 +3765,6 @@ static int __init probe_acpi_namespace_devices(void) mutex_lock(&adev->physical_node_lock); list_for_each_entry(pn, &adev->physical_node_list, node) { - group = iommu_group_get(pn->dev); - if (group) { - iommu_group_put(group); - continue; - } - ret = iommu_probe_device(pn->dev); if (ret) break; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index da340f11c5f5..bdc5ce884b7b 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -351,9 +351,16 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list * but for now enforcing a simple global ordering is fine. */ mutex_lock(&iommu_probe_device_lock); + + /* Device is probed already if in a group */ + if (dev->iommu_group) { + ret = 0; + goto out_unlock; + } + if (!dev_iommu_get(dev)) { ret = -ENOMEM; - goto err_unlock; + goto out_unlock; } if (!try_module_get(ops->owner)) { @@ -399,7 +406,7 @@ out_module_put: err_free: dev_iommu_free(dev); -err_unlock: +out_unlock: mutex_unlock(&iommu_probe_device_lock); return ret; @@ -1711,16 +1718,8 @@ 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; diff --git a/drivers/iommu/of_iommu.c b/drivers/iommu/of_iommu.c index 40f57d293a79..157b286e36bf 100644 --- a/drivers/iommu/of_iommu.c +++ b/drivers/iommu/of_iommu.c @@ -159,7 +159,7 @@ const struct iommu_ops *of_iommu_configure(struct device *dev, * If we have reason to believe the IOMMU driver missed the initial * probe for dev, replay it to get things in order. */ - if (!err && dev->bus && !device_iommu_mapped(dev)) + if (!err && dev->bus) err = iommu_probe_device(dev); /* Ignore all other errors apart from EPROBE_DEFER */ -- cgit v1.2.3 From 5665d15d3cb796d363a2dc0d2ed17855a3cb5942 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:40 -0300 Subject: iommu: Use iommu_group_ref_get/put() for dev->iommu_group No reason to open code this, use the proper helper functions. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/2-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index bdc5ce884b7b..2f16f988ab36 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -500,7 +500,7 @@ static void __iommu_group_release_device(struct iommu_group *group, kfree(grp_dev->name); kfree(grp_dev); dev->iommu_group = NULL; - kobject_put(group->devices_kobj); + iommu_group_put(group); } static void iommu_release_device(struct device *dev) @@ -1067,8 +1067,7 @@ rename: goto err_free_name; } - kobject_get(group->devices_kobj); - + iommu_group_ref_get(group); dev->iommu_group = group; mutex_lock(&group->mutex); -- cgit v1.2.3 From 7bdb99622f7e7dcaa58bfc2fa98caf23cfc40994 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:41 -0300 Subject: iommu: Inline iommu_group_get_for_dev() into __iommu_probe_device() This is the only caller, and it doesn't need the generality of the function. We already know there is no iommu_group, so it is simply two function calls. Moving it here allows the following patches to split the logic in these functions. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/3-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 50 +++++++++----------------------------------------- 1 file changed, 9 insertions(+), 41 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 2f16f988ab36..bd56cf02bc99 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -127,7 +127,6 @@ 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); @@ -379,12 +378,18 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list if (ops->is_attach_deferred) dev->iommu->attach_deferred = ops->is_attach_deferred(dev); - 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; } + ret = iommu_group_add_device(group, dev); + if (ret) + goto err_put_group; + mutex_lock(&group->mutex); if (group_list && !group->default_domain && list_empty(&group->entry)) list_add_tail(&group->entry, group_list); @@ -396,6 +401,8 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list return 0; +err_put_group: + iommu_group_put(group); out_release: if (ops->release_device) ops->release_device(dev); @@ -1670,45 +1677,6 @@ 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; -- cgit v1.2.3 From df15d76dcacac2126a4d20ba06c9d7e4b18bad8e Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:42 -0300 Subject: iommu: Simplify the __iommu_group_remove_device() flow Instead of returning the struct group_device and then later freeing it, do the entire free under the group->mutex and defer only putting the iommu_group. It is safe to remove the sysfs_links and free memory while holding that mutex. Move the sanity assert of the group status into __iommu_group_free_device(). The next patch will improve upon this and consolidate the group put and the mutex into __iommu_group_remove_device(). __iommu_group_free_device() is close to being the paired undo of iommu_group_add_device(), following patches will improve on that. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/4-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 83 ++++++++++++++++++++++++--------------------------- 1 file changed, 39 insertions(+), 44 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index bd56cf02bc99..6cf4da060e94 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -470,32 +470,8 @@ err_out: } -/* - * 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) -{ - struct group_device *device; - - lockdep_assert_held(&group->mutex); - for_each_group_device(group, device) { - if (device->dev == dev) { - list_del(&device->list); - return device; - } - } - - return NULL; -} - -/* - * 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; @@ -504,16 +480,45 @@ 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; - iommu_group_put(group); } -static void iommu_release_device(struct device *dev) +/* + * Remove the iommu_group from the struct device. The attached group must be put + * by the caller after releaseing the group->mutex. + */ +static void __iommu_group_remove_device(struct device *dev) { struct iommu_group *group = dev->iommu_group; struct group_device *device; + + lockdep_assert_held(&group->mutex); + for_each_group_device(group, device) { + if (device->dev != dev) + continue; + + list_del(&device->list); + __iommu_group_free_device(group, device); + /* Caller must put iommu_group */ + return; + } + WARN(true, "Corrupted iommu_group device_list"); +} + +static void iommu_release_device(struct device *dev) +{ + struct iommu_group *group = dev->iommu_group; const struct iommu_ops *ops; if (!dev->iommu || !group) @@ -522,16 +527,7 @@ static void iommu_release_device(struct device *dev) iommu_device_unlink(dev->iommu->iommu_dev, dev); mutex_lock(&group->mutex); - device = __iommu_group_remove_device(group, 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); + __iommu_group_remove_device(dev); /* * release_device() must stop using any attached domain on the device. @@ -547,8 +543,8 @@ static void iommu_release_device(struct device *dev) ops->release_device(dev); mutex_unlock(&group->mutex); - if (device) - __iommu_group_release_device(group, device); + /* Pairs with the get in iommu_group_add_device() */ + iommu_group_put(group); module_put(ops->owner); dev_iommu_free(dev); @@ -1107,7 +1103,6 @@ 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; @@ -1115,11 +1110,11 @@ void iommu_group_remove_device(struct device *dev) dev_info(dev, "Removing from iommu group %d\n", group->id); mutex_lock(&group->mutex); - device = __iommu_group_remove_device(group, dev); + __iommu_group_remove_device(dev); mutex_unlock(&group->mutex); - if (device) - __iommu_group_release_device(group, device); + /* Pairs with the get in iommu_group_add_device() */ + iommu_group_put(group); } EXPORT_SYMBOL_GPL(iommu_group_remove_device); -- cgit v1.2.3 From aa0958570f24f562422afa41fefd1b3a1fe0f6d0 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:43 -0300 Subject: iommu: Add iommu_init/deinit_device() paired functions Move the driver init and destruction code into two logically paired functions. There is a subtle ordering dependency in how the group's domains are freed, the current code does the kobject_put() on the group which will hopefully trigger the free of the domains before the module_put() that protects the domain->ops. Reorganize this to be explicit and documented. The domains are cleaned up by iommu_deinit_device() if it is the last device to be deinit'd from the group. This must be done in a specific order - after ops->release_device() and before the module_put(). Make it very clear and obvious by putting the order directly in one function. Leave WARN_ON's in case the refcounting gets messed up somehow. This also moves the module_put() and dev_iommu_free() under the group->mutex to keep the code simple. Building paired functions like this helps ensure that error cleanup flows in __iommu_probe_device() are correct because they share the same code that handles the normal flow. These details become relavent as following patches add more error unwind into __iommu_probe_device(), and ultimately a following series adds fine-grained locking to __iommu_probe_device(). Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/5-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 191 +++++++++++++++++++++++++++++--------------------- 1 file changed, 112 insertions(+), 79 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 6cf4da060e94..f052f52958fd 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -332,10 +332,99 @@ static u32 dev_iommu_get_max_pasids(struct device *dev) return min_t(u32, max_pasids, dev->iommu->iommu_dev->max_pasids); } +/* + * 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) +{ + struct iommu_device *iommu_dev; + struct iommu_group *group; + int ret; + + if (!dev_iommu_get(dev)) + return -ENOMEM; + + if (!try_module_get(ops->owner)) { + ret = -EINVAL; + goto err_free; + } + + iommu_dev = ops->probe_device(dev); + if (IS_ERR(iommu_dev)) { + ret = PTR_ERR(iommu_dev); + goto err_module_put; + } + + group = ops->device_group(dev); + if (WARN_ON_ONCE(group == NULL)) + group = ERR_PTR(-EINVAL); + if (IS_ERR(group)) { + ret = PTR_ERR(group); + goto err_release; + } + dev->iommu_group = group; + + 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; + +err_release: + if (ops->release_device) + ops->release_device(dev); +err_module_put: + module_put(ops->owner); +err_free: + dev_iommu_free(dev); + return ret; +} + +static void iommu_deinit_device(struct device *dev) +{ + struct iommu_group *group = dev->iommu_group; + const struct iommu_ops *ops = dev_iommu_ops(dev); + + lockdep_assert_held(&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. + */ + 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); +} + static int __iommu_probe_device(struct device *dev, struct list_head *group_list) { 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; @@ -357,62 +446,30 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list goto out_unlock; } - if (!dev_iommu_get(dev)) { - ret = -ENOMEM; + ret = iommu_init_device(dev, ops); + if (ret) goto out_unlock; - } - - if (!try_module_get(ops->owner)) { - ret = -EINVAL; - goto err_free; - } - - iommu_dev = ops->probe_device(dev); - if (IS_ERR(iommu_dev)) { - ret = PTR_ERR(iommu_dev); - goto out_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); - - 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; - } + group = dev->iommu_group; ret = iommu_group_add_device(group, dev); + mutex_lock(&group->mutex); if (ret) goto err_put_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); + iommu_device_link(dev->iommu->iommu_dev, dev); return 0; err_put_group: + iommu_deinit_device(dev); + mutex_unlock(&group->mutex); iommu_group_put(group); -out_release: - if (ops->release_device) - ops->release_device(dev); - -out_module_put: - module_put(ops->owner); - -err_free: - dev_iommu_free(dev); - out_unlock: mutex_unlock(&iommu_probe_device_lock); @@ -491,63 +548,45 @@ static void __iommu_group_free_device(struct iommu_group *group, kfree(grp_dev->name); kfree(grp_dev); - dev->iommu_group = NULL; } -/* - * Remove the iommu_group from the struct device. The attached group must be put - * by the caller after releaseing the group->mutex. - */ +/* 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; - lockdep_assert_held(&group->mutex); + mutex_lock(&group->mutex); for_each_group_device(group, device) { if (device->dev != dev) continue; list_del(&device->list); __iommu_group_free_device(group, device); - /* Caller must put iommu_group */ - return; + if (dev->iommu && dev->iommu->iommu_dev) + iommu_deinit_device(dev); + else + dev->iommu_group = NULL; + goto out; } WARN(true, "Corrupted iommu_group device_list"); +out: + mutex_unlock(&group->mutex); + + /* Pairs with the get in iommu_group_add_device() */ + iommu_group_put(group); } static void iommu_release_device(struct device *dev) { struct iommu_group *group = dev->iommu_group; - const struct iommu_ops *ops; if (!dev->iommu || !group) return; iommu_device_unlink(dev->iommu->iommu_dev, dev); - mutex_lock(&group->mutex); __iommu_group_remove_device(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. - */ - ops = dev_iommu_ops(dev); - if (ops->release_device) - ops->release_device(dev); - mutex_unlock(&group->mutex); - - /* Pairs with the get in iommu_group_add_device() */ - iommu_group_put(group); - - module_put(ops->owner); - dev_iommu_free(dev); } static int __init iommu_set_def_domain_type(char *str) @@ -808,10 +847,9 @@ 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); @@ -1109,12 +1147,7 @@ void iommu_group_remove_device(struct device *dev) dev_info(dev, "Removing from iommu group %d\n", group->id); - mutex_lock(&group->mutex); __iommu_group_remove_device(dev); - mutex_unlock(&group->mutex); - - /* Pairs with the get in iommu_group_add_device() */ - iommu_group_put(group); } EXPORT_SYMBOL_GPL(iommu_group_remove_device); -- cgit v1.2.3 From 14891af3799e8cd24dee14f78c31fa663cfb5ba9 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:44 -0300 Subject: iommu: Move the iommu driver sysfs setup into iommu_init/deinit_device() It makes logical sense that once the driver is attached to the device the sysfs links appear, even if we haven't fully created the group_device or attached the device to a domain. Fix the missing error handling on sysfs creation since iommu_init_device() can trivially handle this. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/6-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu-sysfs.c | 6 ------ drivers/iommu/iommu.c | 13 +++++++++---- 2 files changed, 9 insertions(+), 10 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu-sysfs.c b/drivers/iommu/iommu-sysfs.c index 99869217fbec..c8aba0e2a30d 100644 --- a/drivers/iommu/iommu-sysfs.c +++ b/drivers/iommu/iommu-sysfs.c @@ -107,9 +107,6 @@ int iommu_device_link(struct iommu_device *iommu, struct device *link) { int ret; - if (!iommu || IS_ERR(iommu)) - return -ENODEV; - ret = sysfs_add_link_to_group(&iommu->dev->kobj, "devices", &link->kobj, dev_name(link)); if (ret) @@ -126,9 +123,6 @@ EXPORT_SYMBOL_GPL(iommu_device_link); void iommu_device_unlink(struct iommu_device *iommu, struct device *link) { - if (!iommu || IS_ERR(iommu)) - return; - sysfs_remove_link(&link->kobj, "iommu"); sysfs_remove_link_from_group(&iommu->dev->kobj, "devices", dev_name(link)); } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index f052f52958fd..4688f61b7a4f 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -356,12 +356,16 @@ static int iommu_init_device(struct device *dev, const struct iommu_ops *ops) goto err_module_put; } + ret = iommu_device_link(iommu_dev, dev); + if (ret) + goto err_release; + group = ops->device_group(dev); if (WARN_ON_ONCE(group == NULL)) group = ERR_PTR(-EINVAL); if (IS_ERR(group)) { ret = PTR_ERR(group); - goto err_release; + goto err_unlink; } dev->iommu_group = group; @@ -371,6 +375,8 @@ static int iommu_init_device(struct device *dev, const struct iommu_ops *ops) dev->iommu->attach_deferred = ops->is_attach_deferred(dev); return 0; +err_unlink: + iommu_device_unlink(iommu_dev, dev); err_release: if (ops->release_device) ops->release_device(dev); @@ -388,6 +394,8 @@ static void iommu_deinit_device(struct device *dev) 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 @@ -462,7 +470,6 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list iommu_group_put(group); mutex_unlock(&iommu_probe_device_lock); - iommu_device_link(dev->iommu->iommu_dev, dev); return 0; @@ -584,8 +591,6 @@ static void iommu_release_device(struct device *dev) if (!dev->iommu || !group) return; - iommu_device_unlink(dev->iommu->iommu_dev, dev); - __iommu_group_remove_device(dev); } -- cgit v1.2.3 From 9a108996b5b39c8b7b4911bd4fe7f9ad87bc0f72 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:45 -0300 Subject: iommu: Do not export iommu_device_link/unlink() These are not used outside iommu.c, they should not be available to modular code. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/7-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu-sysfs.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu-sysfs.c b/drivers/iommu/iommu-sysfs.c index c8aba0e2a30d..cbe378c34ba3 100644 --- a/drivers/iommu/iommu-sysfs.c +++ b/drivers/iommu/iommu-sysfs.c @@ -119,11 +119,9 @@ int iommu_device_link(struct iommu_device *iommu, struct device *link) return ret; } -EXPORT_SYMBOL_GPL(iommu_device_link); void iommu_device_unlink(struct iommu_device *iommu, struct device *link) { sysfs_remove_link(&link->kobj, "iommu"); sysfs_remove_link_from_group(&iommu->dev->kobj, "devices", dev_name(link)); } -EXPORT_SYMBOL_GPL(iommu_device_unlink); -- cgit v1.2.3 From cfb6ee65f7603a0605fa8f5fe5b0782f0731c81c Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:46 -0300 Subject: iommu: Always destroy the iommu_group during iommu_release_device() Have release fully clean up the iommu related parts of the struct device, no matter what state they are in. Split the logic so that the three things owned by the iommu core are always cleaned up: - Any attached iommu_group - Any allocated dev->iommu and its contents including a fwsepc - Any attached driver via a struct group_device This fixes a minor bug where a fwspec created without an iommu_group being probed would not be freed. Reviewed-by: Kevin Tian Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/8-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 4688f61b7a4f..55b916e8da3b 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -574,10 +574,8 @@ static void __iommu_group_remove_device(struct device *dev) iommu_deinit_device(dev); else dev->iommu_group = NULL; - goto out; + break; } - WARN(true, "Corrupted iommu_group device_list"); -out: mutex_unlock(&group->mutex); /* Pairs with the get in iommu_group_add_device() */ @@ -588,10 +586,12 @@ static void iommu_release_device(struct device *dev) { struct iommu_group *group = dev->iommu_group; - if (!dev->iommu || !group) - return; + if (group) + __iommu_group_remove_device(dev); - __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) -- cgit v1.2.3 From fa08280364882c42993dc5d8394c467d76803fdb Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:47 -0300 Subject: iommu: Split iommu_group_add_device() Move the list_add_tail() for the group_device into the critical region that immediately follows in __iommu_probe_device(). This avoids one case of unlocking and immediately re-locking the group->mutex. Consistently make the caller responsible for setting dev->iommu_group, prior patches moved this into iommu_init_device(), make the no-driver path do this in iommu_group_add_device(). This completes making __iommu_group_free_device() and iommu_group_alloc_device() into pair'd functions. Reviewed-by: Lu Baolu Reviewed-by: Kevin Tian Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/9-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 66 +++++++++++++++++++++++++++++++++------------------ 1 file changed, 43 insertions(+), 23 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 55b916e8da3b..3129e5c50c87 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -129,6 +129,8 @@ static int iommu_create_device_direct_mappings(struct iommu_domain *domain, 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); #define IOMMU_GROUP_ATTR(_name, _mode, _show, _store) \ struct iommu_group_attribute iommu_group_attr_##_name = \ @@ -435,6 +437,7 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list 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; if (!ops) @@ -459,16 +462,17 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list goto out_unlock; group = dev->iommu_group; - ret = iommu_group_add_device(group, dev); + gdev = iommu_group_alloc_device(group, dev); mutex_lock(&group->mutex); - if (ret) + if (IS_ERR(gdev)) { + ret = PTR_ERR(gdev); goto err_put_group; + } + list_add_tail(&gdev->list, &group->devices); 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); return 0; @@ -578,7 +582,10 @@ static void __iommu_group_remove_device(struct device *dev) } mutex_unlock(&group->mutex); - /* Pairs with the get in iommu_group_add_device() */ + /* + * Pairs with the get in iommu_init_device() or + * iommu_group_add_device() + */ iommu_group_put(group); } @@ -1067,22 +1074,16 @@ 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; @@ -1113,17 +1114,11 @@ rename: goto err_free_name; } - iommu_group_ref_get(group); - 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); @@ -1132,7 +1127,32 @@ 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); -- cgit v1.2.3 From f188056352bcdcd4bec31c22baa00ba6568f1416 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 5 Jun 2023 21:59:48 -0300 Subject: iommu: Avoid locking/unlocking for iommu_probe_device() Remove the race where a hotplug of a device into an existing group will have the device installed in the group->devices, but not yet attached to the group's current domain. Move the group attachment logic from iommu_probe_device() and put it under the same mutex that updates the group->devices list so everything is atomic under the lock. We retain the two step setup of the default domain for the bus_iommu_probe() case solely so that we have a more complete view of the group when creating the default domain for boot time devices. This is not generally necessary with the current code structure but seems to be supporting some odd corner cases like alias RID's and IOMMU_RESV_DIRECT or driver bugs returning different default_domain types for the same group. During bus_iommu_probe() the group will have a device list but both group->default_domain and group->domain will be NULL. Reviewed-by: Lu Baolu Reviewed-by: Kevin Tian Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/10-v3-328044aa278c+45e49-iommu_probe_jgg@nvidia.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 78 +++++++++++++++++++++++---------------------------- 1 file changed, 35 insertions(+), 43 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 3129e5c50c87..42377f49ab87 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -131,6 +131,8 @@ 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 = \ @@ -469,14 +471,39 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list goto err_put_group; } + /* + * The gdev must be in the list before calling + * iommu_setup_default_domain() + */ list_add_tail(&gdev->list, &group->devices); - if (group_list && !group->default_domain && list_empty(&group->entry)) - list_add_tail(&group->entry, group_list); + 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_remove_gdev; + } else if (!group->default_domain && !group_list) { + ret = iommu_setup_default_domain(group, 0); + if (ret) + 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); mutex_unlock(&iommu_probe_device_lock); return 0; +err_remove_gdev: + list_del(&gdev->list); + __iommu_group_free_device(group, gdev); err_put_group: iommu_deinit_device(dev); mutex_unlock(&group->mutex); @@ -490,52 +517,17 @@ out_unlock: int iommu_probe_device(struct device *dev) { const struct iommu_ops *ops; - struct iommu_group *group; int ret; ret = __iommu_probe_device(dev, NULL); if (ret) - goto err_out; - - group = iommu_group_get(dev); - if (!group) { - ret = -ENODEV; - goto err_release; - } - - mutex_lock(&group->mutex); - - 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) { - ret = iommu_setup_default_domain(group, 0); - if (ret) - goto err_unlock; - } - - mutex_unlock(&group->mutex); - iommu_group_put(group); + return ret; ops = dev_iommu_ops(dev); if (ops->probe_finalize) ops->probe_finalize(dev); return 0; - -err_unlock: - mutex_unlock(&group->mutex); - iommu_group_put(group); -err_release: - iommu_release_device(dev); - -err_out: - return ret; - } static void __iommu_group_free_device(struct iommu_group *group, @@ -1815,11 +1807,6 @@ 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; @@ -1832,6 +1819,11 @@ 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); -- cgit v1.2.3 From 791c2b17fb4023f21c3cbf5f268af01d9b8cb7cc Mon Sep 17 00:00:00 2001 From: Robin Murphy Date: Thu, 13 Apr 2023 14:40:25 +0100 Subject: iommu: Optimise PCI SAC address trick Per the reasoning in commit 4bf7fda4dce2 ("iommu/dma: Add config for PCI SAC address trick") and its subsequent revert, this mechanism no longer serves its original purpose, but now only works around broken hardware/drivers in a way that is unfortunately too impactful to remove. This does not, however, prevent us from solving the performance impact which that workaround has on large-scale systems that don't need it. Once the 32-bit IOVA space fills up and a workload starts allocating and freeing on both sides of the boundary, the opportunistic SAC allocation can then end up spending significant time hunting down scattered fragments of free 32-bit space, or just reestablishing max32_alloc_size. This can easily be exacerbated by a change in allocation pattern, such as by changing the network MTU, which can increase pressure on the 32-bit space by leaving a large quantity of cached IOVAs which are now the wrong size to be recycled, but also won't be freed since the non-opportunistic allocations can still be satisfied from the whole 64-bit space without triggering the reclaim path. However, in the context of a workaround where smaller DMA addresses aren't simply a preference but a necessity, if we get to that point at all then in fact it's already the endgame. The nature of the allocator is currently such that the first IOVA we give to a device after the 32-bit space runs out will be the highest possible address for that device, ever. If that works, then great, we know we can optimise for speed by always allocating from the full range. And if it doesn't, then the worst has already happened and any brokenness is now showing, so there's little point in continuing to try to hide it. To that end, implement a flag to refine the SAC business into a per-device policy that can automatically get itself out of the way if and when it stops being useful. CC: Linus Torvalds CC: Jakub Kicinski Reviewed-by: John Garry Signed-off-by: Robin Murphy Tested-by: Vasant Hegde Tested-by: Jakub Kicinski Link: https://lore.kernel.org/r/b8502b115b915d2a3fabde367e099e39106686c8.1681392791.git.robin.murphy@arm.com Signed-off-by: Joerg Roedel --- drivers/iommu/dma-iommu.c | 26 ++++++++++++++++++++------ drivers/iommu/dma-iommu.h | 8 ++++++++ drivers/iommu/iommu.c | 3 +++ include/linux/iommu.h | 2 ++ 4 files changed, 33 insertions(+), 6 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index e57724163835..4b1a88f514c9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -660,7 +660,7 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain, { struct iommu_dma_cookie *cookie = domain->iova_cookie; struct iova_domain *iovad = &cookie->iovad; - unsigned long shift, iova_len, iova = 0; + unsigned long shift, iova_len, iova; if (cookie->type == IOMMU_DMA_MSI_COOKIE) { cookie->msi_iova += size; @@ -675,15 +675,29 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain, if (domain->geometry.force_aperture) dma_limit = min(dma_limit, (u64)domain->geometry.aperture_end); - /* Try to get PCI devices a SAC address */ - if (dma_limit > DMA_BIT_MASK(32) && !iommu_dma_forcedac && dev_is_pci(dev)) + /* + * Try to use all the 32-bit PCI addresses first. The original SAC vs. + * DAC reasoning loses relevance with PCIe, but enough hardware and + * firmware bugs are still lurking out there that it's safest not to + * venture into the 64-bit space until necessary. + * + * If your device goes wrong after seeing the notice then likely either + * its driver is not setting DMA masks accurately, the hardware has + * some inherent bug in handling >32-bit addresses, or not all the + * expected address bits are wired up between the device and the IOMMU. + */ + if (dma_limit > DMA_BIT_MASK(32) && dev->iommu->pci_32bit_workaround) { iova = alloc_iova_fast(iovad, iova_len, DMA_BIT_MASK(32) >> shift, false); + if (iova) + goto done; - if (!iova) - iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift, - true); + dev->iommu->pci_32bit_workaround = false; + dev_notice(dev, "Using %d-bit DMA addresses\n", bits_per(dma_limit)); + } + iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift, true); +done: return (dma_addr_t)iova << shift; } diff --git a/drivers/iommu/dma-iommu.h b/drivers/iommu/dma-iommu.h index 942790009292..c829f1f82a99 100644 --- a/drivers/iommu/dma-iommu.h +++ b/drivers/iommu/dma-iommu.h @@ -17,6 +17,10 @@ int iommu_dma_init_fq(struct iommu_domain *domain); void iommu_dma_get_resv_regions(struct device *dev, struct list_head *list); extern bool iommu_dma_forcedac; +static inline void iommu_dma_set_pci_32bit_workaround(struct device *dev) +{ + dev->iommu->pci_32bit_workaround = !iommu_dma_forcedac; +} #else /* CONFIG_IOMMU_DMA */ @@ -38,5 +42,9 @@ static inline void iommu_dma_get_resv_regions(struct device *dev, struct list_he { } +static inline void iommu_dma_set_pci_32bit_workaround(struct device *dev) +{ +} + #endif /* CONFIG_IOMMU_DMA */ #endif /* __DMA_IOMMU_H */ diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 42377f49ab87..e67f6562da73 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -499,6 +499,9 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list mutex_unlock(&group->mutex); mutex_unlock(&iommu_probe_device_lock); + if (dev_is_pci(dev)) + iommu_dma_set_pci_32bit_workaround(dev); + return 0; err_remove_gdev: diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d31642596675..b1dcb1b9b170 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -409,6 +409,7 @@ struct iommu_fault_param { * @priv: IOMMU Driver private data * @max_pasids: number of PASIDs this device can consume * @attach_deferred: the dma domain attachment is deferred + * @pci_32bit_workaround: Limit DMA allocations to 32-bit IOVAs * * TODO: migrate other per device data pointers under iommu_dev_data, e.g. * struct iommu_group *iommu_group; @@ -422,6 +423,7 @@ struct dev_iommu { void *priv; u32 max_pasids; u32 attach_deferred:1; + u32 pci_32bit_workaround:1; }; int iommu_device_register(struct iommu_device *iommu, -- cgit v1.2.3 From 534103bcd52ca9c1fecbc70e717b4a538dc4ded8 Mon Sep 17 00:00:00 2001 From: Daniel Marcovitch Date: Fri, 9 Jun 2023 10:51:45 +0000 Subject: iommu/amd/iommu_v2: Fix pasid_state refcount dec hit 0 warning on pasid unbind When unbinding pasid - a race condition exists vs outstanding page faults. To prevent this, the pasid_state object contains a refcount. * set to 1 on pasid bind * incremented on each ppr notification start * decremented on each ppr notification done * decremented on pasid unbind Since refcount_dec assumes that refcount will never reach 0: the current implementation causes the following to be invoked on pasid unbind: REFCOUNT_WARN("decrement hit 0; leaking memory") Fix this issue by changing refcount_dec to refcount_dec_and_test to explicitly handle refcount=1. Fixes: 8bc54824da4e ("iommu/amd: Convert from atomic_t to refcount_t on pasid_state->count") Signed-off-by: Daniel Marcovitch Signed-off-by: Vasant Hegde Link: https://lore.kernel.org/r/20230609105146.7773-2-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/iommu_v2.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/iommu_v2.c b/drivers/iommu/amd/iommu_v2.c index 261352a23271..65d78d7e0440 100644 --- a/drivers/iommu/amd/iommu_v2.c +++ b/drivers/iommu/amd/iommu_v2.c @@ -262,8 +262,8 @@ static void put_pasid_state(struct pasid_state *pasid_state) static void put_pasid_state_wait(struct pasid_state *pasid_state) { - refcount_dec(&pasid_state->count); - wait_event(pasid_state->wq, !refcount_read(&pasid_state->count)); + if (!refcount_dec_and_test(&pasid_state->count)) + wait_event(pasid_state->wq, !refcount_read(&pasid_state->count)); free_pasid_state(pasid_state); } -- cgit v1.2.3 From d269ab61f4f8c56999fd26ade86def26db1fed09 Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Fri, 9 Jun 2023 10:51:46 +0000 Subject: iommu/amd/iommu_v2: Clear pasid state in free path Clear pasid state in device amd_iommu_free_device() path. It will make sure no new ppr notifier is registered in free path. Suggested-by: Jason Gunthorpe Signed-off-by: Vasant Hegde Link: https://lore.kernel.org/r/20230609105146.7773-3-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/iommu_v2.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/amd/iommu_v2.c b/drivers/iommu/amd/iommu_v2.c index 65d78d7e0440..c5825e0a6770 100644 --- a/drivers/iommu/amd/iommu_v2.c +++ b/drivers/iommu/amd/iommu_v2.c @@ -327,6 +327,9 @@ static void free_pasid_states(struct device_state *dev_state) put_pasid_state(pasid_state); + /* Clear the pasid state so that the pasid can be re-used */ + clear_pasid_state(dev_state, pasid_state->pasid); + /* * This will call the mn_release function and * unbind the PASID -- cgit v1.2.3 From 6df63b7ebdaf5fcd75dceedf6967d0761e56eca1 Mon Sep 17 00:00:00 2001 From: Jonas Karlman Date: Sat, 17 Jun 2023 18:25:45 +0000 Subject: iommu: rockchip: Fix directory table address encoding The physical address to the directory table is currently encoded using the following bit layout for IOMMU v2. 31:12 - Address bit 31:0 11: 4 - Address bit 39:32 This is also the bit layout used by the vendor kernel. However, testing has shown that addresses to the directory/page tables and memory pages are all encoded using the same bit layout. IOMMU v1: 31:12 - Address bit 31:0 IOMMU v2: 31:12 - Address bit 31:0 11: 8 - Address bit 35:32 7: 4 - Address bit 39:36 Change to use the mk_dtentries ops to encode the directory table address correctly. The value written to DTE_ADDR may include the valid bit set, a bit that is ignored and DTE_ADDR reg read it back as 0. This also update the bit layout comment for the page address and the number of nybbles that are read back for DTE_ADDR comment. These changes render the dte_addr_phys and dma_addr_dte ops unused and is removed. Fixes: 227014b33f62 ("iommu: rockchip: Add internal ops to handle variants") Fixes: c55356c534aa ("iommu: rockchip: Add support for iommu v2") Fixes: c987b65a574f ("iommu/rockchip: Fix physical address decoding") Signed-off-by: Jonas Karlman Reviewed-by: Robin Murphy Link: https://lore.kernel.org/r/20230617182540.3091374-2-jonas@kwiboo.se Signed-off-by: Joerg Roedel --- drivers/iommu/rockchip-iommu.c | 43 +++++------------------------------------- 1 file changed, 5 insertions(+), 38 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c index 4054030c3237..ae42959bc490 100644 --- a/drivers/iommu/rockchip-iommu.c +++ b/drivers/iommu/rockchip-iommu.c @@ -98,8 +98,6 @@ struct rk_iommu_ops { phys_addr_t (*pt_address)(u32 dte); u32 (*mk_dtentries)(dma_addr_t pt_dma); u32 (*mk_ptentries)(phys_addr_t page, int prot); - phys_addr_t (*dte_addr_phys)(u32 addr); - u32 (*dma_addr_dte)(dma_addr_t dt_dma); u64 dma_bit_mask; }; @@ -278,8 +276,8 @@ static u32 rk_mk_pte(phys_addr_t page, int prot) /* * In v2: * 31:12 - Page address bit 31:0 - * 11:9 - Page address bit 34:32 - * 8:4 - Page address bit 39:35 + * 11: 8 - Page address bit 35:32 + * 7: 4 - Page address bit 39:36 * 3 - Security * 2 - Writable * 1 - Readable @@ -506,7 +504,7 @@ static int rk_iommu_force_reset(struct rk_iommu *iommu) /* * Check if register DTE_ADDR is working by writing DTE_ADDR_DUMMY - * and verifying that upper 5 nybbles are read back. + * and verifying that upper 5 (v1) or 7 (v2) nybbles are read back. */ for (i = 0; i < iommu->num_mmu; i++) { dte_addr = rk_ops->pt_address(DTE_ADDR_DUMMY); @@ -531,33 +529,6 @@ static int rk_iommu_force_reset(struct rk_iommu *iommu) return 0; } -static inline phys_addr_t rk_dte_addr_phys(u32 addr) -{ - return (phys_addr_t)addr; -} - -static inline u32 rk_dma_addr_dte(dma_addr_t dt_dma) -{ - return dt_dma; -} - -#define DT_HI_MASK GENMASK_ULL(39, 32) -#define DTE_BASE_HI_MASK GENMASK(11, 4) -#define DT_SHIFT 28 - -static inline phys_addr_t rk_dte_addr_phys_v2(u32 addr) -{ - u64 addr64 = addr; - return (phys_addr_t)(addr64 & RK_DTE_PT_ADDRESS_MASK) | - ((addr64 & DTE_BASE_HI_MASK) << DT_SHIFT); -} - -static inline u32 rk_dma_addr_dte_v2(dma_addr_t dt_dma) -{ - return (dt_dma & RK_DTE_PT_ADDRESS_MASK) | - ((dt_dma & DT_HI_MASK) >> DT_SHIFT); -} - static void log_iova(struct rk_iommu *iommu, int index, dma_addr_t iova) { void __iomem *base = iommu->bases[index]; @@ -577,7 +548,7 @@ static void log_iova(struct rk_iommu *iommu, int index, dma_addr_t iova) page_offset = rk_iova_page_offset(iova); mmu_dte_addr = rk_iommu_read(base, RK_MMU_DTE_ADDR); - mmu_dte_addr_phys = rk_ops->dte_addr_phys(mmu_dte_addr); + mmu_dte_addr_phys = rk_ops->pt_address(mmu_dte_addr); dte_addr_phys = mmu_dte_addr_phys + (4 * dte_index); dte_addr = phys_to_virt(dte_addr_phys); @@ -967,7 +938,7 @@ static int rk_iommu_enable(struct rk_iommu *iommu) for (i = 0; i < iommu->num_mmu; i++) { rk_iommu_write(iommu->bases[i], RK_MMU_DTE_ADDR, - rk_ops->dma_addr_dte(rk_domain->dt_dma)); + rk_ops->mk_dtentries(rk_domain->dt_dma)); rk_iommu_base_command(iommu->bases[i], RK_MMU_CMD_ZAP_CACHE); rk_iommu_write(iommu->bases[i], RK_MMU_INT_MASK, RK_MMU_IRQ_MASK); } @@ -1405,8 +1376,6 @@ static struct rk_iommu_ops iommu_data_ops_v1 = { .pt_address = &rk_dte_pt_address, .mk_dtentries = &rk_mk_dte, .mk_ptentries = &rk_mk_pte, - .dte_addr_phys = &rk_dte_addr_phys, - .dma_addr_dte = &rk_dma_addr_dte, .dma_bit_mask = DMA_BIT_MASK(32), }; @@ -1414,8 +1383,6 @@ static struct rk_iommu_ops iommu_data_ops_v2 = { .pt_address = &rk_dte_pt_address_v2, .mk_dtentries = &rk_mk_dte_v2, .mk_ptentries = &rk_mk_pte_v2, - .dte_addr_phys = &rk_dte_addr_phys_v2, - .dma_addr_dte = &rk_dma_addr_dte_v2, .dma_bit_mask = DMA_BIT_MASK(40), }; -- cgit v1.2.3 From 2a7e6400f72b9963cbaaac8a60ccc54d2b8c0541 Mon Sep 17 00:00:00 2001 From: Jonas Karlman Date: Sat, 17 Jun 2023 18:25:47 +0000 Subject: iommu: rockchip: Allocate tables from all available memory for IOMMU v2 IOMMU v2 found in newer Rockchip SoCs, e.g. RK356x and RK3588, support placing directory and page tables in up to 40-bit addressable physical memory. Remove the use of GFP_DMA32 flag for IOMMU v2 now that the physical address to the directory table is correctly written to DTE_ADDR. Signed-off-by: Jonas Karlman Reviewed-by: Robin Murphy Link: https://lore.kernel.org/r/20230617182540.3091374-3-jonas@kwiboo.se Signed-off-by: Joerg Roedel --- drivers/iommu/rockchip-iommu.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/rockchip-iommu.c b/drivers/iommu/rockchip-iommu.c index ae42959bc490..8ff69fbf9f65 100644 --- a/drivers/iommu/rockchip-iommu.c +++ b/drivers/iommu/rockchip-iommu.c @@ -99,6 +99,7 @@ struct rk_iommu_ops { u32 (*mk_dtentries)(dma_addr_t pt_dma); u32 (*mk_ptentries)(phys_addr_t page, int prot); u64 dma_bit_mask; + gfp_t gfp_flags; }; struct rk_iommu { @@ -727,7 +728,7 @@ static u32 *rk_dte_get_page_table(struct rk_iommu_domain *rk_domain, if (rk_dte_is_pt_valid(dte)) goto done; - page_table = (u32 *)get_zeroed_page(GFP_ATOMIC | GFP_DMA32); + page_table = (u32 *)get_zeroed_page(GFP_ATOMIC | rk_ops->gfp_flags); if (!page_table) return ERR_PTR(-ENOMEM); @@ -1076,7 +1077,7 @@ static struct iommu_domain *rk_iommu_domain_alloc(unsigned type) * Each level1 (dt) and level2 (pt) table has 1024 4-byte entries. * Allocate one 4 KiB page for each table. */ - rk_domain->dt = (u32 *)get_zeroed_page(GFP_KERNEL | GFP_DMA32); + rk_domain->dt = (u32 *)get_zeroed_page(GFP_KERNEL | rk_ops->gfp_flags); if (!rk_domain->dt) goto err_free_domain; @@ -1377,6 +1378,7 @@ static struct rk_iommu_ops iommu_data_ops_v1 = { .mk_dtentries = &rk_mk_dte, .mk_ptentries = &rk_mk_pte, .dma_bit_mask = DMA_BIT_MASK(32), + .gfp_flags = GFP_DMA32, }; static struct rk_iommu_ops iommu_data_ops_v2 = { @@ -1384,6 +1386,7 @@ static struct rk_iommu_ops iommu_data_ops_v2 = { .mk_dtentries = &rk_mk_dte_v2, .mk_ptentries = &rk_mk_pte_v2, .dma_bit_mask = DMA_BIT_MASK(40), + .gfp_flags = 0, }; static const struct of_device_id rk_iommu_dt_ids[] = { -- cgit v1.2.3 From 386ae59bd7db4eafc74bdfdcf60182c2171aea2a Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:16:23 +0000 Subject: iommu/amd: Generalize log overflow handling Each IOMMU has three log buffers (Event, GA and PPR log). Once a buffer becomes full, IOMMU generates an interrupt with the corresponding overflow status bit, and stop processing the log. To handle an overflow, the IOMMU driver needs to disable the log, clear the overflow status bit, and re-enable the log. This procedure is same among all types of log buffer except it uses different overflow status bit and enabling bit. Hence, to consolidate the log buffer restarting logic, introduce a helper function amd_iommu_restart_log(), which caller can specify parameters specific for each type of log buffer. Also rename MMIO_STATUS_EVT_OVERFLOW_INT_MASK as MMIO_STATUS_EVT_OVERFLOW_MASK. Reviewed-by: Jerry Snitselaar Reviewed-by: Suravee Suthikulpanit Reviewed-by: Joao Martins Signed-off-by: Vasant Hegde Link: https://lore.kernel.org/r/20230628051624.5792-2-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu_types.h | 3 ++- drivers/iommu/amd/init.c | 51 +++++++++++++++++++++++-------------- drivers/iommu/amd/iommu.c | 4 +-- 3 files changed, 36 insertions(+), 22 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index dc1db6167927..7c436ba2a0a2 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -120,9 +120,10 @@ #define PASID_MASK 0x0000ffff /* MMIO status bits */ -#define MMIO_STATUS_EVT_OVERFLOW_INT_MASK BIT(0) +#define MMIO_STATUS_EVT_OVERFLOW_MASK BIT(0) #define MMIO_STATUS_EVT_INT_MASK BIT(1) #define MMIO_STATUS_COM_WAIT_INT_MASK BIT(2) +#define MMIO_STATUS_EVT_RUN_MASK BIT(3) #define MMIO_STATUS_PPR_INT_MASK BIT(6) #define MMIO_STATUS_GALOG_RUN_MASK BIT(8) #define MMIO_STATUS_GALOG_OVERFLOW_MASK BIT(9) diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index ea0f1ab94178..7fab6ecb6295 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -752,38 +752,51 @@ static int __init alloc_command_buffer(struct amd_iommu *iommu) return iommu->cmd_buf ? 0 : -ENOMEM; } +/* + * Interrupt handler has processed all pending events and adjusted head + * and tail pointer. Reset overflow mask and restart logging again. + */ +static void amd_iommu_restart_log(struct amd_iommu *iommu, const char *evt_type, + u8 cntrl_intr, u8 cntrl_log, + u32 status_run_mask, u32 status_overflow_mask) +{ + u32 status; + + status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); + if (status & status_run_mask) + return; + + pr_info_ratelimited("IOMMU %s log restarting\n", evt_type); + + iommu_feature_disable(iommu, cntrl_log); + iommu_feature_disable(iommu, cntrl_intr); + + writel(status_overflow_mask, iommu->mmio_base + MMIO_STATUS_OFFSET); + + iommu_feature_enable(iommu, cntrl_intr); + iommu_feature_enable(iommu, cntrl_log); +} + /* * This function restarts event logging in case the IOMMU experienced * an event log buffer overflow. */ void amd_iommu_restart_event_logging(struct amd_iommu *iommu) { - iommu_feature_disable(iommu, CONTROL_EVT_LOG_EN); - iommu_feature_enable(iommu, CONTROL_EVT_LOG_EN); + amd_iommu_restart_log(iommu, "Event", CONTROL_EVT_INT_EN, + CONTROL_EVT_LOG_EN, MMIO_STATUS_EVT_RUN_MASK, + MMIO_STATUS_EVT_OVERFLOW_MASK); } /* * This function restarts event logging in case the IOMMU experienced - * an GA log overflow. + * GA log overflow. */ void amd_iommu_restart_ga_log(struct amd_iommu *iommu) { - u32 status; - - status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); - if (status & MMIO_STATUS_GALOG_RUN_MASK) - return; - - pr_info_ratelimited("IOMMU GA Log restarting\n"); - - iommu_feature_disable(iommu, CONTROL_GALOG_EN); - iommu_feature_disable(iommu, CONTROL_GAINT_EN); - - writel(MMIO_STATUS_GALOG_OVERFLOW_MASK, - iommu->mmio_base + MMIO_STATUS_OFFSET); - - iommu_feature_enable(iommu, CONTROL_GAINT_EN); - iommu_feature_enable(iommu, CONTROL_GALOG_EN); + amd_iommu_restart_log(iommu, "GA", CONTROL_GAINT_EN, + CONTROL_GALOG_EN, MMIO_STATUS_GALOG_RUN_MASK, + MMIO_STATUS_GALOG_OVERFLOW_MASK); } /* diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index c3b58a8389b9..86c0ae34373b 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -842,7 +842,7 @@ amd_iommu_set_pci_msi_domain(struct device *dev, struct amd_iommu *iommu) { } #endif /* !CONFIG_IRQ_REMAP */ #define AMD_IOMMU_INT_MASK \ - (MMIO_STATUS_EVT_OVERFLOW_INT_MASK | \ + (MMIO_STATUS_EVT_OVERFLOW_MASK | \ MMIO_STATUS_EVT_INT_MASK | \ MMIO_STATUS_PPR_INT_MASK | \ MMIO_STATUS_GALOG_OVERFLOW_MASK | \ @@ -881,7 +881,7 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data) } #endif - if (status & MMIO_STATUS_EVT_OVERFLOW_INT_MASK) { + if (status & MMIO_STATUS_EVT_OVERFLOW_MASK) { pr_info_ratelimited("IOMMU event log overflow\n"); amd_iommu_restart_event_logging(iommu); } -- cgit v1.2.3 From 274c2218b8b29ea2e0404f97d52040deb16d42dd Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:16:24 +0000 Subject: iommu/amd: Handle PPR log overflow Some ATS-capable peripherals can issue requests to the processor to service peripheral page requests using PCIe PRI (the Page Request Interface). IOMMU supports PRI using PPR log buffer. IOMMU writes PRI request to PPR log buffer and sends PPR interrupt to host. When there is no space in the PPR log buffer (PPR log overflow) it will set PprOverflow bit in 'MMIO Offset 2020h IOMMU Status Register'. When this happens PPR log needs to be restarted as specified in IOMMU spec [1] section 2.6.2. When handling the event it just resumes the PPR log without resizing (similar to the way event and GA log overflow is handled). Failing to handle PPR overflow means device may not work properly as IOMMU stops processing new PPR events from device. [1] https://www.amd.com/system/files/TechDocs/48882_3.07_PUB.pdf Reviewed-by: Jerry Snitselaar Reviewed-by: Suravee Suthikulpanit Reviewed-by: Joao Martins Signed-off-by: Vasant Hegde Link: https://lore.kernel.org/r/20230628051624.5792-3-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu.h | 1 + drivers/iommu/amd/amd_iommu_types.h | 2 ++ drivers/iommu/amd/init.c | 11 +++++++++++ drivers/iommu/amd/iommu.c | 9 ++++++++- 4 files changed, 22 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h index 0c35018239ce..8c61c19dabc4 100644 --- a/drivers/iommu/amd/amd_iommu.h +++ b/drivers/iommu/amd/amd_iommu.h @@ -16,6 +16,7 @@ irqreturn_t amd_iommu_int_handler(int irq, void *data); void amd_iommu_apply_erratum_63(struct amd_iommu *iommu, u16 devid); void amd_iommu_restart_event_logging(struct amd_iommu *iommu); void amd_iommu_restart_ga_log(struct amd_iommu *iommu); +void amd_iommu_restart_ppr_log(struct amd_iommu *iommu); int amd_iommu_init_devices(void); void amd_iommu_uninit_devices(void); void amd_iommu_init_notifier(void); diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index 7c436ba2a0a2..c8150eaf5c0e 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -124,7 +124,9 @@ #define MMIO_STATUS_EVT_INT_MASK BIT(1) #define MMIO_STATUS_COM_WAIT_INT_MASK BIT(2) #define MMIO_STATUS_EVT_RUN_MASK BIT(3) +#define MMIO_STATUS_PPR_OVERFLOW_MASK BIT(5) #define MMIO_STATUS_PPR_INT_MASK BIT(6) +#define MMIO_STATUS_PPR_RUN_MASK BIT(7) #define MMIO_STATUS_GALOG_RUN_MASK BIT(8) #define MMIO_STATUS_GALOG_OVERFLOW_MASK BIT(9) #define MMIO_STATUS_GALOG_INT_MASK BIT(10) diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index 7fab6ecb6295..e78d7c4f41bd 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -799,6 +799,17 @@ void amd_iommu_restart_ga_log(struct amd_iommu *iommu) MMIO_STATUS_GALOG_OVERFLOW_MASK); } +/* + * This function restarts ppr logging in case the IOMMU experienced + * PPR log overflow. + */ +void amd_iommu_restart_ppr_log(struct amd_iommu *iommu) +{ + amd_iommu_restart_log(iommu, "PPR", CONTROL_PPRINT_EN, + CONTROL_PPRLOG_EN, MMIO_STATUS_PPR_RUN_MASK, + MMIO_STATUS_PPR_OVERFLOW_MASK); +} + /* * This function resets the command buffer if the IOMMU stopped fetching * commands from it. diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index 86c0ae34373b..f8316901a178 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -844,6 +844,7 @@ amd_iommu_set_pci_msi_domain(struct device *dev, struct amd_iommu *iommu) { } #define AMD_IOMMU_INT_MASK \ (MMIO_STATUS_EVT_OVERFLOW_MASK | \ MMIO_STATUS_EVT_INT_MASK | \ + MMIO_STATUS_PPR_OVERFLOW_MASK | \ MMIO_STATUS_PPR_INT_MASK | \ MMIO_STATUS_GALOG_OVERFLOW_MASK | \ MMIO_STATUS_GALOG_INT_MASK) @@ -863,11 +864,17 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data) iommu_poll_events(iommu); } - if (status & MMIO_STATUS_PPR_INT_MASK) { + if (status & (MMIO_STATUS_PPR_INT_MASK | + MMIO_STATUS_PPR_OVERFLOW_MASK)) { pr_devel("Processing IOMMU PPR Log\n"); iommu_poll_ppr_log(iommu); } + if (status & MMIO_STATUS_PPR_OVERFLOW_MASK) { + pr_info_ratelimited("IOMMU PPR log overflow\n"); + amd_iommu_restart_ppr_log(iommu); + } + #ifdef CONFIG_IRQ_REMAP if (status & (MMIO_STATUS_GALOG_INT_MASK | MMIO_STATUS_GALOG_OVERFLOW_MASK)) { -- cgit v1.2.3 From 2379f3485239095b8782353c2431f3c97c275986 Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:32:21 +0000 Subject: iommu/amd: Refactor IOMMU interrupt handling logic for Event, PPR, and GA logs The AMD IOMMU has three log buffers (i.e. Event, PPR, and GA). The IOMMU driver processes these log entries when it receive an IOMMU interrupt. Then, it needs to clear the corresponding interrupt status bits. Also, when an overflow occurs, it needs to handle the log overflow by clearing the specific overflow status bit and restart the log. Since, logic for handling these logs is the same, refactor the code into a helper function called amd_iommu_handle_irq(), which handles the steps described. Then, reuse it for all types of log. Reviewed-by: Suravee Suthikulpanit Signed-off-by: Vasant Hegde Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230628053222.5962-2-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu.h | 3 ++ drivers/iommu/amd/iommu.c | 93 +++++++++++++++++++++++-------------------- 2 files changed, 53 insertions(+), 43 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h index 8c61c19dabc4..f23ad6043f04 100644 --- a/drivers/iommu/amd/amd_iommu.h +++ b/drivers/iommu/amd/amd_iommu.h @@ -12,6 +12,9 @@ #include "amd_iommu_types.h" irqreturn_t amd_iommu_int_thread(int irq, void *data); +irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data); +irqreturn_t amd_iommu_int_thread_pprlog(int irq, void *data); +irqreturn_t amd_iommu_int_thread_galog(int irq, void *data); irqreturn_t amd_iommu_int_handler(int irq, void *data); void amd_iommu_apply_erratum_63(struct amd_iommu *iommu, u16 devid); void amd_iommu_restart_event_logging(struct amd_iommu *iommu); diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index f8316901a178..56b6cf8bf03f 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -841,57 +841,27 @@ static inline void amd_iommu_set_pci_msi_domain(struct device *dev, struct amd_iommu *iommu) { } #endif /* !CONFIG_IRQ_REMAP */ -#define AMD_IOMMU_INT_MASK \ - (MMIO_STATUS_EVT_OVERFLOW_MASK | \ - MMIO_STATUS_EVT_INT_MASK | \ - MMIO_STATUS_PPR_OVERFLOW_MASK | \ - MMIO_STATUS_PPR_INT_MASK | \ - MMIO_STATUS_GALOG_OVERFLOW_MASK | \ - MMIO_STATUS_GALOG_INT_MASK) - -irqreturn_t amd_iommu_int_thread(int irq, void *data) +static void amd_iommu_handle_irq(void *data, const char *evt_type, + u32 int_mask, u32 overflow_mask, + void (*int_handler)(struct amd_iommu *), + void (*overflow_handler)(struct amd_iommu *)) { struct amd_iommu *iommu = (struct amd_iommu *) data; u32 status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); + u32 mask = int_mask | overflow_mask; - while (status & AMD_IOMMU_INT_MASK) { + while (status & mask) { /* Enable interrupt sources again */ - writel(AMD_IOMMU_INT_MASK, - iommu->mmio_base + MMIO_STATUS_OFFSET); - - if (status & MMIO_STATUS_EVT_INT_MASK) { - pr_devel("Processing IOMMU Event Log\n"); - iommu_poll_events(iommu); - } - - if (status & (MMIO_STATUS_PPR_INT_MASK | - MMIO_STATUS_PPR_OVERFLOW_MASK)) { - pr_devel("Processing IOMMU PPR Log\n"); - iommu_poll_ppr_log(iommu); - } - - if (status & MMIO_STATUS_PPR_OVERFLOW_MASK) { - pr_info_ratelimited("IOMMU PPR log overflow\n"); - amd_iommu_restart_ppr_log(iommu); - } - -#ifdef CONFIG_IRQ_REMAP - if (status & (MMIO_STATUS_GALOG_INT_MASK | - MMIO_STATUS_GALOG_OVERFLOW_MASK)) { - pr_devel("Processing IOMMU GA Log\n"); - iommu_poll_ga_log(iommu); - } + writel(mask, iommu->mmio_base + MMIO_STATUS_OFFSET); - if (status & MMIO_STATUS_GALOG_OVERFLOW_MASK) { - pr_info_ratelimited("IOMMU GA Log overflow\n"); - amd_iommu_restart_ga_log(iommu); + if (int_handler) { + pr_devel("Processing IOMMU (ivhd%d) %s Log\n", + iommu->index, evt_type); + int_handler(iommu); } -#endif - if (status & MMIO_STATUS_EVT_OVERFLOW_MASK) { - pr_info_ratelimited("IOMMU event log overflow\n"); - amd_iommu_restart_event_logging(iommu); - } + if ((status & overflow_mask) && overflow_handler) + overflow_handler(iommu); /* * Hardware bug: ERBT1312 @@ -908,6 +878,43 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data) */ status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET); } +} + +irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data) +{ + amd_iommu_handle_irq(data, "Evt", MMIO_STATUS_EVT_INT_MASK, + MMIO_STATUS_EVT_OVERFLOW_MASK, + iommu_poll_events, amd_iommu_restart_event_logging); + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread_pprlog(int irq, void *data) +{ + amd_iommu_handle_irq(data, "PPR", MMIO_STATUS_PPR_INT_MASK, + MMIO_STATUS_PPR_OVERFLOW_MASK, + iommu_poll_ppr_log, amd_iommu_restart_ppr_log); + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread_galog(int irq, void *data) +{ +#ifdef CONFIG_IRQ_REMAP + amd_iommu_handle_irq(data, "GA", MMIO_STATUS_GALOG_INT_MASK, + MMIO_STATUS_GALOG_OVERFLOW_MASK, + iommu_poll_ga_log, amd_iommu_restart_ga_log); +#endif + + return IRQ_HANDLED; +} + +irqreturn_t amd_iommu_int_thread(int irq, void *data) +{ + amd_iommu_int_thread_evtlog(irq, data); + amd_iommu_int_thread_pprlog(irq, data); + amd_iommu_int_thread_galog(irq, data); + return IRQ_HANDLED; } -- cgit v1.2.3 From e5ebd90d1b5cbfa0df9b911cd3f24c1bb84c05dd Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:32:22 +0000 Subject: iommu/amd: Enable separate interrupt for PPR and GA log AMD IOMMU has three log buffers (i.e. Event, PPR, and GA). These logs can be configured to generate different interrupts when an entry is inserted into a log buffer. However, current implementation share single interrupt to handle all three logs. With increasing usages of the GA (for IOMMU AVIC) and PPR logs (for IOMMUv2 APIs and SVA), interrupt sharing could potentially become performance bottleneck. Hence, separate IOMMU interrupt into use three separate vectors and irq threads with corresponding name, which will be displayed in the /proc/interrupts as "AMD-Vi-[Evt/PPR/GA]", where "x" is an IOMMU id. Note that this patch changes interrupt handling only in IOMMU x2apic mode (MMIO 0x18[IntCapXTEn]=1). In legacy mode it will continue to use single MSI interrupt. Signed-off-by: Vasant Hegde Reviewed-by: Alexey Kardashevskiy Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230628053222.5962-3-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu_types.h | 9 +++++++ drivers/iommu/amd/init.c | 50 ++++++++++++++++++++++++++++--------- 2 files changed, 47 insertions(+), 12 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index c8150eaf5c0e..781ab9c2ea70 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -705,12 +705,21 @@ struct amd_iommu { /* event buffer virtual address */ u8 *evt_buf; + /* Name for event log interrupt */ + unsigned char evt_irq_name[16]; + /* Base of the PPR log, if present */ u8 *ppr_log; + /* Name for PPR log interrupt */ + unsigned char ppr_irq_name[16]; + /* Base of the GA log, if present */ u8 *ga_log; + /* Name for GA log interrupt */ + unsigned char ga_irq_name[16]; + /* Tail of the GA log, if present */ u8 *ga_log_tail; diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index e78d7c4f41bd..30f09e786f7b 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -2335,6 +2335,7 @@ static int intcapxt_irqdomain_alloc(struct irq_domain *domain, unsigned int virq struct irq_data *irqd = irq_domain_get_irq_data(domain, i); irqd->chip = &intcapxt_controller; + irqd->hwirq = info->hwirq; irqd->chip_data = info->data; __irq_set_handler(i, handle_edge_irq, 0, "edge"); } @@ -2361,22 +2362,14 @@ static void intcapxt_unmask_irq(struct irq_data *irqd) xt.destid_0_23 = cfg->dest_apicid & GENMASK(23, 0); xt.destid_24_31 = cfg->dest_apicid >> 24; - /** - * Current IOMMU implementation uses the same IRQ for all - * 3 IOMMU interrupts. - */ - writeq(xt.capxt, iommu->mmio_base + MMIO_INTCAPXT_EVT_OFFSET); - writeq(xt.capxt, iommu->mmio_base + MMIO_INTCAPXT_PPR_OFFSET); - writeq(xt.capxt, iommu->mmio_base + MMIO_INTCAPXT_GALOG_OFFSET); + writeq(xt.capxt, iommu->mmio_base + irqd->hwirq); } static void intcapxt_mask_irq(struct irq_data *irqd) { struct amd_iommu *iommu = irqd->chip_data; - writeq(0, iommu->mmio_base + MMIO_INTCAPXT_EVT_OFFSET); - writeq(0, iommu->mmio_base + MMIO_INTCAPXT_PPR_OFFSET); - writeq(0, iommu->mmio_base + MMIO_INTCAPXT_GALOG_OFFSET); + writeq(0, iommu->mmio_base + irqd->hwirq); } @@ -2439,7 +2432,8 @@ static struct irq_domain *iommu_get_irqdomain(void) return iommu_irqdomain; } -static int iommu_setup_intcapxt(struct amd_iommu *iommu) +static int __iommu_setup_intcapxt(struct amd_iommu *iommu, const char *devname, + int hwirq, irq_handler_t thread_fn) { struct irq_domain *domain; struct irq_alloc_info info; @@ -2453,6 +2447,7 @@ static int iommu_setup_intcapxt(struct amd_iommu *iommu) init_irq_alloc_info(&info, NULL); info.type = X86_IRQ_ALLOC_TYPE_AMDVI; info.data = iommu; + info.hwirq = hwirq; irq = irq_domain_alloc_irqs(domain, 1, node, &info); if (irq < 0) { @@ -2461,7 +2456,7 @@ static int iommu_setup_intcapxt(struct amd_iommu *iommu) } ret = request_threaded_irq(irq, amd_iommu_int_handler, - amd_iommu_int_thread, 0, "AMD-Vi", iommu); + thread_fn, 0, devname, iommu); if (ret) { irq_domain_free_irqs(irq, 1); irq_domain_remove(domain); @@ -2471,6 +2466,37 @@ static int iommu_setup_intcapxt(struct amd_iommu *iommu) return 0; } +static int iommu_setup_intcapxt(struct amd_iommu *iommu) +{ + int ret; + + snprintf(iommu->evt_irq_name, sizeof(iommu->evt_irq_name), + "AMD-Vi%d-Evt", iommu->index); + ret = __iommu_setup_intcapxt(iommu, iommu->evt_irq_name, + MMIO_INTCAPXT_EVT_OFFSET, + amd_iommu_int_thread_evtlog); + if (ret) + return ret; + + snprintf(iommu->ppr_irq_name, sizeof(iommu->ppr_irq_name), + "AMD-Vi%d-PPR", iommu->index); + ret = __iommu_setup_intcapxt(iommu, iommu->ppr_irq_name, + MMIO_INTCAPXT_PPR_OFFSET, + amd_iommu_int_thread_pprlog); + if (ret) + return ret; + +#ifdef CONFIG_IRQ_REMAP + snprintf(iommu->ga_irq_name, sizeof(iommu->ga_irq_name), + "AMD-Vi%d-GA", iommu->index); + ret = __iommu_setup_intcapxt(iommu, iommu->ga_irq_name, + MMIO_INTCAPXT_GALOG_OFFSET, + amd_iommu_int_thread_galog); +#endif + + return ret; +} + static int iommu_init_irq(struct amd_iommu *iommu) { int ret; -- cgit v1.2.3 From 7827a2689e79bcd3096b4b49b8a8beb8574fd4aa Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:45:52 +0000 Subject: iommu/amd: Disable PPR log/interrupt in iommu_disable() Similar to other logs, disable PPR log/interrupt in iommu_disable() path. Reviewed-by: Suravee Suthikulpanit Signed-off-by: Vasant Hegde Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230628054554.6131-2-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/init.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index 30f09e786f7b..94c91b6ee603 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -483,6 +483,10 @@ static void iommu_disable(struct amd_iommu *iommu) iommu_feature_disable(iommu, CONTROL_GALOG_EN); iommu_feature_disable(iommu, CONTROL_GAINT_EN); + /* Disable IOMMU PPR logging */ + iommu_feature_disable(iommu, CONTROL_PPRLOG_EN); + iommu_feature_disable(iommu, CONTROL_PPRINT_EN); + /* Disable IOMMU hardware itself */ iommu_feature_disable(iommu, CONTROL_IOMMU_EN); -- cgit v1.2.3 From f52c895a2de8697108c6385e9695061585690dc8 Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:45:53 +0000 Subject: iommu/amd: Consolidate PPR log enablement Move PPR log interrupt bit setting to iommu_enable_ppr_log(). Also rearrange iommu_enable_ppr_log() such that PPREn bit is enabled before enabling PPRLog and PPRInt bits. So that when PPRLog bit is set it will clear the PPRLogOverflow bit and sets the PPRLogRun bit in the IOMMU Status Register [MMIO Offset 2020h]. Reviewed-by: Suravee Suthikulpanit Signed-off-by: Vasant Hegde Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230628054554.6131-3-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/init.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index 94c91b6ee603..f26c077eb9ea 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -934,6 +934,8 @@ static void iommu_enable_ppr_log(struct amd_iommu *iommu) if (iommu->ppr_log == NULL) return; + iommu_feature_enable(iommu, CONTROL_PPR_EN); + entry = iommu_virt_to_phys(iommu->ppr_log) | PPR_LOG_SIZE_512; memcpy_toio(iommu->mmio_base + MMIO_PPR_LOG_OFFSET, @@ -944,7 +946,7 @@ static void iommu_enable_ppr_log(struct amd_iommu *iommu) writel(0x00, iommu->mmio_base + MMIO_PPR_TAIL_OFFSET); iommu_feature_enable(iommu, CONTROL_PPRLOG_EN); - iommu_feature_enable(iommu, CONTROL_PPR_EN); + iommu_feature_enable(iommu, CONTROL_PPRINT_EN); } static void __init free_ppr_log(struct amd_iommu *iommu) @@ -2526,8 +2528,6 @@ enable_faults: iommu_feature_enable(iommu, CONTROL_EVT_INT_EN); - if (iommu->ppr_log != NULL) - iommu_feature_enable(iommu, CONTROL_PPRINT_EN); return 0; } -- cgit v1.2.3 From a48130e92f1c86638295a53d1735dfed7f55a2c4 Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Wed, 28 Jun 2023 05:45:54 +0000 Subject: iommu/amd: Enable PPR/GA interrupt after interrupt handler setup Current code enables PPR and GA interrupts before setting up the interrupt handler (in state_next()). Make sure interrupt handler is in place before enabling these interrupt. amd_iommu_enable_interrupts() gets called in normal boot, kdump as well as in suspend/resume path. Hence moving interrupt enablement to this function works fine. Reviewed-by: Suravee Suthikulpanit Signed-off-by: Vasant Hegde Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230628054554.6131-4-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/init.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index f26c077eb9ea..45efb7e5d725 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -2943,8 +2943,6 @@ static void enable_iommus_vapic(void) static void enable_iommus(void) { early_enable_iommus(); - enable_iommus_vapic(); - enable_iommus_v2(); } static void disable_iommus(void) @@ -3208,6 +3206,13 @@ static int amd_iommu_enable_interrupts(void) goto out; } + /* + * Interrupt handler is ready to process interrupts. Enable + * PPR and GA log interrupt for all IOMMUs. + */ + enable_iommus_vapic(); + enable_iommus_v2(); + out: return ret; } @@ -3287,8 +3292,6 @@ static int __init state_next(void) register_syscore_ops(&amd_iommu_syscore_ops); ret = amd_iommu_init_pci(); init_state = ret ? IOMMU_INIT_ERROR : IOMMU_PCI_INIT; - enable_iommus_vapic(); - enable_iommus_v2(); break; case IOMMU_PCI_INIT: ret = amd_iommu_enable_interrupts(); -- cgit v1.2.3 From 1672730cffaf56ea3f7b697d499d9697a34b77f1 Mon Sep 17 00:00:00 2001 From: Dawei Li Date: Sun, 16 Jul 2023 00:16:21 +0800 Subject: iommu/arm-smmu-v3: Change vmid alloc strategy from bitmap to ida For current implementation of vmid allocation of arm smmu-v3, a per-smmu devide bitmap of 64K bits(8K bytes) is allocated on behalf of possible VMID range, which is two pages for some architectures. Besides that, its memory consumption is 'static', despite of how many VMIDs are allocated actually. That's memory inefficient and lack of scalability. So an IDA based implementation is introduced to address this issue, which is capable of self-expanding on the actual need of VMID allocation. Signed-off-by: Dawei Li Reviewed-by: Jason Gunthorpe Link: https://lore.kernel.org/r/TYCP286MB2323E0C525FF9F94E3B07C7ACA35A@TYCP286MB2323.JPNP286.PROD.OUTLOOK.COM Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 29 +++++++---------------------- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h | 2 +- 2 files changed, 8 insertions(+), 23 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c index 9b0dc3505601..d9487602701f 100644 --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c @@ -2055,24 +2055,6 @@ static struct iommu_domain *arm_smmu_domain_alloc(unsigned type) return &smmu_domain->domain; } -static int arm_smmu_bitmap_alloc(unsigned long *map, int span) -{ - int idx, size = 1 << span; - - do { - idx = find_first_zero_bit(map, size); - if (idx == size) - return -ENOSPC; - } while (test_and_set_bit(idx, map)); - - return idx; -} - -static void arm_smmu_bitmap_free(unsigned long *map, int idx) -{ - clear_bit(idx, map); -} - static void arm_smmu_domain_free(struct iommu_domain *domain) { struct arm_smmu_domain *smmu_domain = to_smmu_domain(domain); @@ -2093,7 +2075,7 @@ static void arm_smmu_domain_free(struct iommu_domain *domain) } else { struct arm_smmu_s2_cfg *cfg = &smmu_domain->s2_cfg; if (cfg->vmid) - arm_smmu_bitmap_free(smmu->vmid_map, cfg->vmid); + ida_free(&smmu->vmid_map, cfg->vmid); } kfree(smmu_domain); @@ -2167,7 +2149,9 @@ static int arm_smmu_domain_finalise_s2(struct arm_smmu_domain *smmu_domain, struct arm_smmu_s2_cfg *cfg = &smmu_domain->s2_cfg; typeof(&pgtbl_cfg->arm_lpae_s2_cfg.vtcr) vtcr; - vmid = arm_smmu_bitmap_alloc(smmu->vmid_map, smmu->vmid_bits); + /* Reserve VMID 0 for stage-2 bypass STEs */ + vmid = ida_alloc_range(&smmu->vmid_map, 1, (1 << smmu->vmid_bits) - 1, + GFP_KERNEL); if (vmid < 0) return vmid; @@ -3098,8 +3082,8 @@ static int arm_smmu_init_strtab(struct arm_smmu_device *smmu) reg |= STRTAB_BASE_RA; smmu->strtab_cfg.strtab_base = reg; - /* Allocate the first VMID for stage-2 bypass STEs */ - set_bit(0, smmu->vmid_map); + ida_init(&smmu->vmid_map); + return 0; } @@ -3923,6 +3907,7 @@ static void arm_smmu_device_remove(struct platform_device *pdev) iommu_device_sysfs_remove(&smmu->iommu); arm_smmu_device_disable(smmu); iopf_queue_free(smmu->evtq.iopf); + ida_destroy(&smmu->vmid_map); } static void arm_smmu_device_shutdown(struct platform_device *pdev) diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h index dcab85698a4e..9915850dd4db 100644 --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.h @@ -670,7 +670,7 @@ struct arm_smmu_device { #define ARM_SMMU_MAX_VMIDS (1 << 16) unsigned int vmid_bits; - DECLARE_BITMAP(vmid_map, ARM_SMMU_MAX_VMIDS); + struct ida vmid_map; unsigned int ssid_bits; unsigned int sid_bits; -- cgit v1.2.3 From 0a8c264d51ad3196a3239e731be51d2cc98f9337 Mon Sep 17 00:00:00 2001 From: Yangtao Li Date: Wed, 5 Jul 2023 21:04:11 +0800 Subject: iommu/arm-smmu: Clean up resource handling during Qualcomm context probe Convert to use devm_platform_ioremap_resource() and fix return value when platform_get_irq fails. Signed-off-by: Yangtao Li Link: https://lore.kernel.org/r/20230705130416.46710-1-frank.li@vivo.com Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/qcom_iommu.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index a503ed758ec3..abf798104830 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -633,7 +633,6 @@ static int qcom_iommu_ctx_probe(struct platform_device *pdev) struct qcom_iommu_ctx *ctx; struct device *dev = &pdev->dev; struct qcom_iommu_dev *qcom_iommu = dev_get_drvdata(dev->parent); - struct resource *res; int ret, irq; ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL); @@ -643,14 +642,13 @@ static int qcom_iommu_ctx_probe(struct platform_device *pdev) ctx->dev = dev; platform_set_drvdata(pdev, ctx); - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); - ctx->base = devm_ioremap_resource(dev, res); + ctx->base = devm_platform_ioremap_resource(pdev, 0); if (IS_ERR(ctx->base)) return PTR_ERR(ctx->base); irq = platform_get_irq(pdev, 0); if (irq < 0) - return -ENODEV; + return irq; /* clear IRQs before registering fault handler, just in case the * boot-loader left us a surprise: -- cgit v1.2.3 From cf69ef46dbd980a0b1c956d668e066a73e0acd0f Mon Sep 17 00:00:00 2001 From: "Chengci.Xu" Date: Fri, 2 Jun 2023 17:02:22 +0800 Subject: iommu/mediatek: Fix two IOMMU share pagetable issue Prepare for mt8188 to fix a two IOMMU HWs share pagetable issue. We have two MM IOMMU HWs in mt8188, one is VPP-IOMMU, the other is VDO-IOMMU. The 2 MM IOMMU HWs share pagetable don't work in this case: a) VPP-IOMMU probe firstly. b) VDO-IOMMU probe. c) The master for VDO-IOMMU probe (means frstdata is vpp-iommu). d) The master in another domain probe. No matter it is vdo or vpp. Then it still create a new pagetable in step d). The problem is "frstdata->bank[0]->m4u_dom" was not initialized. Then when d) enter, it still create a new one. In this patch, we create a new variable "share_dom" for this share pgtable case, it should be helpful for readable. and put all the share pgtable logic in the mtk_iommu_domain_finalise. In mt8195, the master of VPP-IOMMU probes before than VDO-IOMMU from its dtsi node sequence, we don't see this issue in it. Prepare for mt8188. Fixes: 645b87c190c9 ("iommu/mediatek: Fix 2 HW sharing pgtable issue") Signed-off-by: Chengci.Xu Signed-off-by: Yong Wu Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230602090227.7264-3-yong.wu@mediatek.com Signed-off-by: Joerg Roedel --- drivers/iommu/mtk_iommu.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index e93906d6e112..c2764891a779 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -258,6 +258,8 @@ struct mtk_iommu_data { struct device *smicomm_dev; struct mtk_iommu_bank_data *bank; + struct mtk_iommu_domain *share_dom; /* For 2 HWs share pgtable */ + struct regmap *pericfg; struct mutex mutex; /* Protect m4u_group/m4u_dom above */ @@ -620,15 +622,14 @@ static int mtk_iommu_domain_finalise(struct mtk_iommu_domain *dom, struct mtk_iommu_data *data, unsigned int region_id) { + struct mtk_iommu_domain *share_dom = data->share_dom; const struct mtk_iommu_iova_region *region; - struct mtk_iommu_domain *m4u_dom; - - /* Always use bank0 in sharing pgtable case */ - m4u_dom = data->bank[0].m4u_dom; - if (m4u_dom) { - dom->iop = m4u_dom->iop; - dom->cfg = m4u_dom->cfg; - dom->domain.pgsize_bitmap = m4u_dom->cfg.pgsize_bitmap; + + /* Always use share domain in sharing pgtable case */ + if (MTK_IOMMU_HAS_FLAG(data->plat_data, SHARE_PGTABLE) && share_dom) { + dom->iop = share_dom->iop; + dom->cfg = share_dom->cfg; + dom->domain.pgsize_bitmap = share_dom->cfg.pgsize_bitmap; goto update_iova_region; } @@ -658,6 +659,9 @@ static int mtk_iommu_domain_finalise(struct mtk_iommu_domain *dom, /* Update our support page sizes bitmap */ dom->domain.pgsize_bitmap = dom->cfg.pgsize_bitmap; + if (MTK_IOMMU_HAS_FLAG(data->plat_data, SHARE_PGTABLE)) + data->share_dom = dom; + update_iova_region: /* Update the iova region for this domain */ region = data->plat_data->iova_region + region_id; @@ -708,7 +712,9 @@ static int mtk_iommu_attach_device(struct iommu_domain *domain, /* Data is in the frstdata in sharing pgtable case. */ frstdata = mtk_iommu_get_frst_data(hw_list); + mutex_lock(&frstdata->mutex); ret = mtk_iommu_domain_finalise(dom, frstdata, region_id); + mutex_unlock(&frstdata->mutex); if (ret) { mutex_unlock(&dom->mutex); return ret; -- cgit v1.2.3 From 9a89051084af8ef566f3a2127cf5baa433dea497 Mon Sep 17 00:00:00 2001 From: "Chengci.Xu" Date: Fri, 2 Jun 2023 17:02:23 +0800 Subject: iommu/mediatek: Adjust mtk_iommu_config flow If there are many ports in a infra master, current flow will update the INFRA register many times. This patch saves all ports to portid_msk in the front of mtk_iommu_config(), then update only once for the IOMMU configure. After this, we could avoid send too many SMC calls to ATF in MT8188. Prepare for MT8188, also reduce the indention without functional change. Signed-off-by: Chengci.Xu Signed-off-by: Yong Wu Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230602090227.7264-4-yong.wu@mediatek.com Signed-off-by: Joerg Roedel --- drivers/iommu/mtk_iommu.c | 58 ++++++++++++++++++++++++++--------------------- 1 file changed, 32 insertions(+), 26 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index c2764891a779..e1146a595732 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -579,41 +579,47 @@ static int mtk_iommu_config(struct mtk_iommu_data *data, struct device *dev, unsigned int larbid, portid; struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); const struct mtk_iommu_iova_region *region; - u32 peri_mmuen, peri_mmuen_msk; + unsigned long portid_msk = 0; int i, ret = 0; for (i = 0; i < fwspec->num_ids; ++i) { - larbid = MTK_M4U_TO_LARB(fwspec->ids[i]); portid = MTK_M4U_TO_PORT(fwspec->ids[i]); + portid_msk |= BIT(portid); + } - if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_MM)) { - larb_mmu = &data->larb_imu[larbid]; + if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_MM)) { + /* All ports should be in the same larb. just use 0 here */ + larbid = MTK_M4U_TO_LARB(fwspec->ids[0]); + larb_mmu = &data->larb_imu[larbid]; + region = data->plat_data->iova_region + regionid; - region = data->plat_data->iova_region + regionid; + for_each_set_bit(portid, &portid_msk, 32) larb_mmu->bank[portid] = upper_32_bits(region->iova_base); - dev_dbg(dev, "%s iommu for larb(%s) port %d region %d rgn-bank %d.\n", - enable ? "enable" : "disable", dev_name(larb_mmu->dev), - portid, regionid, larb_mmu->bank[portid]); - - if (enable) - larb_mmu->mmu |= MTK_SMI_MMU_EN(portid); - else - larb_mmu->mmu &= ~MTK_SMI_MMU_EN(portid); - } else if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_INFRA)) { - peri_mmuen_msk = BIT(portid); - /* PCI dev has only one output id, enable the next writing bit for PCIe */ - if (dev_is_pci(dev)) - peri_mmuen_msk |= BIT(portid + 1); - - peri_mmuen = enable ? peri_mmuen_msk : 0; - ret = regmap_update_bits(data->pericfg, PERICFG_IOMMU_1, - peri_mmuen_msk, peri_mmuen); - if (ret) - dev_err(dev, "%s iommu(%s) inframaster 0x%x fail(%d).\n", - enable ? "enable" : "disable", - dev_name(data->dev), peri_mmuen_msk, ret); + dev_dbg(dev, "%s iommu for larb(%s) port 0x%lx region %d rgn-bank %d.\n", + enable ? "enable" : "disable", dev_name(larb_mmu->dev), + portid_msk, regionid, upper_32_bits(region->iova_base)); + + if (enable) + larb_mmu->mmu |= portid_msk; + else + larb_mmu->mmu &= ~portid_msk; + } else if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_INFRA)) { + /* PCI dev has only one output id, enable the next writing bit for PCIe */ + if (dev_is_pci(dev)) { + if (fwspec->num_ids != 1) { + dev_err(dev, "PCI dev can only have one port.\n"); + return -ENODEV; + } + portid_msk |= BIT(portid + 1); } + + ret = regmap_update_bits(data->pericfg, PERICFG_IOMMU_1, + (u32)portid_msk, enable ? (u32)portid_msk : 0); + if (ret) + dev_err(dev, "%s iommu(%s) inframaster 0x%lx fail(%d).\n", + enable ? "enable" : "disable", + dev_name(data->dev), portid_msk, ret); } return ret; } -- cgit v1.2.3 From 946e719ce621fc409001bc05c942d08a37b9052e Mon Sep 17 00:00:00 2001 From: "Chengci.Xu" Date: Fri, 2 Jun 2023 17:02:24 +0800 Subject: iommu/mediatek: Add enable IOMMU SMC command for INFRA masters Prepare for MT8188. In MT8188, the register which enables IOMMU for INFRA masters are in the secure world for security concerns, therefore we add a SMC command for INFRA masters to enable IOMMU in ATF. Signed-off-by: Chengci.Xu Signed-off-by: Yong Wu Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230602090227.7264-5-yong.wu@mediatek.com Signed-off-by: Joerg Roedel --- drivers/iommu/mtk_iommu.c | 32 ++++++++++++++++++++++---------- include/soc/mediatek/smi.h | 1 + 2 files changed, 23 insertions(+), 10 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index e1146a595732..fe3839bb14f7 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -3,6 +3,7 @@ * Copyright (c) 2015-2016 MediaTek Inc. * Author: Yong Wu */ +#include #include #include #include @@ -27,6 +28,7 @@ #include #include #include +#include #include #include @@ -143,6 +145,7 @@ #define PGTABLE_PA_35_EN BIT(17) #define TF_PORT_TO_ADDR_MT8173 BIT(18) #define INT_ID_PORT_WIDTH_6 BIT(19) +#define CFG_IFA_MASTER_IN_ATF BIT(20) #define MTK_IOMMU_HAS_FLAG_MASK(pdata, _x, mask) \ ((((pdata)->flags) & (mask)) == (_x)) @@ -580,6 +583,7 @@ static int mtk_iommu_config(struct mtk_iommu_data *data, struct device *dev, struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); const struct mtk_iommu_iova_region *region; unsigned long portid_msk = 0; + struct arm_smccc_res res; int i, ret = 0; for (i = 0; i < fwspec->num_ids; ++i) { @@ -605,17 +609,24 @@ static int mtk_iommu_config(struct mtk_iommu_data *data, struct device *dev, else larb_mmu->mmu &= ~portid_msk; } else if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_INFRA)) { - /* PCI dev has only one output id, enable the next writing bit for PCIe */ - if (dev_is_pci(dev)) { - if (fwspec->num_ids != 1) { - dev_err(dev, "PCI dev can only have one port.\n"); - return -ENODEV; + if (MTK_IOMMU_HAS_FLAG(data->plat_data, CFG_IFA_MASTER_IN_ATF)) { + arm_smccc_smc(MTK_SIP_KERNEL_IOMMU_CONTROL, + IOMMU_ATF_CMD_CONFIG_INFRA_IOMMU, + portid_msk, enable, 0, 0, 0, 0, &res); + ret = res.a0; + } else { + /* PCI dev has only one output id, enable the next writing bit for PCIe */ + if (dev_is_pci(dev)) { + if (fwspec->num_ids != 1) { + dev_err(dev, "PCI dev can only have one port.\n"); + return -ENODEV; + } + portid_msk |= BIT(portid + 1); } - portid_msk |= BIT(portid + 1); - } - ret = regmap_update_bits(data->pericfg, PERICFG_IOMMU_1, - (u32)portid_msk, enable ? (u32)portid_msk : 0); + ret = regmap_update_bits(data->pericfg, PERICFG_IOMMU_1, + (u32)portid_msk, enable ? (u32)portid_msk : 0); + } if (ret) dev_err(dev, "%s iommu(%s) inframaster 0x%lx fail(%d).\n", enable ? "enable" : "disable", @@ -1330,7 +1341,8 @@ static int mtk_iommu_probe(struct platform_device *pdev) dev_err_probe(dev, ret, "mm dts parse fail\n"); goto out_runtime_disable; } - } else if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_INFRA)) { + } else if (MTK_IOMMU_IS_TYPE(data->plat_data, MTK_IOMMU_TYPE_INFRA) && + !MTK_IOMMU_HAS_FLAG(data->plat_data, CFG_IFA_MASTER_IN_ATF)) { p = data->plat_data->pericfg_comp_str; data->pericfg = syscon_regmap_lookup_by_compatible(p); if (IS_ERR(data->pericfg)) { diff --git a/include/soc/mediatek/smi.h b/include/soc/mediatek/smi.h index dfd8efca5e60..000eb1cf68b7 100644 --- a/include/soc/mediatek/smi.h +++ b/include/soc/mediatek/smi.h @@ -13,6 +13,7 @@ enum iommu_atf_cmd { IOMMU_ATF_CMD_CONFIG_SMI_LARB, /* For mm master to en/disable iommu */ + IOMMU_ATF_CMD_CONFIG_INFRA_IOMMU, /* For infra master to enable iommu */ IOMMU_ATF_CMD_MAX, }; -- cgit v1.2.3 From a09e84034d87d3aa2f9ad584bebd92b0bab6e0be Mon Sep 17 00:00:00 2001 From: "Chengci.Xu" Date: Fri, 2 Jun 2023 17:02:25 +0800 Subject: iommu/mediatek: Add MT8188 IOMMU Support MT8188 has 3 IOMMU, containing 2 MM IOMMUs, one is for vdo, the other is for vpp. and 1 INFRA IOMMU. Signed-off-by: Chengci.Xu Signed-off-by: Yong Wu Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230602090227.7264-6-yong.wu@mediatek.com Signed-off-by: Joerg Roedel --- drivers/iommu/mtk_iommu.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 49 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index fe3839bb14f7..33cfb226d949 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -170,6 +170,7 @@ enum mtk_iommu_plat { M4U_MT8173, M4U_MT8183, M4U_MT8186, + M4U_MT8188, M4U_MT8192, M4U_MT8195, M4U_MT8365, @@ -1594,6 +1595,51 @@ static const struct mtk_iommu_plat_data mt8186_data_mm = { .iova_region_larb_msk = mt8186_larb_region_msk, }; +static const struct mtk_iommu_plat_data mt8188_data_infra = { + .m4u_plat = M4U_MT8188, + .flags = WR_THROT_EN | DCM_DISABLE | STD_AXI_MODE | PM_CLK_AO | + MTK_IOMMU_TYPE_INFRA | IFA_IOMMU_PCIE_SUPPORT | + PGTABLE_PA_35_EN | CFG_IFA_MASTER_IN_ATF, + .inv_sel_reg = REG_MMU_INV_SEL_GEN2, + .banks_num = 1, + .banks_enable = {true}, + .iova_region = single_domain, + .iova_region_nr = ARRAY_SIZE(single_domain), +}; + +static const struct mtk_iommu_plat_data mt8188_data_vdo = { + .m4u_plat = M4U_MT8188, + .flags = HAS_BCLK | HAS_SUB_COMM_3BITS | OUT_ORDER_WR_EN | + WR_THROT_EN | IOVA_34_EN | SHARE_PGTABLE | + PGTABLE_PA_35_EN | MTK_IOMMU_TYPE_MM, + .hw_list = &m4ulist, + .inv_sel_reg = REG_MMU_INV_SEL_GEN2, + .banks_num = 1, + .banks_enable = {true}, + .iova_region = mt8192_multi_dom, + .iova_region_nr = ARRAY_SIZE(mt8192_multi_dom), + .larbid_remap = {{2}, {0}, {21}, {0}, {19}, {9, 10, + 11 /* 11a */, 25 /* 11c */}, + {13, 0, 29 /* 16b */, 30 /* 17b */, 0}, {5}}, +}; + +static const struct mtk_iommu_plat_data mt8188_data_vpp = { + .m4u_plat = M4U_MT8188, + .flags = HAS_BCLK | HAS_SUB_COMM_3BITS | OUT_ORDER_WR_EN | + WR_THROT_EN | IOVA_34_EN | SHARE_PGTABLE | + PGTABLE_PA_35_EN | MTK_IOMMU_TYPE_MM, + .hw_list = &m4ulist, + .inv_sel_reg = REG_MMU_INV_SEL_GEN2, + .banks_num = 1, + .banks_enable = {true}, + .iova_region = mt8192_multi_dom, + .iova_region_nr = ARRAY_SIZE(mt8192_multi_dom), + .larbid_remap = {{1}, {3}, {23}, {7}, {MTK_INVALID_LARBID}, + {12, 15, 24 /* 11b */}, {14, MTK_INVALID_LARBID, + 16 /* 16a */, 17 /* 17a */, MTK_INVALID_LARBID, + 27, 28 /* ccu0 */, MTK_INVALID_LARBID}, {4, 6}}, +}; + static const unsigned int mt8192_larb_region_msk[MT8192_MULTI_REGION_NR_MAX][MTK_LARB_NR_MAX] = { [0] = {~0, ~0}, /* Region0: larb0/1 */ [1] = {0, 0, 0, 0, ~0, ~0, 0, ~0}, /* Region1: larb4/5/7 */ @@ -1702,6 +1748,9 @@ static const struct of_device_id mtk_iommu_of_ids[] = { { .compatible = "mediatek,mt8173-m4u", .data = &mt8173_data}, { .compatible = "mediatek,mt8183-m4u", .data = &mt8183_data}, { .compatible = "mediatek,mt8186-iommu-mm", .data = &mt8186_data_mm}, /* mm: m4u */ + { .compatible = "mediatek,mt8188-iommu-infra", .data = &mt8188_data_infra}, + { .compatible = "mediatek,mt8188-iommu-vdo", .data = &mt8188_data_vdo}, + { .compatible = "mediatek,mt8188-iommu-vpp", .data = &mt8188_data_vpp}, { .compatible = "mediatek,mt8192-m4u", .data = &mt8192_data}, { .compatible = "mediatek,mt8195-iommu-infra", .data = &mt8195_data_infra}, { .compatible = "mediatek,mt8195-iommu-vdo", .data = &mt8195_data_vdo}, -- cgit v1.2.3 From 1e8a46393a9579d2fe390d330edcd85623d2119d Mon Sep 17 00:00:00 2001 From: Yong Wu Date: Fri, 2 Jun 2023 17:02:26 +0800 Subject: iommu/mediatek: mt8188: Add iova_region_larb_msk Add iova_region_larb_msk for mt8188. We separate the 16GB iova regions by each device's larbid/portid. Refer to include/dt-bindings/memory/mediatek,mt8188-memory-port.h As commented in the code, larb19(21) means it's larb19 while its SW index is 21. Signed-off-by: Yong Wu Reviewed-by: AngeloGioacchino Del Regno Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230602090227.7264-7-yong.wu@mediatek.com Signed-off-by: Joerg Roedel --- drivers/iommu/mtk_iommu.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index 33cfb226d949..640275873a27 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -1607,6 +1607,20 @@ static const struct mtk_iommu_plat_data mt8188_data_infra = { .iova_region_nr = ARRAY_SIZE(single_domain), }; +static const u32 mt8188_larb_region_msk[MT8192_MULTI_REGION_NR_MAX][MTK_LARB_NR_MAX] = { + [0] = {~0, ~0, ~0, ~0}, /* Region0: all ports for larb0/1/2/3 */ + [1] = {0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, ~0, ~0, ~0}, /* Region1: larb19(21)/21(22)/23 */ + [2] = {0, 0, 0, 0, ~0, ~0, ~0, ~0, /* Region2: the other larbs. */ + ~0, ~0, ~0, ~0, ~0, ~0, ~0, ~0, + ~0, ~0, ~0, ~0, ~0, 0, 0, 0, + 0, ~0}, + [3] = {0}, + [4] = {[24] = BIT(0) | BIT(1)}, /* Only larb27(24) port0/1 */ + [5] = {[24] = BIT(2) | BIT(3)}, /* Only larb27(24) port2/3 */ +}; + static const struct mtk_iommu_plat_data mt8188_data_vdo = { .m4u_plat = M4U_MT8188, .flags = HAS_BCLK | HAS_SUB_COMM_3BITS | OUT_ORDER_WR_EN | @@ -1618,6 +1632,7 @@ static const struct mtk_iommu_plat_data mt8188_data_vdo = { .banks_enable = {true}, .iova_region = mt8192_multi_dom, .iova_region_nr = ARRAY_SIZE(mt8192_multi_dom), + .iova_region_larb_msk = mt8188_larb_region_msk, .larbid_remap = {{2}, {0}, {21}, {0}, {19}, {9, 10, 11 /* 11a */, 25 /* 11c */}, {13, 0, 29 /* 16b */, 30 /* 17b */, 0}, {5}}, @@ -1634,6 +1649,7 @@ static const struct mtk_iommu_plat_data mt8188_data_vpp = { .banks_enable = {true}, .iova_region = mt8192_multi_dom, .iova_region_nr = ARRAY_SIZE(mt8192_multi_dom), + .iova_region_larb_msk = mt8188_larb_region_msk, .larbid_remap = {{1}, {3}, {23}, {7}, {MTK_INVALID_LARBID}, {12, 15, 24 /* 11b */}, {14, MTK_INVALID_LARBID, 16 /* 16a */, 17 /* 17a */, MTK_INVALID_LARBID, -- cgit v1.2.3 From aedd11e01db069123f3927ddce79f5376251c949 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Mon, 17 Jul 2023 15:14:40 +0200 Subject: iommu/ipmmu-vmsa: Convert to read_poll_timeout_atomic() Use read_poll_timeout_atomic() instead of open-coding the same operation. Signed-off-by: Geert Uytterhoeven Link: https://lore.kernel.org/r/649c7e09841b998c5c8d7fc274884a85e4b5bfe9.1689599528.git.geert+renesas@glider.be Signed-off-by: Joerg Roedel --- drivers/iommu/ipmmu-vmsa.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c index 9f64c5c9f5b9..3b58a8ea3bde 100644 --- a/drivers/iommu/ipmmu-vmsa.c +++ b/drivers/iommu/ipmmu-vmsa.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -253,17 +254,13 @@ static void ipmmu_imuctr_write(struct ipmmu_vmsa_device *mmu, /* Wait for any pending TLB invalidations to complete */ static void ipmmu_tlb_sync(struct ipmmu_vmsa_domain *domain) { - unsigned int count = 0; + u32 val; - while (ipmmu_ctx_read_root(domain, IMCTR) & IMCTR_FLUSH) { - cpu_relax(); - if (++count == TLB_LOOP_TIMEOUT) { - dev_err_ratelimited(domain->mmu->dev, + if (read_poll_timeout_atomic(ipmmu_ctx_read_root, val, + !(val & IMCTR_FLUSH), 1, TLB_LOOP_TIMEOUT, + false, domain, IMCTR)) + dev_err_ratelimited(domain->mmu->dev, "TLB sync timed out -- MMU may be deadlocked\n"); - return; - } - udelay(1); - } } static void ipmmu_tlb_invalidate(struct ipmmu_vmsa_domain *domain) -- cgit v1.2.3 From ed8c975b0f64ffc1e546fca26519974d8a446831 Mon Sep 17 00:00:00 2001 From: Min-Hua Chen Date: Fri, 21 Jul 2023 07:21:54 +0800 Subject: iommu/apple-dart: mark apple_dart_pm_ops static This patch fixes the following sparse warning: drivers/iommu/apple-dart.c:1279:1: sparse: warning: symbol 'apple_dart_pm_ops' was not declared. Should it be static? No functional change intended. Signed-off-by: Min-Hua Chen Acked-by: Sven Peter Link: https://lore.kernel.org/r/20230720232155.3923-1-minhuadotchen@gmail.com Signed-off-by: Joerg Roedel --- drivers/iommu/apple-dart.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/iommu/apple-dart.c b/drivers/iommu/apple-dart.c index 8af64b57f048..2082081402d3 100644 --- a/drivers/iommu/apple-dart.c +++ b/drivers/iommu/apple-dart.c @@ -1276,7 +1276,7 @@ static __maybe_unused int apple_dart_resume(struct device *dev) return 0; } -DEFINE_SIMPLE_DEV_PM_OPS(apple_dart_pm_ops, apple_dart_suspend, apple_dart_resume); +static DEFINE_SIMPLE_DEV_PM_OPS(apple_dart_pm_ops, apple_dart_suspend, apple_dart_resume); static const struct of_device_id apple_dart_of_match[] = { { .compatible = "apple,t8103-dart", .data = &apple_dart_hw_t8103 }, -- cgit v1.2.3 From d48a51286c698f7fe8efc688f23a532f4fe9a904 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Mon, 24 Jul 2023 14:36:05 -0300 Subject: iommu/sprd: Add missing force_aperture force_aperture was intended to false only by GART drivers that have an identity translation outside the aperture. This does not describe sprd, so add the missing 'force_aperture = true'. Fixes: b23e4fc4e3fa ("iommu: add Unisoc IOMMU basic driver") Signed-off-by: Jason Gunthorpe Acked-by: Chunyan Zhang Signed-off-by: Joerg Roedel --- drivers/iommu/sprd-iommu.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/iommu/sprd-iommu.c b/drivers/iommu/sprd-iommu.c index 39e34fdeccda..eb684d8807ca 100644 --- a/drivers/iommu/sprd-iommu.c +++ b/drivers/iommu/sprd-iommu.c @@ -148,6 +148,7 @@ static struct iommu_domain *sprd_iommu_domain_alloc(unsigned int domain_type) dom->domain.geometry.aperture_start = 0; dom->domain.geometry.aperture_end = SZ_256M - 1; + dom->domain.geometry.force_aperture = true; return &dom->domain; } -- cgit v1.2.3 From 52a8fd24d129fb0e82fe6476dea35480d771365a Mon Sep 17 00:00:00 2001 From: Yoshihiro Shimoda Date: Fri, 28 Jul 2023 10:46:59 +0900 Subject: iommu/ipmmu-vmsa: Allow PCIe devices IPMMU hardware on R-Car Gen3 and RZ/G2 is simple. Each bus-master device like eMMC host and PCIe controllers has a micro-TLB of The IPMMU, and after enabled it, all transactions of the device are under the IPMMU. eMMC host ---(micro-TLB of eMMC)--- IPMMU cache --- IPMMU main PCIe --------(micro-TLB of PCIe)--- IPMMU cache --- IPMMU main Now this IPMMU driver allows eMMC host, and it is safe to use the IPMMU. So, we can assume that it is safe to use the IPMMU from PCIe devices too, because all PCIe devices transactions will go to the micro-TLB of PCIe. So, add a new condition whether the device is a PCIe device or not in the ipmmu_device_is_allowed() which will be called if the PCIe host controller has iommu-map property. This can improve CPU load because the PCIe controllers only have a capability for lower 32-bit memory area so that this can avoid using swiotlb. Note that IPMMU on R-Car Gen4 is different than R-Car Gen3 and RZ/G2's one, especially OS-ID. But, for now, the IPMMU driver takes care of OS-ID 0 only. In other words, all PCIe devices will go to the micro-TLB of PCIe. Signed-off-by: Yoshihiro Shimoda Reviewed-by: Geert Uytterhoeven Link: https://lore.kernel.org/r/20230728014659.411751-1-yoshihiro.shimoda.uh@renesas.com Signed-off-by: Joerg Roedel --- drivers/iommu/ipmmu-vmsa.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c index 3b58a8ea3bde..c31552d6a5fa 100644 --- a/drivers/iommu/ipmmu-vmsa.c +++ b/drivers/iommu/ipmmu-vmsa.c @@ -20,6 +20,7 @@ #include #include #include +#include #include #include #include @@ -720,6 +721,10 @@ static bool ipmmu_device_is_allowed(struct device *dev) if (soc_device_match(soc_denylist)) return false; + /* Check whether this device is a PCI device */ + if (dev_is_pci(dev)) + return true; + /* Check whether this device can work with the IPMMU */ for (i = 0; i < ARRAY_SIZE(devices_allowlist); i++) { if (!strcmp(dev_name(dev), devices_allowlist[i])) -- cgit v1.2.3 From 6b7867b5b8a6b14c487bf04a693ab424c7a8718d Mon Sep 17 00:00:00 2001 From: Zhu Wang Date: Mon, 31 Jul 2023 19:27:58 +0800 Subject: iommu: Remove kernel-doc warnings Remove kernel-doc warnings: drivers/iommu/iommu.c:3261: warning: Function parameter or member 'group' not described in 'iommu_group_release_dma_owner' drivers/iommu/iommu.c:3261: warning: Excess function parameter 'dev' description in 'iommu_group_release_dma_owner' drivers/iommu/iommu.c:3275: warning: Function parameter or member 'dev' not described in 'iommu_device_release_dma_owner' drivers/iommu/iommu.c:3275: warning: Excess function parameter 'group' description in 'iommu_device_release_dma_owner' Signed-off-by: Zhu Wang Fixes: 89395ccedbc1 ("iommu: Add device-centric DMA ownership interfaces") Reviewed-by: Jason Gunthorpe Link: https://lore.kernel.org/r/20230731112758.214775-1-wangzhu9@huawei.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index e67f6562da73..c53a4942f745 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -3216,7 +3216,7 @@ 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(). */ @@ -3230,7 +3230,7 @@ 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(). */ -- cgit v1.2.3 From 8e11876a1127f27c5e5e07072bfc590e53509a84 Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Mon, 19 Jun 2023 13:19:08 +0000 Subject: iommu/amd: Rearrange DTE bit definations Rearrage according to 64bit word they are in. Note that I have not rearranged gcr3 related macros even though they belong to different 64bit word as its easy to read it in current format. No functional changes intended. Suggested-by: Jerry Snitselaar Signed-off-by: Vasant Hegde Reviewed-by: Jerry Snitselaar Link: https://lore.kernel.org/r/20230619131908.5887-1-vasant.hegde@amd.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu_types.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index 781ab9c2ea70..7dc30c2b56b3 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -384,15 +384,15 @@ */ #define DTE_FLAG_V BIT_ULL(0) #define DTE_FLAG_TV BIT_ULL(1) +#define DTE_FLAG_GIOV BIT_ULL(54) +#define DTE_FLAG_GV BIT_ULL(55) +#define DTE_GLX_SHIFT (56) +#define DTE_GLX_MASK (3) #define DTE_FLAG_IR BIT_ULL(61) #define DTE_FLAG_IW BIT_ULL(62) #define DTE_FLAG_IOTLB BIT_ULL(32) -#define DTE_FLAG_GIOV BIT_ULL(54) -#define DTE_FLAG_GV BIT_ULL(55) #define DTE_FLAG_MASK (0x3ffULL << 32) -#define DTE_GLX_SHIFT (56) -#define DTE_GLX_MASK (3) #define DEV_DOMID_MASK 0xffffULL #define DTE_GCR3_VAL_A(x) (((x) >> 12) & 0x00007ULL) -- cgit v1.2.3 From fcf226f1f7083cba76af47bf8dd764b68b149cd2 Mon Sep 17 00:00:00 2001 From: AngeloGioacchino Del Regno Date: Thu, 22 Jun 2023 11:27:38 +0200 Subject: iommu/qcom: Use the asid read from device-tree if specified As specified in this driver, the context banks are 0x1000 apart but on some SoCs the context number does not necessarily match this logic, hence we end up using the wrong ASID: keeping in mind that this IOMMU implementation relies heavily on SCM (TZ) calls, it is mandatory that we communicate the right context number. Since this is all about how context banks are mapped in firmware, which may be board dependent (as a different firmware version may eventually change the expected context bank numbers), introduce a new property "qcom,ctx-asid": when found, the ASID will be forced as read from the devicetree. When "qcom,ctx-asid" is not found, this driver retains the previous behavior as to avoid breaking older devicetrees or systems that do not require forcing ASID numbers. Signed-off-by: Marijn Suijten [Marijn: Rebased over next-20221111] Signed-off-by: AngeloGioacchino Del Regno Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230622092742.74819-3-angelogioacchino.delregno@collabora.com Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/qcom_iommu.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index abf798104830..8c717bc2d9fb 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -531,7 +531,8 @@ static int qcom_iommu_of_xlate(struct device *dev, struct of_phandle_args *args) * index into qcom_iommu->ctxs: */ if (WARN_ON(asid < 1) || - WARN_ON(asid > qcom_iommu->num_ctxs)) { + WARN_ON(asid > qcom_iommu->num_ctxs) || + WARN_ON(qcom_iommu->ctxs[asid - 1] == NULL)) { put_device(&iommu_pdev->dev); return -EINVAL; } @@ -617,7 +618,8 @@ free_mem: static int get_asid(const struct device_node *np) { - u32 reg; + u32 reg, val; + int asid; /* read the "reg" property directly to get the relative address * of the context bank, and calculate the asid from that: @@ -625,7 +627,17 @@ static int get_asid(const struct device_node *np) if (of_property_read_u32_index(np, "reg", 0, ®)) return -ENODEV; - return reg / 0x1000; /* context banks are 0x1000 apart */ + /* + * Context banks are 0x1000 apart but, in some cases, the ASID + * number doesn't match to this logic and needs to be passed + * from the DT configuration explicitly. + */ + if (!of_property_read_u32(np, "qcom,ctx-asid", &val)) + asid = val; + else + asid = reg / 0x1000; + + return asid; } static int qcom_iommu_ctx_probe(struct platform_device *pdev) -- cgit v1.2.3 From 9f3fef23d9b5a858a6e6d5f478bb1b6b76265e76 Mon Sep 17 00:00:00 2001 From: AngeloGioacchino Del Regno Date: Thu, 22 Jun 2023 11:27:39 +0200 Subject: iommu/qcom: Disable and reset context bank before programming Writing the new TTBRs, TCRs and MAIRs on a previously enabled context bank may trigger a context fault, resulting in firmware driven AP resets: change the domain initialization programming sequence to disable the context bank(s) and to also clear the related fault address (CB_FAR) and fault status (CB_FSR) registers before writing new values to TTBR0/1, TCR/TCR2, MAIR0/1. Fixes: 0ae349a0f33f ("iommu/qcom: Add qcom_iommu") Signed-off-by: AngeloGioacchino Del Regno Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230622092742.74819-4-angelogioacchino.delregno@collabora.com Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/qcom_iommu.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index 8c717bc2d9fb..a352859edea8 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -273,6 +273,13 @@ static int qcom_iommu_init_domain(struct iommu_domain *domain, ctx->secure_init = true; } + /* Disable context bank before programming */ + iommu_writel(ctx, ARM_SMMU_CB_SCTLR, 0); + + /* Clear context bank fault address fault status registers */ + iommu_writel(ctx, ARM_SMMU_CB_FAR, 0); + iommu_writel(ctx, ARM_SMMU_CB_FSR, ARM_SMMU_FSR_FAULT); + /* TTBRs */ iommu_writeq(ctx, ARM_SMMU_CB_TTBR0, pgtbl_cfg.arm_lpae_s1_cfg.ttbr | -- cgit v1.2.3 From ec5601661bfcdc206e6ceba1b97837e763dab1ba Mon Sep 17 00:00:00 2001 From: AngeloGioacchino Del Regno Date: Thu, 22 Jun 2023 11:27:40 +0200 Subject: iommu/qcom: Index contexts by asid number to allow asid 0 This driver was indexing the contexts by asid-1, which is probably done under the assumption that the first ASID is always 1. Unfortunately this is not always true: at least for MSM8956 and MSM8976's GPU IOMMU, the gpu_user context's ASID number is zero. To allow using a zero asid number, index the contexts by `asid` instead of by `asid - 1`. While at it, also enhance human readability by renaming the `num_ctxs` member of struct qcom_iommu_dev to `max_asid`. Signed-off-by: AngeloGioacchino Del Regno Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230622092742.74819-5-angelogioacchino.delregno@collabora.com Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/qcom_iommu.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index a352859edea8..60e4959cc9fd 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -51,8 +51,8 @@ struct qcom_iommu_dev { struct clk_bulk_data clks[CLK_NUM]; void __iomem *local_base; u32 sec_id; - u8 num_ctxs; - struct qcom_iommu_ctx *ctxs[]; /* indexed by asid-1 */ + u8 max_asid; + struct qcom_iommu_ctx *ctxs[]; /* indexed by asid */ }; struct qcom_iommu_ctx { @@ -94,7 +94,7 @@ static struct qcom_iommu_ctx * to_ctx(struct qcom_iommu_domain *d, unsigned asid struct qcom_iommu_dev *qcom_iommu = d->iommu; if (!qcom_iommu) return NULL; - return qcom_iommu->ctxs[asid - 1]; + return qcom_iommu->ctxs[asid]; } static inline void @@ -534,12 +534,10 @@ static int qcom_iommu_of_xlate(struct device *dev, struct of_phandle_args *args) qcom_iommu = platform_get_drvdata(iommu_pdev); /* make sure the asid specified in dt is valid, so we don't have - * to sanity check this elsewhere, since 'asid - 1' is used to - * index into qcom_iommu->ctxs: + * to sanity check this elsewhere: */ - if (WARN_ON(asid < 1) || - WARN_ON(asid > qcom_iommu->num_ctxs) || - WARN_ON(qcom_iommu->ctxs[asid - 1] == NULL)) { + if (WARN_ON(asid > qcom_iommu->max_asid) || + WARN_ON(qcom_iommu->ctxs[asid] == NULL)) { put_device(&iommu_pdev->dev); return -EINVAL; } @@ -694,7 +692,7 @@ static int qcom_iommu_ctx_probe(struct platform_device *pdev) dev_dbg(dev, "found asid %u\n", ctx->asid); - qcom_iommu->ctxs[ctx->asid - 1] = ctx; + qcom_iommu->ctxs[ctx->asid] = ctx; return 0; } @@ -706,7 +704,7 @@ static void qcom_iommu_ctx_remove(struct platform_device *pdev) platform_set_drvdata(pdev, NULL); - qcom_iommu->ctxs[ctx->asid - 1] = NULL; + qcom_iommu->ctxs[ctx->asid] = NULL; } static const struct of_device_id ctx_of_match[] = { @@ -753,11 +751,11 @@ static int qcom_iommu_device_probe(struct platform_device *pdev) for_each_child_of_node(dev->of_node, child) max_asid = max(max_asid, get_asid(child)); - qcom_iommu = devm_kzalloc(dev, struct_size(qcom_iommu, ctxs, max_asid), + qcom_iommu = devm_kzalloc(dev, struct_size(qcom_iommu, ctxs, max_asid + 1), GFP_KERNEL); if (!qcom_iommu) return -ENOMEM; - qcom_iommu->num_ctxs = max_asid; + qcom_iommu->max_asid = max_asid; qcom_iommu->dev = dev; res = platform_get_resource(pdev, IORESOURCE_MEM, 0); -- cgit v1.2.3 From e30c960d3f44b7aaa0861ce4afa8ba8b8f4b0f03 Mon Sep 17 00:00:00 2001 From: AngeloGioacchino Del Regno Date: Thu, 22 Jun 2023 11:27:42 +0200 Subject: iommu/qcom: Add support for QSMMUv2 and QSMMU-500 secured contexts On some SoCs like MSM8956, MSM8976 and others, secure contexts are also secured: these get programmed by the bootloader or TZ (as usual) but their "interesting" registers are locked out by the hypervisor, disallowing direct register writes from Linux and, in many cases, completely disallowing the reprogramming of TTBR, TCR, MAIR and other registers including, but not limited to, resetting contexts. This is referred downstream as a "v2" IOMMU but this is effectively a "v2 firmware configuration" instead. Luckily, the described behavior of version 2 is effective only on secure contexts and not on non-secure ones: add support for that, finally getting a completely working IOMMU on at least MSM8956/76. Signed-off-by: Marijn Suijten [Marijn: Rebased over next-20221111] Signed-off-by: AngeloGioacchino Del Regno Reviewed-by: Dmitry Baryshkov Reviewed-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230622092742.74819-7-angelogioacchino.delregno@collabora.com Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/qcom_iommu.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index 60e4959cc9fd..3f788912b639 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -59,6 +59,7 @@ struct qcom_iommu_ctx { struct device *dev; void __iomem *base; bool secure_init; + bool secured_ctx; u8 asid; /* asid and ctx bank # are 1:1 */ struct iommu_domain *domain; }; @@ -273,6 +274,12 @@ static int qcom_iommu_init_domain(struct iommu_domain *domain, ctx->secure_init = true; } + /* Secured QSMMU-500/QSMMU-v2 contexts cannot be programmed */ + if (ctx->secured_ctx) { + ctx->domain = domain; + continue; + } + /* Disable context bank before programming */ iommu_writel(ctx, ARM_SMMU_CB_SCTLR, 0); @@ -667,10 +674,14 @@ static int qcom_iommu_ctx_probe(struct platform_device *pdev) if (irq < 0) return irq; + if (of_device_is_compatible(dev->of_node, "qcom,msm-iommu-v2-sec")) + ctx->secured_ctx = true; + /* clear IRQs before registering fault handler, just in case the * boot-loader left us a surprise: */ - iommu_writel(ctx, ARM_SMMU_CB_FSR, iommu_readl(ctx, ARM_SMMU_CB_FSR)); + if (!ctx->secured_ctx) + iommu_writel(ctx, ARM_SMMU_CB_FSR, iommu_readl(ctx, ARM_SMMU_CB_FSR)); ret = devm_request_irq(dev, irq, qcom_iommu_fault, @@ -710,6 +721,8 @@ static void qcom_iommu_ctx_remove(struct platform_device *pdev) static const struct of_device_id ctx_of_match[] = { { .compatible = "qcom,msm-iommu-v1-ns" }, { .compatible = "qcom,msm-iommu-v1-sec" }, + { .compatible = "qcom,msm-iommu-v2-ns" }, + { .compatible = "qcom,msm-iommu-v2-sec" }, { /* sentinel */ } }; @@ -727,7 +740,8 @@ static bool qcom_iommu_has_secure_context(struct qcom_iommu_dev *qcom_iommu) struct device_node *child; for_each_child_of_node(qcom_iommu->dev->of_node, child) { - if (of_device_is_compatible(child, "qcom,msm-iommu-v1-sec")) { + if (of_device_is_compatible(child, "qcom,msm-iommu-v1-sec") || + of_device_is_compatible(child, "qcom,msm-iommu-v2-sec")) { of_node_put(child); return true; } @@ -871,6 +885,7 @@ static const struct dev_pm_ops qcom_iommu_pm_ops = { static const struct of_device_id qcom_iommu_of_match[] = { { .compatible = "qcom,msm-iommu-v1" }, + { .compatible = "qcom,msm-iommu-v2" }, { /* sentinel */ } }; -- cgit v1.2.3 From 4298780126c298f20ae4bc8676591eaf8c48767e Mon Sep 17 00:00:00 2001 From: Jacob Pan Date: Wed, 9 Aug 2023 20:47:54 +0800 Subject: iommu: Generalize PASID 0 for normal DMA w/o PASID PCIe Process address space ID (PASID) is used to tag DMA traffic, it provides finer grained isolation than requester ID (RID). For each device/RID, 0 is a special PASID for the normal DMA (no PASID). This is universal across all architectures that supports PASID, therefore warranted to be reserved globally and declared in the common header. Consequently, we can avoid the conflict between different PASID use cases in the generic code. e.g. SVA and DMA API with PASIDs. This paved away for device drivers to choose global PASID policy while continue doing normal DMA. Noting that VT-d could support none-zero RID/NO_PASID, but currently not used. Reviewed-by: Lu Baolu Reviewed-by: Kevin Tian Reviewed-by: Jean-Philippe Brucker Reviewed-by: Jason Gunthorpe Signed-off-by: Jacob Pan Link: https://lore.kernel.org/r/20230802212427.1497170-2-jacob.jun.pan@linux.intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-sva.c | 2 +- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 16 ++++++++-------- drivers/iommu/intel/iommu.c | 24 ++++++++++++------------ drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/intel/pasid.h | 2 -- include/linux/iommu.h | 1 + 6 files changed, 23 insertions(+), 24 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-sva.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-sva.c index a5a63b1c947e..5e6b39881c04 100644 --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-sva.c +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3-sva.c @@ -80,7 +80,7 @@ arm_smmu_share_asid(struct mm_struct *mm, u16 asid) * be some overlap between use of both ASIDs, until we invalidate the * TLB. */ - arm_smmu_write_ctx_desc(smmu_domain, 0, cd); + arm_smmu_write_ctx_desc(smmu_domain, IOMMU_NO_PASID, cd); /* Invalidate TLB entries previously associated with that context */ arm_smmu_tlb_inv_asid(smmu, asid); diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c index 9b0dc3505601..ee70687f060b 100644 --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c @@ -1059,7 +1059,7 @@ int arm_smmu_write_ctx_desc(struct arm_smmu_domain *smmu_domain, int ssid, /* * This function handles the following cases: * - * (1) Install primary CD, for normal DMA traffic (SSID = 0). + * (1) Install primary CD, for normal DMA traffic (SSID = IOMMU_NO_PASID = 0). * (2) Install a secondary CD, for SID+SSID traffic. * (3) Update ASID of a CD. Atomically write the first 64 bits of the * CD, then invalidate the old entry and mappings. @@ -1607,7 +1607,7 @@ static void arm_smmu_handle_ppr(struct arm_smmu_device *smmu, u64 *evt) sid = FIELD_GET(PRIQ_0_SID, evt[0]); ssv = FIELD_GET(PRIQ_0_SSID_V, evt[0]); - ssid = ssv ? FIELD_GET(PRIQ_0_SSID, evt[0]) : 0; + ssid = ssv ? FIELD_GET(PRIQ_0_SSID, evt[0]) : IOMMU_NO_PASID; last = FIELD_GET(PRIQ_0_PRG_LAST, evt[0]); grpid = FIELD_GET(PRIQ_1_PRG_IDX, evt[1]); @@ -1748,7 +1748,7 @@ arm_smmu_atc_inv_to_cmd(int ssid, unsigned long iova, size_t size, */ *cmd = (struct arm_smmu_cmdq_ent) { .opcode = CMDQ_OP_ATC_INV, - .substream_valid = !!ssid, + .substream_valid = (ssid != IOMMU_NO_PASID), .atc.ssid = ssid, }; @@ -1795,7 +1795,7 @@ static int arm_smmu_atc_inv_master(struct arm_smmu_master *master) struct arm_smmu_cmdq_ent cmd; struct arm_smmu_cmdq_batch cmds; - arm_smmu_atc_inv_to_cmd(0, 0, 0, &cmd); + arm_smmu_atc_inv_to_cmd(IOMMU_NO_PASID, 0, 0, &cmd); cmds.num = 0; for (i = 0; i < master->num_streams; i++) { @@ -1875,7 +1875,7 @@ static void arm_smmu_tlb_inv_context(void *cookie) cmd.tlbi.vmid = smmu_domain->s2_cfg.vmid; arm_smmu_cmdq_issue_cmd_with_sync(smmu, &cmd); } - arm_smmu_atc_inv_domain(smmu_domain, 0, 0, 0); + arm_smmu_atc_inv_domain(smmu_domain, IOMMU_NO_PASID, 0, 0); } static void __arm_smmu_tlb_inv_range(struct arm_smmu_cmdq_ent *cmd, @@ -1968,7 +1968,7 @@ static void arm_smmu_tlb_inv_range_domain(unsigned long iova, size_t size, * Unfortunately, this can't be leaf-only since we may have * zapped an entire table. */ - arm_smmu_atc_inv_domain(smmu_domain, 0, iova, size); + arm_smmu_atc_inv_domain(smmu_domain, IOMMU_NO_PASID, iova, size); } void arm_smmu_tlb_inv_range_asid(unsigned long iova, size_t size, int asid, @@ -2142,7 +2142,7 @@ static int arm_smmu_domain_finalise_s1(struct arm_smmu_domain *smmu_domain, * the master has been added to the devices list for this domain. * This isn't an issue because the STE hasn't been installed yet. */ - ret = arm_smmu_write_ctx_desc(smmu_domain, 0, &cfg->cd); + ret = arm_smmu_write_ctx_desc(smmu_domain, IOMMU_NO_PASID, &cfg->cd); if (ret) goto out_free_cd_tables; @@ -2328,7 +2328,7 @@ static void arm_smmu_enable_ats(struct arm_smmu_master *master) pdev = to_pci_dev(master->dev); atomic_inc(&smmu_domain->nr_ats_masters); - arm_smmu_atc_inv_domain(smmu_domain, 0, 0, 0); + arm_smmu_atc_inv_domain(smmu_domain, IOMMU_NO_PASID, 0, 0); if (pci_enable_ats(pdev, stu)) dev_err(master->dev, "Failed to enable ATS (STU %zu)\n", stu); } diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 5c8c5cdc36cf..ddff43def3ab 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -877,7 +877,7 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, } /* For request-without-pasid, get the pasid from context entry */ if (intel_iommu_sm && pasid == IOMMU_PASID_INVALID) - pasid = PASID_RID2PASID; + pasid = IOMMU_NO_PASID; dir_index = pasid >> PASID_PDE_SHIFT; pde = &dir[dir_index]; @@ -1449,7 +1449,7 @@ static void __iommu_flush_dev_iotlb(struct device_domain_info *info, qdep = info->ats_qdep; qi_flush_dev_iotlb(info->iommu, sid, info->pfsid, qdep, addr, mask); - quirk_extra_dev_tlb_flush(info, addr, mask, PASID_RID2PASID, qdep); + quirk_extra_dev_tlb_flush(info, addr, mask, IOMMU_NO_PASID, qdep); } static void iommu_flush_dev_iotlb(struct dmar_domain *domain, @@ -1484,7 +1484,7 @@ static void iommu_flush_iotlb_psi(struct intel_iommu *iommu, ih = 1 << 6; if (domain->use_first_level) { - qi_flush_piotlb(iommu, did, PASID_RID2PASID, addr, pages, ih); + qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, pages, ih); } else { unsigned long bitmask = aligned_pages - 1; @@ -1554,7 +1554,7 @@ static void intel_flush_iotlb_all(struct iommu_domain *domain) u16 did = domain_id_iommu(dmar_domain, iommu); if (dmar_domain->use_first_level) - qi_flush_piotlb(iommu, did, PASID_RID2PASID, 0, -1, 0); + qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, 0, -1, 0); else iommu->flush.flush_iotlb(iommu, did, 0, 0, DMA_TLB_DSI_FLUSH); @@ -1940,7 +1940,7 @@ static int domain_context_mapping_one(struct dmar_domain *domain, context_pdts(pds); /* Setup the RID_PASID field: */ - context_set_sm_rid2pasid(context, PASID_RID2PASID); + context_set_sm_rid2pasid(context, IOMMU_NO_PASID); /* * Setup the Device-TLB enable bit and Page request @@ -2420,13 +2420,13 @@ static int dmar_domain_attach_device(struct dmar_domain *domain, /* Setup the PASID entry for requests without PASID: */ if (hw_pass_through && domain_type_is_si(domain)) ret = intel_pasid_setup_pass_through(iommu, domain, - dev, PASID_RID2PASID); + dev, IOMMU_NO_PASID); else if (domain->use_first_level) ret = domain_setup_first_level(iommu, domain, dev, - PASID_RID2PASID); + IOMMU_NO_PASID); else ret = intel_pasid_setup_second_level(iommu, domain, - dev, PASID_RID2PASID); + dev, IOMMU_NO_PASID); if (ret) { dev_err(dev, "Setup RID2PASID failed\n"); device_block_translation(dev); @@ -3968,7 +3968,7 @@ static void dmar_remove_one_dev_info(struct device *dev) if (!dev_is_real_dma_subdevice(info->dev)) { if (dev_is_pci(info->dev) && sm_supported(iommu)) intel_pasid_tear_down_entry(iommu, info->dev, - PASID_RID2PASID, false); + IOMMU_NO_PASID, false); iommu_disable_pci_caps(info); domain_context_clear(info); @@ -3997,7 +3997,7 @@ static void device_block_translation(struct device *dev) if (!dev_is_real_dma_subdevice(dev)) { if (sm_supported(iommu)) intel_pasid_tear_down_entry(iommu, dev, - PASID_RID2PASID, false); + IOMMU_NO_PASID, false); else domain_context_clear(info); } @@ -4331,7 +4331,7 @@ static void domain_set_force_snooping(struct dmar_domain *domain) list_for_each_entry(info, &domain->devices, link) intel_pasid_setup_page_snoop_control(info->iommu, info->dev, - PASID_RID2PASID); + IOMMU_NO_PASID); } static bool intel_iommu_enforce_cache_coherency(struct iommu_domain *domain) @@ -4987,7 +4987,7 @@ void quirk_extra_dev_tlb_flush(struct device_domain_info *info, return; sid = PCI_DEVID(info->bus, info->devfn); - if (pasid == PASID_RID2PASID) { + if (pasid == IOMMU_NO_PASID) { qi_flush_dev_iotlb(info->iommu, sid, info->pfsid, qdep, address, mask); } else { diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index c5d479770e12..23dca3bc319d 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -438,7 +438,7 @@ devtlb_invalidation_with_pasid(struct intel_iommu *iommu, * SVA usage, device could do DMA with multiple PASIDs. It is more * efficient to flush devTLB specific to the PASID. */ - if (pasid == PASID_RID2PASID) + if (pasid == IOMMU_NO_PASID) qi_flush_dev_iotlb(iommu, sid, pfsid, qdep, 0, 64 - VTD_PAGE_SHIFT); else qi_flush_dev_iotlb_pasid(iommu, sid, pfsid, pasid, qdep, 0, 64 - VTD_PAGE_SHIFT); diff --git a/drivers/iommu/intel/pasid.h b/drivers/iommu/intel/pasid.h index d6b7d21244b1..4e9e68c3c388 100644 --- a/drivers/iommu/intel/pasid.h +++ b/drivers/iommu/intel/pasid.h @@ -10,8 +10,6 @@ #ifndef __INTEL_PASID_H #define __INTEL_PASID_H -#define PASID_RID2PASID 0x0 -#define PASID_MIN 0x1 #define PASID_MAX 0x100000 #define PASID_PTE_MASK 0x3F #define PASID_PTE_PRESENT 1 diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d31642596675..2870bc29d456 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -196,6 +196,7 @@ enum iommu_dev_features { IOMMU_DEV_FEAT_IOPF, }; +#define IOMMU_NO_PASID (0U) /* Reserved for DMA w/o PASID */ #define IOMMU_PASID_INVALID (-1U) typedef unsigned int ioasid_t; -- cgit v1.2.3 From 2dcebc7ddce7ffd4015824227c7623558b89d721 Mon Sep 17 00:00:00 2001 From: Jacob Pan Date: Wed, 9 Aug 2023 20:47:55 +0800 Subject: iommu: Move global PASID allocation from SVA to core Intel ENQCMD requires a single PASID to be shared between multiple devices, as the PASID is stored in a single MSR register per-process and userspace can use only that one PASID. This means that the PASID allocation for any ENQCMD using device driver must always come from a shared global pool, regardless of what kind of domain the PASID will be used with. Split the code for the global PASID allocator into iommu_alloc/free_global_pasid() so that drivers can attach non-SVA domains to PASIDs as well. This patch moves global PASID allocation APIs from SVA to IOMMU APIs. Reserved PASIDs, currently only RID_PASID, are excluded from the global PASID allocation. It is expected that device drivers will use the allocated PASIDs to attach to appropriate IOMMU domains for use. Reviewed-by: Lu Baolu Reviewed-by: Kevin Tian Reviewed-by: Jason Gunthorpe Signed-off-by: Jacob Pan Link: https://lore.kernel.org/r/20230802212427.1497170-3-jacob.jun.pan@linux.intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/iommu-sva.c | 29 ++++++++++------------------- drivers/iommu/iommu.c | 28 ++++++++++++++++++++++++++++ include/linux/iommu.h | 10 ++++++++++ 3 files changed, 48 insertions(+), 19 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu-sva.c b/drivers/iommu/iommu-sva.c index 05c0fb2acbc4..b78671a8a914 100644 --- a/drivers/iommu/iommu-sva.c +++ b/drivers/iommu/iommu-sva.c @@ -10,34 +10,30 @@ #include "iommu-sva.h" static DEFINE_MUTEX(iommu_sva_lock); -static DEFINE_IDA(iommu_global_pasid_ida); /* Allocate a PASID for the mm within range (inclusive) */ -static int iommu_sva_alloc_pasid(struct mm_struct *mm, ioasid_t min, ioasid_t max) +static int iommu_sva_alloc_pasid(struct mm_struct *mm, struct device *dev) { + ioasid_t pasid; int ret = 0; - if (min == IOMMU_PASID_INVALID || - max == IOMMU_PASID_INVALID || - min == 0 || max < min) - return -EINVAL; - if (!arch_pgtable_dma_compat(mm)) return -EBUSY; mutex_lock(&iommu_sva_lock); /* Is a PASID already associated with this mm? */ if (mm_valid_pasid(mm)) { - if (mm->pasid < min || mm->pasid > max) + if (mm->pasid >= dev->iommu->max_pasids) ret = -EOVERFLOW; goto out; } - ret = ida_alloc_range(&iommu_global_pasid_ida, min, max, GFP_KERNEL); - if (ret < 0) + pasid = iommu_alloc_global_pasid(dev); + if (pasid == IOMMU_PASID_INVALID) { + ret = -ENOSPC; goto out; - - mm->pasid = ret; + } + mm->pasid = pasid; ret = 0; out: mutex_unlock(&iommu_sva_lock); @@ -64,15 +60,10 @@ struct iommu_sva *iommu_sva_bind_device(struct device *dev, struct mm_struct *mm { struct iommu_domain *domain; struct iommu_sva *handle; - ioasid_t max_pasids; int ret; - max_pasids = dev->iommu->max_pasids; - if (!max_pasids) - return ERR_PTR(-EOPNOTSUPP); - /* Allocate mm->pasid if necessary. */ - ret = iommu_sva_alloc_pasid(mm, 1, max_pasids - 1); + ret = iommu_sva_alloc_pasid(mm, dev); if (ret) return ERR_PTR(ret); @@ -217,5 +208,5 @@ void mm_pasid_drop(struct mm_struct *mm) if (likely(!mm_valid_pasid(mm))) return; - ida_free(&iommu_global_pasid_ida, mm->pasid); + iommu_free_global_pasid(mm->pasid); } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index caaf563d38ae..d199d144460c 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -39,6 +39,7 @@ 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); @@ -3400,3 +3401,30 @@ 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); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 2870bc29d456..6b821cb715d5 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -197,6 +197,7 @@ enum iommu_dev_features { }; #define IOMMU_NO_PASID (0U) /* Reserved for DMA w/o PASID */ +#define IOMMU_FIRST_GLOBAL_PASID (1U) /*starting range for allocation */ #define IOMMU_PASID_INVALID (-1U) typedef unsigned int ioasid_t; @@ -728,6 +729,8 @@ void iommu_detach_device_pasid(struct iommu_domain *domain, struct iommu_domain * iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, unsigned int type); +ioasid_t iommu_alloc_global_pasid(struct device *dev); +void iommu_free_global_pasid(ioasid_t pasid); #else /* CONFIG_IOMMU_API */ struct iommu_ops {}; @@ -1089,6 +1092,13 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, { return NULL; } + +static inline ioasid_t iommu_alloc_global_pasid(struct device *dev) +{ + return IOMMU_PASID_INVALID; +} + +static inline void iommu_free_global_pasid(ioasid_t pasid) {} #endif /* CONFIG_IOMMU_API */ /** -- cgit v1.2.3 From ac1a3483febd0ec8fced1ccb15da349dcc360fbb Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:47:56 +0800 Subject: iommu/vt-d: Add domain_flush_pasid_iotlb() The VT-d spec requires to use PASID-based-IOTLB invalidation descriptor to invalidate IOTLB and the paging-structure caches for a first-stage page table. Add a generic helper to do this. RID2PASID is used if the domain has been attached to a physical device, otherwise real PASIDs that the domain has been attached to will be used. The 'real' PASID attachment is handled in the subsequent change. Signed-off-by: Lu Baolu Signed-off-by: Jacob Pan Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230802212427.1497170-4-jacob.jun.pan@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index ddff43def3ab..ecb7e44377a8 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -1467,6 +1467,18 @@ static void iommu_flush_dev_iotlb(struct dmar_domain *domain, spin_unlock_irqrestore(&domain->lock, flags); } +static void domain_flush_pasid_iotlb(struct intel_iommu *iommu, + struct dmar_domain *domain, u64 addr, + unsigned long npages, bool ih) +{ + u16 did = domain_id_iommu(domain, iommu); + unsigned long flags; + + spin_lock_irqsave(&domain->lock, flags); + qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, npages, ih); + spin_unlock_irqrestore(&domain->lock, flags); +} + static void iommu_flush_iotlb_psi(struct intel_iommu *iommu, struct dmar_domain *domain, unsigned long pfn, unsigned int pages, @@ -1484,7 +1496,7 @@ static void iommu_flush_iotlb_psi(struct intel_iommu *iommu, ih = 1 << 6; if (domain->use_first_level) { - qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, pages, ih); + domain_flush_pasid_iotlb(iommu, domain, addr, pages, ih); } else { unsigned long bitmask = aligned_pages - 1; @@ -1554,7 +1566,7 @@ static void intel_flush_iotlb_all(struct iommu_domain *domain) u16 did = domain_id_iommu(dmar_domain, iommu); if (dmar_domain->use_first_level) - qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, 0, -1, 0); + domain_flush_pasid_iotlb(iommu, dmar_domain, 0, -1, 0); else iommu->flush.flush_iotlb(iommu, did, 0, 0, DMA_TLB_DSI_FLUSH); -- cgit v1.2.3 From b61701881f2ff809e9cd229f129bdc4e09ca688e Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:47:57 +0800 Subject: iommu/vt-d: Remove pasid_mutex The pasid_mutex was used to protect the paths of set/remove_dev_pasid(). It's duplicate with iommu_sva_lock. Remove it to avoid duplicate code. Signed-off-by: Lu Baolu Signed-off-by: Jacob Pan Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230802212427.1497170-5-jacob.jun.pan@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/svm.c | 45 +++++---------------------------------------- 1 file changed, 5 insertions(+), 40 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index e95b339e9cdc..2a82864e9d57 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -259,8 +259,6 @@ static const struct mmu_notifier_ops intel_mmuops = { .invalidate_range = intel_invalidate_range, }; -static DEFINE_MUTEX(pasid_mutex); - static int pasid_to_svm_sdev(struct device *dev, unsigned int pasid, struct intel_svm **rsvm, struct intel_svm_dev **rsdev) @@ -268,10 +266,6 @@ static int pasid_to_svm_sdev(struct device *dev, unsigned int pasid, struct intel_svm_dev *sdev = NULL; struct intel_svm *svm; - /* The caller should hold the pasid_mutex lock */ - if (WARN_ON(!mutex_is_locked(&pasid_mutex))) - return -EINVAL; - if (pasid == IOMMU_PASID_INVALID || pasid >= PASID_MAX) return -EINVAL; @@ -371,22 +365,19 @@ free_svm: return ret; } -/* Caller must hold pasid_mutex */ -static int intel_svm_unbind_mm(struct device *dev, u32 pasid) +void intel_svm_remove_dev_pasid(struct device *dev, u32 pasid) { struct intel_svm_dev *sdev; struct intel_iommu *iommu; struct intel_svm *svm; struct mm_struct *mm; - int ret = -EINVAL; iommu = device_to_iommu(dev, NULL, NULL); if (!iommu) - goto out; + return; - ret = pasid_to_svm_sdev(dev, pasid, &svm, &sdev); - if (ret) - goto out; + if (pasid_to_svm_sdev(dev, pasid, &svm, &sdev)) + return; mm = svm->mm; if (sdev) { @@ -418,8 +409,6 @@ static int intel_svm_unbind_mm(struct device *dev, u32 pasid) kfree(svm); } } -out: - return ret; } /* Page request queue descriptor */ @@ -520,19 +509,7 @@ prq_retry: goto prq_retry; } - /* - * A work in IO page fault workqueue may try to lock pasid_mutex now. - * Holding pasid_mutex while waiting in iopf_queue_flush_dev() for - * all works in the workqueue to finish may cause deadlock. - * - * It's unnecessary to hold pasid_mutex in iopf_queue_flush_dev(). - * Unlock it to allow the works to be handled while waiting for - * them to finish. - */ - lockdep_assert_held(&pasid_mutex); - mutex_unlock(&pasid_mutex); iopf_queue_flush_dev(dev); - mutex_lock(&pasid_mutex); /* * Perform steps described in VT-d spec CH7.10 to drain page @@ -827,26 +804,14 @@ out: return ret; } -void intel_svm_remove_dev_pasid(struct device *dev, ioasid_t pasid) -{ - mutex_lock(&pasid_mutex); - intel_svm_unbind_mm(dev, pasid); - mutex_unlock(&pasid_mutex); -} - static int intel_svm_set_dev_pasid(struct iommu_domain *domain, struct device *dev, ioasid_t pasid) { struct device_domain_info *info = dev_iommu_priv_get(dev); struct intel_iommu *iommu = info->iommu; struct mm_struct *mm = domain->mm; - int ret; - mutex_lock(&pasid_mutex); - ret = intel_svm_bind_mm(iommu, dev, mm); - mutex_unlock(&pasid_mutex); - - return ret; + return intel_svm_bind_mm(iommu, dev, mm); } static void intel_svm_domain_free(struct iommu_domain *domain) -- cgit v1.2.3 From 154786235d0156ddce9575f13e23f3169e688435 Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:47:58 +0800 Subject: iommu/vt-d: Make prq draining code generic Currently draining page requests and responses for a pasid is part of SVA implementation. This is because the driver only supports attaching an SVA domain to a device pasid. As we are about to support attaching other types of domains to a device pasid, the prq draining code becomes generic. Signed-off-by: Lu Baolu Signed-off-by: Jacob Pan Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230802212427.1497170-6-jacob.jun.pan@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 30 +++++++++++++++++++----------- drivers/iommu/intel/iommu.h | 2 ++ drivers/iommu/intel/svm.c | 17 ++--------------- 3 files changed, 23 insertions(+), 26 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index ecb7e44377a8..574021c36db5 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -4727,21 +4727,29 @@ static void intel_iommu_remove_dev_pasid(struct device *dev, ioasid_t pasid) struct intel_iommu *iommu = device_to_iommu(dev, NULL, NULL); struct iommu_domain *domain; - /* Domain type specific cleanup: */ domain = iommu_get_domain_for_dev_pasid(dev, pasid, 0); - if (domain) { - switch (domain->type) { - case IOMMU_DOMAIN_SVA: - intel_svm_remove_dev_pasid(dev, pasid); - break; - default: - /* should never reach here */ - WARN_ON(1); - break; - } + if (WARN_ON_ONCE(!domain)) + goto out_tear_down; + + /* + * The SVA implementation needs to handle its own stuffs like the mm + * notification. Before consolidating that code into iommu core, let + * the intel sva code handle it. + */ + if (domain->type == IOMMU_DOMAIN_SVA) { + intel_svm_remove_dev_pasid(dev, pasid); + goto out_tear_down; } + /* + * Should never reach here until we add support for attaching + * non-SVA domain to a pasid. + */ + WARN_ON(1); + +out_tear_down: intel_pasid_tear_down_entry(iommu, dev, pasid, false); + intel_drain_pasid_prq(dev, pasid); } const struct iommu_ops intel_iommu_ops = { diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 1c5e1d88862b..6d94a29f5d52 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -844,6 +844,7 @@ int intel_svm_page_response(struct device *dev, struct iommu_fault_event *evt, struct iommu_page_response *msg); struct iommu_domain *intel_svm_domain_alloc(void); void intel_svm_remove_dev_pasid(struct device *dev, ioasid_t pasid); +void intel_drain_pasid_prq(struct device *dev, u32 pasid); struct intel_svm_dev { struct list_head list; @@ -862,6 +863,7 @@ struct intel_svm { }; #else static inline void intel_svm_check(struct intel_iommu *iommu) {} +static inline void intel_drain_pasid_prq(struct device *dev, u32 pasid) {} static inline struct iommu_domain *intel_svm_domain_alloc(void) { return NULL; diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index 2a82864e9d57..9fbae9af6615 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -26,8 +26,6 @@ #include "trace.h" static irqreturn_t prq_event_thread(int irq, void *d); -static void intel_svm_drain_prq(struct device *dev, u32 pasid); -#define to_intel_svm_dev(handle) container_of(handle, struct intel_svm_dev, sva) static DEFINE_XARRAY_ALLOC(pasid_private_array); static int pasid_private_add(ioasid_t pasid, void *priv) @@ -382,17 +380,6 @@ void intel_svm_remove_dev_pasid(struct device *dev, u32 pasid) if (sdev) { list_del_rcu(&sdev->list); - /* - * Flush the PASID cache and IOTLB for this device. - * Note that we do depend on the hardware *not* using - * the PASID any more. Just as we depend on other - * devices never using PASIDs that they have no right - * to use. We have a *shared* PASID table, because it's - * large and has to be physically contiguous. So it's - * hard to be as defensive as we might like. - */ - intel_pasid_tear_down_entry(iommu, dev, svm->pasid, false); - intel_svm_drain_prq(dev, svm->pasid); kfree_rcu(sdev, rcu); if (list_empty(&svm->devs)) { @@ -449,7 +436,7 @@ static bool is_canonical_address(u64 addr) } /** - * intel_svm_drain_prq - Drain page requests and responses for a pasid + * intel_drain_pasid_prq - Drain page requests and responses for a pasid * @dev: target device * @pasid: pasid for draining * @@ -463,7 +450,7 @@ static bool is_canonical_address(u64 addr) * described in VT-d spec CH7.10 to drain all page requests and page * responses pending in the hardware. */ -static void intel_svm_drain_prq(struct device *dev, u32 pasid) +void intel_drain_pasid_prq(struct device *dev, u32 pasid) { struct device_domain_info *info; struct dmar_domain *domain; -- cgit v1.2.3 From 37f900e7180ac67c73dcfae308625c87194025c0 Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:47:59 +0800 Subject: iommu/vt-d: Prepare for set_dev_pasid callback The domain_flush_pasid_iotlb() helper function is used to flush the IOTLB entries for a given PASID. Previously, this function assumed that RID2PASID was only used for the first-level DMA translation. However, with the introduction of the set_dev_pasid callback, this assumption is no longer valid. Add a check before using the RID2PASID for PASID invalidation. This check ensures that the domain has been attached to a physical device before using RID2PASID. Signed-off-by: Lu Baolu Signed-off-by: Jacob Pan Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230802212427.1497170-7-jacob.jun.pan@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 574021c36db5..19585800dcc9 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -1475,7 +1475,8 @@ static void domain_flush_pasid_iotlb(struct intel_iommu *iommu, unsigned long flags; spin_lock_irqsave(&domain->lock, flags); - qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, npages, ih); + if (!list_empty(&domain->devices)) + qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, npages, ih); spin_unlock_irqrestore(&domain->lock, flags); } -- cgit v1.2.3 From 7d0c9da6c1509664d96488042bacc02308ca33b2 Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:48:00 +0800 Subject: iommu/vt-d: Add set_dev_pasid callback for dma domain This allows the upper layers to set a domain to a PASID of a device if the PASID feature is supported by the IOMMU hardware. The typical use cases are, for example, kernel DMA with PASID and hardware assisted mediated device drivers. The attaching device and pasid information is tracked in a per-domain list and is used for IOTLB and devTLB invalidation. Signed-off-by: Lu Baolu Signed-off-by: Jacob Pan Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230802212427.1497170-8-jacob.jun.pan@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 104 +++++++++++++++++++++++++++++++++++++++++--- drivers/iommu/intel/iommu.h | 7 +++ 2 files changed, 106 insertions(+), 5 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 19585800dcc9..4b5ed58ee422 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -1359,6 +1359,7 @@ domain_lookup_dev_info(struct dmar_domain *domain, static void domain_update_iotlb(struct dmar_domain *domain) { + struct dev_pasid_info *dev_pasid; struct device_domain_info *info; bool has_iotlb_device = false; unsigned long flags; @@ -1370,6 +1371,14 @@ static void domain_update_iotlb(struct dmar_domain *domain) break; } } + + list_for_each_entry(dev_pasid, &domain->dev_pasids, link_domain) { + info = dev_iommu_priv_get(dev_pasid->dev); + if (info->ats_enabled) { + has_iotlb_device = true; + break; + } + } domain->has_iotlb_device = has_iotlb_device; spin_unlock_irqrestore(&domain->lock, flags); } @@ -1455,6 +1464,7 @@ static void __iommu_flush_dev_iotlb(struct device_domain_info *info, static void iommu_flush_dev_iotlb(struct dmar_domain *domain, u64 addr, unsigned mask) { + struct dev_pasid_info *dev_pasid; struct device_domain_info *info; unsigned long flags; @@ -1464,6 +1474,19 @@ static void iommu_flush_dev_iotlb(struct dmar_domain *domain, spin_lock_irqsave(&domain->lock, flags); list_for_each_entry(info, &domain->devices, link) __iommu_flush_dev_iotlb(info, addr, mask); + + list_for_each_entry(dev_pasid, &domain->dev_pasids, link_domain) { + info = dev_iommu_priv_get(dev_pasid->dev); + + if (!info->ats_enabled) + continue; + + qi_flush_dev_iotlb_pasid(info->iommu, + PCI_DEVID(info->bus, info->devfn), + info->pfsid, dev_pasid->pasid, + info->ats_qdep, addr, + mask); + } spin_unlock_irqrestore(&domain->lock, flags); } @@ -1472,9 +1495,13 @@ static void domain_flush_pasid_iotlb(struct intel_iommu *iommu, unsigned long npages, bool ih) { u16 did = domain_id_iommu(domain, iommu); + struct dev_pasid_info *dev_pasid; unsigned long flags; spin_lock_irqsave(&domain->lock, flags); + list_for_each_entry(dev_pasid, &domain->dev_pasids, link_domain) + qi_flush_piotlb(iommu, did, dev_pasid->pasid, addr, npages, ih); + if (!list_empty(&domain->devices)) qi_flush_piotlb(iommu, did, IOMMU_NO_PASID, addr, npages, ih); spin_unlock_irqrestore(&domain->lock, flags); @@ -1739,6 +1766,7 @@ static struct dmar_domain *alloc_domain(unsigned int type) domain->use_first_level = true; domain->has_iotlb_device = false; INIT_LIST_HEAD(&domain->devices); + INIT_LIST_HEAD(&domain->dev_pasids); spin_lock_init(&domain->lock); xa_init(&domain->iommu_array); @@ -4726,7 +4754,10 @@ static void intel_iommu_iotlb_sync_map(struct iommu_domain *domain, static void intel_iommu_remove_dev_pasid(struct device *dev, ioasid_t pasid) { struct intel_iommu *iommu = device_to_iommu(dev, NULL, NULL); + struct dev_pasid_info *curr, *dev_pasid = NULL; + struct dmar_domain *dmar_domain; struct iommu_domain *domain; + unsigned long flags; domain = iommu_get_domain_for_dev_pasid(dev, pasid, 0); if (WARN_ON_ONCE(!domain)) @@ -4742,17 +4773,79 @@ static void intel_iommu_remove_dev_pasid(struct device *dev, ioasid_t pasid) goto out_tear_down; } - /* - * Should never reach here until we add support for attaching - * non-SVA domain to a pasid. - */ - WARN_ON(1); + dmar_domain = to_dmar_domain(domain); + spin_lock_irqsave(&dmar_domain->lock, flags); + list_for_each_entry(curr, &dmar_domain->dev_pasids, link_domain) { + if (curr->dev == dev && curr->pasid == pasid) { + list_del(&curr->link_domain); + dev_pasid = curr; + break; + } + } + WARN_ON_ONCE(!dev_pasid); + spin_unlock_irqrestore(&dmar_domain->lock, flags); + domain_detach_iommu(dmar_domain, iommu); + kfree(dev_pasid); out_tear_down: intel_pasid_tear_down_entry(iommu, dev, pasid, false); intel_drain_pasid_prq(dev, pasid); } +static int intel_iommu_set_dev_pasid(struct iommu_domain *domain, + struct device *dev, ioasid_t pasid) +{ + struct device_domain_info *info = dev_iommu_priv_get(dev); + struct dmar_domain *dmar_domain = to_dmar_domain(domain); + struct intel_iommu *iommu = info->iommu; + struct dev_pasid_info *dev_pasid; + unsigned long flags; + int ret; + + if (!pasid_supported(iommu) || dev_is_real_dma_subdevice(dev)) + return -EOPNOTSUPP; + + if (context_copied(iommu, info->bus, info->devfn)) + return -EBUSY; + + ret = prepare_domain_attach_device(domain, dev); + if (ret) + return ret; + + dev_pasid = kzalloc(sizeof(*dev_pasid), GFP_KERNEL); + if (!dev_pasid) + return -ENOMEM; + + ret = domain_attach_iommu(dmar_domain, iommu); + if (ret) + goto out_free; + + if (domain_type_is_si(dmar_domain)) + ret = intel_pasid_setup_pass_through(iommu, dmar_domain, + dev, pasid); + else if (dmar_domain->use_first_level) + ret = domain_setup_first_level(iommu, dmar_domain, + dev, pasid); + else + ret = intel_pasid_setup_second_level(iommu, dmar_domain, + dev, pasid); + if (ret) + goto out_detach_iommu; + + dev_pasid->dev = dev; + dev_pasid->pasid = pasid; + spin_lock_irqsave(&dmar_domain->lock, flags); + list_add(&dev_pasid->link_domain, &dmar_domain->dev_pasids); + spin_unlock_irqrestore(&dmar_domain->lock, flags); + + return 0; +out_detach_iommu: + domain_detach_iommu(dmar_domain, iommu); +out_free: + kfree(dev_pasid); + return ret; +} + const struct iommu_ops intel_iommu_ops = { .capable = intel_iommu_capable, .domain_alloc = intel_iommu_domain_alloc, @@ -4772,6 +4865,7 @@ const struct iommu_ops intel_iommu_ops = { #endif .default_domain_ops = &(const struct iommu_domain_ops) { .attach_dev = intel_iommu_attach_device, + .set_dev_pasid = intel_iommu_set_dev_pasid, .map_pages = intel_iommu_map_pages, .unmap_pages = intel_iommu_unmap_pages, .iotlb_sync_map = intel_iommu_iotlb_sync_map, diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 6d94a29f5d52..c18fb699c87a 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -595,6 +595,7 @@ struct dmar_domain { spinlock_t lock; /* Protect device tracking lists */ struct list_head devices; /* all devices' list */ + struct list_head dev_pasids; /* all attached pasids */ struct dma_pte *pgd; /* virtual address */ int gaw; /* max guest address width */ @@ -717,6 +718,12 @@ struct device_domain_info { struct pasid_table *pasid_table; /* pasid table */ }; +struct dev_pasid_info { + struct list_head link_domain; /* link to domain siblings */ + struct device *dev; + ioasid_t pasid; +}; + static inline void __iommu_flush_cache( struct intel_iommu *iommu, void *addr, int size) { -- cgit v1.2.3 From f5ccf55e10281ae4523b83fe87f2c27fd634dc9d Mon Sep 17 00:00:00 2001 From: Jacob Pan Date: Wed, 9 Aug 2023 20:48:01 +0800 Subject: dmaengine/idxd: Re-enable kernel workqueue under DMA API Kernel workqueues were disabled due to flawed use of kernel VA and SVA API. Now that we have the support for attaching PASID to the device's default domain and the ability to reserve global PASIDs from SVA APIs, we can re-enable the kernel work queues and use them under DMA API. We also use non-privileged access for in-kernel DMA to be consistent with the IOMMU settings. Consequently, interrupt for user privilege is enabled for work completion IRQs. Link: https://lore.kernel.org/linux-iommu/20210511194726.GP1002214@nvidia.com/ Tested-by: Tony Zhu Reviewed-by: Dave Jiang Reviewed-by: Fenghua Yu Reviewed-by: Lu Baolu Reviewed-by: Jason Gunthorpe Acked-by: Vinod Koul Signed-off-by: Jacob Pan Link: https://lore.kernel.org/r/20230802212427.1497170-9-jacob.jun.pan@linux.intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/dma/idxd/device.c | 39 ++++++++++++---------------------- drivers/dma/idxd/dma.c | 5 +++-- drivers/dma/idxd/idxd.h | 9 ++++++++ drivers/dma/idxd/init.c | 54 +++++++++++++++++++++++++++++++++++++++++++---- drivers/dma/idxd/sysfs.c | 7 ------ 5 files changed, 76 insertions(+), 38 deletions(-) (limited to 'drivers') diff --git a/drivers/dma/idxd/device.c b/drivers/dma/idxd/device.c index 5abbcc61c528..169b7ade8919 100644 --- a/drivers/dma/idxd/device.c +++ b/drivers/dma/idxd/device.c @@ -299,21 +299,6 @@ void idxd_wqs_unmap_portal(struct idxd_device *idxd) } } -static void __idxd_wq_set_priv_locked(struct idxd_wq *wq, int priv) -{ - struct idxd_device *idxd = wq->idxd; - union wqcfg wqcfg; - unsigned int offset; - - offset = WQCFG_OFFSET(idxd, wq->id, WQCFG_PRIVL_IDX); - spin_lock(&idxd->dev_lock); - wqcfg.bits[WQCFG_PRIVL_IDX] = ioread32(idxd->reg_base + offset); - wqcfg.priv = priv; - wq->wqcfg->bits[WQCFG_PRIVL_IDX] = wqcfg.bits[WQCFG_PRIVL_IDX]; - iowrite32(wqcfg.bits[WQCFG_PRIVL_IDX], idxd->reg_base + offset); - spin_unlock(&idxd->dev_lock); -} - static void __idxd_wq_set_pasid_locked(struct idxd_wq *wq, int pasid) { struct idxd_device *idxd = wq->idxd; @@ -1423,15 +1408,14 @@ int drv_enable_wq(struct idxd_wq *wq) } /* - * In the event that the WQ is configurable for pasid and priv bits. - * For kernel wq, the driver should setup the pasid, pasid_en, and priv bit. - * However, for non-kernel wq, the driver should only set the pasid_en bit for - * shared wq. A dedicated wq that is not 'kernel' type will configure pasid and + * In the event that the WQ is configurable for pasid, the driver + * should setup the pasid, pasid_en bit. This is true for both kernel + * and user shared workqueues. There is no need to setup priv bit in + * that in-kernel DMA will also do user privileged requests. + * A dedicated wq that is not 'kernel' type will configure pasid and * pasid_en later on so there is no need to setup. */ if (test_bit(IDXD_FLAG_CONFIGURABLE, &idxd->flags)) { - int priv = 0; - if (wq_pasid_enabled(wq)) { if (is_idxd_wq_kernel(wq) || wq_shared(wq)) { u32 pasid = wq_dedicated(wq) ? idxd->pasid : 0; @@ -1439,10 +1423,6 @@ int drv_enable_wq(struct idxd_wq *wq) __idxd_wq_set_pasid_locked(wq, pasid); } } - - if (is_idxd_wq_kernel(wq)) - priv = 1; - __idxd_wq_set_priv_locked(wq, priv); } rc = 0; @@ -1550,6 +1530,15 @@ int idxd_device_drv_probe(struct idxd_dev *idxd_dev) if (rc < 0) return -ENXIO; + /* + * System PASID is preserved across device disable/enable cycle, but + * genconfig register content gets cleared during device reset. We + * need to re-enable user interrupts for kernel work queue completion + * IRQ to function. + */ + if (idxd->pasid != IOMMU_PASID_INVALID) + idxd_set_user_intr(idxd, 1); + rc = idxd_device_evl_setup(idxd); if (rc < 0) { idxd->cmd_status = IDXD_SCMD_DEV_EVL_ERR; diff --git a/drivers/dma/idxd/dma.c b/drivers/dma/idxd/dma.c index eb35ca313684..07623fb0f52f 100644 --- a/drivers/dma/idxd/dma.c +++ b/drivers/dma/idxd/dma.c @@ -75,9 +75,10 @@ static inline void idxd_prep_desc_common(struct idxd_wq *wq, hw->xfer_size = len; /* * For dedicated WQ, this field is ignored and HW will use the WQCFG.priv - * field instead. This field should be set to 1 for kernel descriptors. + * field instead. This field should be set to 0 for kernel descriptors + * since kernel DMA on VT-d supports "user" privilege only. */ - hw->priv = 1; + hw->priv = 0; hw->completion_addr = compl; } diff --git a/drivers/dma/idxd/idxd.h b/drivers/dma/idxd/idxd.h index 5428a2e1b1ec..502be9db63f4 100644 --- a/drivers/dma/idxd/idxd.h +++ b/drivers/dma/idxd/idxd.h @@ -473,6 +473,15 @@ static inline struct idxd_device *ie_to_idxd(struct idxd_irq_entry *ie) return container_of(ie, struct idxd_device, ie); } +static inline void idxd_set_user_intr(struct idxd_device *idxd, bool enable) +{ + union gencfg_reg reg; + + reg.bits = ioread32(idxd->reg_base + IDXD_GENCFG_OFFSET); + reg.user_int_en = enable; + iowrite32(reg.bits, idxd->reg_base + IDXD_GENCFG_OFFSET); +} + extern struct bus_type dsa_bus_type; extern bool support_enqcmd; diff --git a/drivers/dma/idxd/init.c b/drivers/dma/idxd/init.c index 1aa823974cda..0eb1c827a215 100644 --- a/drivers/dma/idxd/init.c +++ b/drivers/dma/idxd/init.c @@ -550,14 +550,59 @@ static struct idxd_device *idxd_alloc(struct pci_dev *pdev, struct idxd_driver_d static int idxd_enable_system_pasid(struct idxd_device *idxd) { - return -EOPNOTSUPP; + struct pci_dev *pdev = idxd->pdev; + struct device *dev = &pdev->dev; + struct iommu_domain *domain; + ioasid_t pasid; + int ret; + + /* + * Attach a global PASID to the DMA domain so that we can use ENQCMDS + * to submit work on buffers mapped by DMA API. + */ + domain = iommu_get_domain_for_dev(dev); + if (!domain) + return -EPERM; + + pasid = iommu_alloc_global_pasid(dev); + if (pasid == IOMMU_PASID_INVALID) + return -ENOSPC; + + /* + * DMA domain is owned by the driver, it should support all valid + * types such as DMA-FQ, identity, etc. + */ + ret = iommu_attach_device_pasid(domain, dev, pasid); + if (ret) { + dev_err(dev, "failed to attach device pasid %d, domain type %d", + pasid, domain->type); + iommu_free_global_pasid(pasid); + return ret; + } + + /* Since we set user privilege for kernel DMA, enable completion IRQ */ + idxd_set_user_intr(idxd, 1); + idxd->pasid = pasid; + + return ret; } static void idxd_disable_system_pasid(struct idxd_device *idxd) { + struct pci_dev *pdev = idxd->pdev; + struct device *dev = &pdev->dev; + struct iommu_domain *domain; + + domain = iommu_get_domain_for_dev(dev); + if (!domain) + return; + + iommu_detach_device_pasid(domain, dev, idxd->pasid); + iommu_free_global_pasid(idxd->pasid); - iommu_sva_unbind_device(idxd->sva); + idxd_set_user_intr(idxd, 0); idxd->sva = NULL; + idxd->pasid = IOMMU_PASID_INVALID; } static int idxd_enable_sva(struct pci_dev *pdev) @@ -600,8 +645,9 @@ static int idxd_probe(struct idxd_device *idxd) } else { set_bit(IDXD_FLAG_USER_PASID_ENABLED, &idxd->flags); - if (idxd_enable_system_pasid(idxd)) - dev_warn(dev, "No in-kernel DMA with PASID.\n"); + rc = idxd_enable_system_pasid(idxd); + if (rc) + dev_warn(dev, "No in-kernel DMA with PASID. %d\n", rc); else set_bit(IDXD_FLAG_PASID_ENABLED, &idxd->flags); } diff --git a/drivers/dma/idxd/sysfs.c b/drivers/dma/idxd/sysfs.c index 293739ac5596..63f6966c51aa 100644 --- a/drivers/dma/idxd/sysfs.c +++ b/drivers/dma/idxd/sysfs.c @@ -948,13 +948,6 @@ static ssize_t wq_name_store(struct device *dev, if (strlen(buf) > WQ_NAME_SIZE || strlen(buf) == 0) return -EINVAL; - /* - * This is temporarily placed here until we have SVM support for - * dmaengine. - */ - if (wq->type == IDXD_WQT_KERNEL && device_pasid_enabled(wq->idxd)) - return -EOPNOTSUPP; - input = kstrndup(buf, count, GFP_KERNEL); if (!input) return -ENOMEM; -- cgit v1.2.3 From a48ce36e2786f8bb33e9f42ea2d0c23ea4af3e0d Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:48:02 +0800 Subject: iommu: Prevent RESV_DIRECT devices from blocking domains The IOMMU_RESV_DIRECT flag indicates that a memory region must be mapped 1:1 at all times. This means that the region must always be accessible to the device, even if the device is attached to a blocking domain. This is equal to saying that IOMMU_RESV_DIRECT flag prevents devices from being attached to blocking domains. This also implies that devices that implement RESV_DIRECT regions will be prevented from being assigned to user space since taking the DMA ownership immediately switches to a blocking domain. The rule of preventing devices with the IOMMU_RESV_DIRECT regions from being assigned to user space has existed in the Intel IOMMU driver for a long time. Now, this rule is being lifted up to a general core rule, as other architectures like AMD and ARM also have RMRR-like reserved regions. This has been discussed in the community mailing list and refer to below link for more details. Other places using unmanaged domains for kernel DMA must follow the iommu_get_resv_regions() and setup IOMMU_RESV_DIRECT - we do not restrict them in the core code. Cc: Robin Murphy Cc: Alex Williamson Cc: Kevin Tian Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/linux-iommu/BN9PR11MB5276E84229B5BD952D78E9598C639@BN9PR11MB5276.namprd11.prod.outlook.com Signed-off-by: Lu Baolu Reviewed-by: Jason Gunthorpe Acked-by: Joerg Roedel Link: https://lore.kernel.org/r/20230724060352.113458-2-baolu.lu@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 37 +++++++++++++++++++++++++++---------- include/linux/iommu.h | 2 ++ 2 files changed, 29 insertions(+), 10 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index d199d144460c..e385a99e25e1 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -960,14 +960,12 @@ 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 */ @@ -975,13 +973,17 @@ 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; @@ -2122,6 +2124,21 @@ 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; diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 6b821cb715d5..54bae452975f 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -411,6 +411,7 @@ struct iommu_fault_param { * @priv: IOMMU Driver private data * @max_pasids: number of PASIDs this device can consume * @attach_deferred: the dma domain attachment is deferred + * @require_direct: device requires IOMMU_RESV_DIRECT regions * * TODO: migrate other per device data pointers under iommu_dev_data, e.g. * struct iommu_group *iommu_group; @@ -424,6 +425,7 @@ struct dev_iommu { void *priv; u32 max_pasids; u32 attach_deferred:1; + u32 require_direct:1; }; int iommu_device_register(struct iommu_device *iommu, -- cgit v1.2.3 From d3aedf94f480971f7ffe88d337ba72e6f7d32497 Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Wed, 9 Aug 2023 20:48:03 +0800 Subject: iommu/vt-d: Remove rmrr check in domain attaching device path The core code now prevents devices with RMRR regions from being assigned to user space. There is no need to check for this condition in individual drivers. Remove it to avoid duplicate code. Signed-off-by: Lu Baolu Reviewed-by: Jason Gunthorpe Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20230724060352.113458-3-baolu.lu@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 58 --------------------------------------------- 1 file changed, 58 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 4b5ed58ee422..9369a8b02bd6 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -2487,30 +2487,6 @@ static int dmar_domain_attach_device(struct dmar_domain *domain, return 0; } -static bool device_has_rmrr(struct device *dev) -{ - struct dmar_rmrr_unit *rmrr; - struct device *tmp; - int i; - - rcu_read_lock(); - for_each_rmrr_units(rmrr) { - /* - * Return TRUE if this RMRR contains the device that - * is passed in. - */ - for_each_active_dev_scope(rmrr->devices, - rmrr->devices_cnt, i, tmp) - if (tmp == dev || - is_downstream_to_pci_bridge(dev, tmp)) { - rcu_read_unlock(); - return true; - } - } - rcu_read_unlock(); - return false; -} - /** * device_rmrr_is_relaxable - Test whether the RMRR of this device * is relaxable (ie. is allowed to be not enforced under some conditions) @@ -2540,34 +2516,6 @@ static bool device_rmrr_is_relaxable(struct device *dev) return false; } -/* - * There are a couple cases where we need to restrict the functionality of - * devices associated with RMRRs. The first is when evaluating a device for - * identity mapping because problems exist when devices are moved in and out - * of domains and their respective RMRR information is lost. This means that - * a device with associated RMRRs will never be in a "passthrough" domain. - * The second is use of the device through the IOMMU API. This interface - * expects to have full control of the IOVA space for the device. We cannot - * satisfy both the requirement that RMRR access is maintained and have an - * unencumbered IOVA space. We also have no ability to quiesce the device's - * use of the RMRR space or even inform the IOMMU API user of the restriction. - * We therefore prevent devices associated with an RMRR from participating in - * the IOMMU API, which eliminates them from device assignment. - * - * In both cases, devices which have relaxable RMRRs are not concerned by this - * restriction. See device_rmrr_is_relaxable comment. - */ -static bool device_is_rmrr_locked(struct device *dev) -{ - if (!device_has_rmrr(dev)) - return false; - - if (device_rmrr_is_relaxable(dev)) - return false; - - return true; -} - /* * Return the required default domain type for a specific device. * @@ -4180,12 +4128,6 @@ static int intel_iommu_attach_device(struct iommu_domain *domain, struct device_domain_info *info = dev_iommu_priv_get(dev); int ret; - if (domain->type == IOMMU_DOMAIN_UNMANAGED && - device_is_rmrr_locked(dev)) { - dev_warn(dev, "Device is ineligible for IOMMU domain attach due to platform RMRR requirement. Contact your platform vendor.\n"); - return -EPERM; - } - if (info->domain) device_block_translation(dev); -- cgit v1.2.3 From 8a3b8e63f8371c1247b7aa24ff9c5312f1a6948b Mon Sep 17 00:00:00 2001 From: Yanfei Xu Date: Wed, 9 Aug 2023 20:48:04 +0800 Subject: iommu/vt-d: Fix to flush cache of PASID directory table Even the PCI devices don't support pasid capability, PASID table is mandatory for a PCI device in scalable mode. However flushing cache of pasid directory table for these devices are not taken after pasid table is allocated as the "size" of table is zero. Fix it by calculating the size by page order. Found this when reading the code, no real problem encountered for now. Fixes: 194b3348bdbb ("iommu/vt-d: Fix PASID directory pointer coherency") Suggested-by: Lu Baolu Signed-off-by: Yanfei Xu Link: https://lore.kernel.org/r/20230616081045.721873-1-yanfei.xu@intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/intel/pasid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index 23dca3bc319d..8f92b92f3d2a 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -129,7 +129,7 @@ int intel_pasid_alloc_table(struct device *dev) info->pasid_table = pasid_table; if (!ecap_coherent(info->iommu->ecap)) - clflush_cache_range(pasid_table->table, size); + clflush_cache_range(pasid_table->table, (1 << order) * PAGE_SIZE); return 0; } -- cgit v1.2.3 From fb5f50a43d9fd44fd6bc4f4dbcf9d3ec5b556558 Mon Sep 17 00:00:00 2001 From: Yanfei Xu Date: Wed, 9 Aug 2023 20:48:05 +0800 Subject: iommu/vt-d: Fix to convert mm pfn to dma pfn For the case that VT-d page is smaller than mm page, converting dma pfn should be handled in two cases which are for start pfn and for end pfn. Currently the calculation of end dma pfn is incorrect and the result is less than real page frame number which is causing the mapping of iova always misses some page frames. Rename the mm_to_dma_pfn() to mm_to_dma_pfn_start() and add a new helper for converting end dma pfn named mm_to_dma_pfn_end(). Signed-off-by: Yanfei Xu Link: https://lore.kernel.org/r/20230625082046.979742-1-yanfei.xu@intel.com Signed-off-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/intel/iommu.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 9369a8b02bd6..dbf9270e565e 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -113,13 +113,17 @@ static inline unsigned long lvl_to_nr_pages(unsigned int lvl) /* VT-d pages must always be _smaller_ than MM pages. Otherwise things are never going to work. */ -static inline unsigned long mm_to_dma_pfn(unsigned long mm_pfn) +static inline unsigned long mm_to_dma_pfn_start(unsigned long mm_pfn) { return mm_pfn << (PAGE_SHIFT - VTD_PAGE_SHIFT); } +static inline unsigned long mm_to_dma_pfn_end(unsigned long mm_pfn) +{ + return ((mm_pfn + 1) << (PAGE_SHIFT - VTD_PAGE_SHIFT)) - 1; +} static inline unsigned long page_to_dma_pfn(struct page *pg) { - return mm_to_dma_pfn(page_to_pfn(pg)); + return mm_to_dma_pfn_start(page_to_pfn(pg)); } static inline unsigned long virt_to_dma_pfn(void *p) { @@ -2403,8 +2407,8 @@ static int __init si_domain_init(int hw) for_each_mem_pfn_range(i, nid, &start_pfn, &end_pfn, NULL) { ret = iommu_domain_identity_map(si_domain, - mm_to_dma_pfn(start_pfn), - mm_to_dma_pfn(end_pfn)); + mm_to_dma_pfn_start(start_pfn), + mm_to_dma_pfn_end(end_pfn)); if (ret) return ret; } @@ -2425,8 +2429,8 @@ static int __init si_domain_init(int hw) continue; ret = iommu_domain_identity_map(si_domain, - mm_to_dma_pfn(start >> PAGE_SHIFT), - mm_to_dma_pfn(end >> PAGE_SHIFT)); + mm_to_dma_pfn_start(start >> PAGE_SHIFT), + mm_to_dma_pfn_end(end >> PAGE_SHIFT)); if (ret) return ret; } @@ -3549,8 +3553,8 @@ static int intel_iommu_memory_notifier(struct notifier_block *nb, unsigned long val, void *v) { struct memory_notify *mhp = v; - unsigned long start_vpfn = mm_to_dma_pfn(mhp->start_pfn); - unsigned long last_vpfn = mm_to_dma_pfn(mhp->start_pfn + + unsigned long start_vpfn = mm_to_dma_pfn_start(mhp->start_pfn); + unsigned long last_vpfn = mm_to_dma_pfn_end(mhp->start_pfn + mhp->nr_pages - 1); switch (val) { @@ -4254,7 +4258,7 @@ static void intel_iommu_tlb_sync(struct iommu_domain *domain, unsigned long i; nrpages = aligned_nrpages(gather->start, size); - start_pfn = mm_to_dma_pfn(iova_pfn); + start_pfn = mm_to_dma_pfn_start(iova_pfn); xa_for_each(&dmar_domain->iommu_array, i, info) iommu_flush_iotlb_psi(info->iommu, dmar_domain, -- cgit v1.2.3 From ec2ff4d8160f5d6fdf1b5be7a47e3ff29563b0ba Mon Sep 17 00:00:00 2001 From: Konrad Dybcio Date: Thu, 10 Aug 2023 19:45:04 +0200 Subject: iommu/arm-smmu-qcom: Sort the compatible list alphabetically It got broken at some point, fix it up. Reviewed-by: Dmitry Baryshkov Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230810-topic-lost_smmu_compats-v1-1-64a0d8749404@linaro.org Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c index c71afda79d64..3800ab478216 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c @@ -251,10 +251,10 @@ static const struct of_device_id qcom_smmu_client_of_match[] __maybe_unused = { { .compatible = "qcom,sc7280-mss-pil" }, { .compatible = "qcom,sc8180x-mdss" }, { .compatible = "qcom,sc8280xp-mdss" }, - { .compatible = "qcom,sm8150-mdss" }, - { .compatible = "qcom,sm8250-mdss" }, { .compatible = "qcom,sdm845-mdss" }, { .compatible = "qcom,sdm845-mss-pil" }, + { .compatible = "qcom,sm8150-mdss" }, + { .compatible = "qcom,sm8250-mdss" }, { } }; -- cgit v1.2.3 From 6ebaa77ce483effb3e21a22bf534517e12af8c39 Mon Sep 17 00:00:00 2001 From: Konrad Dybcio Date: Thu, 10 Aug 2023 19:45:05 +0200 Subject: iommu/arm-smmu-qcom: Add SM6375 DPU compatible Add the SM6375 DPU compatible to clients compatible list, as it also needs the workarounds. Acked-by: Dmitry Baryshkov Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230810-topic-lost_smmu_compats-v1-2-64a0d8749404@linaro.org Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c index 3800ab478216..cc574928c707 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c @@ -253,6 +253,7 @@ static const struct of_device_id qcom_smmu_client_of_match[] __maybe_unused = { { .compatible = "qcom,sc8280xp-mdss" }, { .compatible = "qcom,sdm845-mdss" }, { .compatible = "qcom,sdm845-mss-pil" }, + { .compatible = "qcom,sm6375-mdss" }, { .compatible = "qcom,sm8150-mdss" }, { .compatible = "qcom,sm8250-mdss" }, { } -- cgit v1.2.3 From 7e85676a4523c096ec02ff1abcd1a6dd52604002 Mon Sep 17 00:00:00 2001 From: Konrad Dybcio Date: Thu, 10 Aug 2023 19:45:06 +0200 Subject: iommu/arm-smmu-qcom: Add SM6350 DPU compatible Add the SM6350 DPU compatible to clients compatible list, as it also needs the workarounds. Signed-off-by: Konrad Dybcio Acked-by: Dmitry Baryshkov Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230810-topic-lost_smmu_compats-v1-3-64a0d8749404@linaro.org Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c index cc574928c707..bdeb587552c0 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c @@ -253,6 +253,7 @@ static const struct of_device_id qcom_smmu_client_of_match[] __maybe_unused = { { .compatible = "qcom,sc8280xp-mdss" }, { .compatible = "qcom,sdm845-mdss" }, { .compatible = "qcom,sdm845-mss-pil" }, + { .compatible = "qcom,sm6350-mdss" }, { .compatible = "qcom,sm6375-mdss" }, { .compatible = "qcom,sm8150-mdss" }, { .compatible = "qcom,sm8250-mdss" }, -- cgit v1.2.3 From 757d591d965f5fc1cc9888ab480988b6ec8e58fa Mon Sep 17 00:00:00 2001 From: Konrad Dybcio Date: Thu, 10 Aug 2023 19:45:07 +0200 Subject: iommu/arm-smmu-qcom: Add SM6375 SMMUv2 SM6375 uses a qcom,smmu-v2-style SMMU just for Adreno and friends. Add a compatible for it. Signed-off-by: Konrad Dybcio Link: https://lore.kernel.org/r/20230810-topic-lost_smmu_compats-v1-4-64a0d8749404@linaro.org Signed-off-by: Will Deacon --- drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c index bdeb587552c0..7f52ac67495f 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c @@ -530,6 +530,7 @@ static const struct of_device_id __maybe_unused qcom_smmu_impl_of_match[] = { { .compatible = "qcom,sm6125-smmu-500", .data = &qcom_smmu_500_impl0_data }, { .compatible = "qcom,sm6350-smmu-v2", .data = &qcom_smmu_v2_data }, { .compatible = "qcom,sm6350-smmu-500", .data = &qcom_smmu_500_impl0_data }, + { .compatible = "qcom,sm6375-smmu-v2", .data = &qcom_smmu_v2_data }, { .compatible = "qcom,sm6375-smmu-500", .data = &qcom_smmu_500_impl0_data }, { .compatible = "qcom,sm8150-smmu-500", .data = &qcom_smmu_500_impl0_data }, { .compatible = "qcom,sm8250-smmu-500", .data = &qcom_smmu_500_impl0_data }, -- cgit v1.2.3 From 0652cf98e0f09e9dc0e83e53b055947950b0d894 Mon Sep 17 00:00:00 2001 From: Yue Haibing Date: Mon, 14 Aug 2023 21:55:02 +0800 Subject: iommu/amd: Remove unused declarations Commit aafd8ba0ca74 ("iommu/amd: Implement add_device and remove_device") removed the implementations but left declarations in place. Remove it. Signed-off-by: Yue Haibing Link: https://lore.kernel.org/r/20230814135502.4808-1-yuehaibing@huawei.com Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h index f23ad6043f04..e2857109e966 100644 --- a/drivers/iommu/amd/amd_iommu.h +++ b/drivers/iommu/amd/amd_iommu.h @@ -20,9 +20,6 @@ void amd_iommu_apply_erratum_63(struct amd_iommu *iommu, u16 devid); void amd_iommu_restart_event_logging(struct amd_iommu *iommu); void amd_iommu_restart_ga_log(struct amd_iommu *iommu); void amd_iommu_restart_ppr_log(struct amd_iommu *iommu); -int amd_iommu_init_devices(void); -void amd_iommu_uninit_devices(void); -void amd_iommu_init_notifier(void); void amd_iommu_set_rlookup_table(struct amd_iommu *iommu, u16 devid); #ifdef CONFIG_AMD_IOMMU_DEBUGFS -- cgit v1.2.3 From d477f603f37197474437091bf66badddcf76f4c7 Mon Sep 17 00:00:00 2001 From: Rob Herring Date: Fri, 14 Jul 2023 11:46:39 -0600 Subject: iommu: Explicitly include correct DT includes The DT of_device.h and of_platform.h date back to the separate of_platform_bus_type before it as merged into the regular platform bus. As part of that merge prepping Arm DT support 13 years ago, they "temporarily" include each other. They also include platform_device.h and of.h. As a result, there's a pretty much random mix of those include files used throughout the tree. In order to detangle these headers and replace the implicit includes with struct declarations, users need to explicitly include the correct includes. Signed-off-by: Rob Herring Acked-by: Robin Murphy Link: https://lore.kernel.org/r/20230714174640.4058404-1-robh@kernel.org Signed-off-by: Joerg Roedel --- drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c | 2 +- drivers/iommu/arm/arm-smmu/arm-smmu.c | 1 - drivers/iommu/arm/arm-smmu/qcom_iommu.c | 3 +-- drivers/iommu/ipmmu-vmsa.c | 1 - drivers/iommu/sprd-iommu.c | 1 + drivers/iommu/tegra-smmu.c | 2 +- drivers/iommu/virtio-iommu.c | 2 +- 7 files changed, 5 insertions(+), 7 deletions(-) (limited to 'drivers') diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c index b5b14108e086..bb89d49adf8d 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom-debug.c @@ -3,7 +3,7 @@ * Copyright (c) 2022 Qualcomm Innovation Center, Inc. All rights reserved. */ -#include +#include #include #include diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu.c b/drivers/iommu/arm/arm-smmu/arm-smmu.c index a86acd76c1df..d6d1a2a55cc0 100644 --- a/drivers/iommu/arm/arm-smmu/arm-smmu.c +++ b/drivers/iommu/arm/arm-smmu/arm-smmu.c @@ -29,7 +29,6 @@ #include #include #include -#include #include #include #include diff --git a/drivers/iommu/arm/arm-smmu/qcom_iommu.c b/drivers/iommu/arm/arm-smmu/qcom_iommu.c index a503ed758ec3..cc3f68a3516c 100644 --- a/drivers/iommu/arm/arm-smmu/qcom_iommu.c +++ b/drivers/iommu/arm/arm-smmu/qcom_iommu.c @@ -22,8 +22,7 @@ #include #include #include -#include -#include +#include #include #include #include diff --git a/drivers/iommu/ipmmu-vmsa.c b/drivers/iommu/ipmmu-vmsa.c index 9f64c5c9f5b9..0aeedd3e1494 100644 --- a/drivers/iommu/ipmmu-vmsa.c +++ b/drivers/iommu/ipmmu-vmsa.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include #include diff --git a/drivers/iommu/sprd-iommu.c b/drivers/iommu/sprd-iommu.c index 39e34fdeccda..51144c232474 100644 --- a/drivers/iommu/sprd-iommu.c +++ b/drivers/iommu/sprd-iommu.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include diff --git a/drivers/iommu/tegra-smmu.c b/drivers/iommu/tegra-smmu.c index 1cbf063ccf14..e445f80d0226 100644 --- a/drivers/iommu/tegra-smmu.c +++ b/drivers/iommu/tegra-smmu.c @@ -9,7 +9,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/drivers/iommu/virtio-iommu.c b/drivers/iommu/virtio-iommu.c index 3551ed057774..17dcd826f5c2 100644 --- a/drivers/iommu/virtio-iommu.c +++ b/drivers/iommu/virtio-iommu.c @@ -13,7 +13,7 @@ #include #include #include -#include +#include #include #include #include -- cgit v1.2.3