iommu: Remove __iommu_group_for_each_dev()
authorJason Gunthorpe <jgg@nvidia.com>
Thu, 11 May 2023 04:42:14 +0000 (01:42 -0300)
committerJoerg Roedel <jroedel@suse.de>
Tue, 23 May 2023 06:15:58 +0000 (08:15 +0200)
The last two users of it are quite trivial, just open code the one line
loop.

Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/16-v5-1b99ae392328+44574-iommu_err_unwind_jgg@nvidia.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
drivers/iommu/iommu.c

index 197d46a..1aaf3eb 100644 (file)
@@ -1110,20 +1110,6 @@ void iommu_group_remove_device(struct device *dev)
 }
 EXPORT_SYMBOL_GPL(iommu_group_remove_device);
 
-static int __iommu_group_for_each_dev(struct iommu_group *group, void *data,
-                                     int (*fn)(struct device *, void *))
-{
-       struct group_device *device;
-       int ret = 0;
-
-       for_each_group_device(group, device) {
-               ret = fn(device->dev, data);
-               if (ret)
-                       break;
-       }
-       return ret;
-}
-
 /**
  * iommu_group_for_each_dev - iterate over each device in the group
  * @group: the group
@@ -1138,10 +1124,15 @@ static int __iommu_group_for_each_dev(struct iommu_group *group, void *data,
 int iommu_group_for_each_dev(struct iommu_group *group, void *data,
                             int (*fn)(struct device *, void *))
 {
-       int ret;
+       struct group_device *device;
+       int ret = 0;
 
        mutex_lock(&group->mutex);
-       ret = __iommu_group_for_each_dev(group, data, fn);
+       for_each_group_device(group, device) {
+               ret = fn(device->dev, data);
+               if (ret)
+                       break;
+       }
        mutex_unlock(&group->mutex);
 
        return ret;
@@ -1791,20 +1782,12 @@ static int iommu_get_default_domain_type(struct iommu_group *group,
        return best_type;
 }
 
-static int iommu_group_do_probe_finalize(struct device *dev, void *data)
+static void iommu_group_do_probe_finalize(struct device *dev)
 {
        const struct iommu_ops *ops = dev_iommu_ops(dev);
 
        if (ops->probe_finalize)
                ops->probe_finalize(dev);
-
-       return 0;
-}
-
-static void __iommu_group_dma_finalize(struct iommu_group *group)
-{
-       __iommu_group_for_each_dev(group, group->default_domain,
-                                  iommu_group_do_probe_finalize);
 }
 
 int bus_iommu_probe(const struct bus_type *bus)
@@ -1823,6 +1806,8 @@ int bus_iommu_probe(const struct bus_type *bus)
                return ret;
 
        list_for_each_entry_safe(group, next, &group_list, entry) {
+               struct group_device *gdev;
+
                mutex_lock(&group->mutex);
 
                /* Remove item from the list */
@@ -1834,7 +1819,15 @@ int bus_iommu_probe(const struct bus_type *bus)
                        return ret;
                }
                mutex_unlock(&group->mutex);
-               __iommu_group_dma_finalize(group);
+
+               /*
+                * FIXME: Mis-locked because the ops->probe_finalize() call-back
+                * of some IOMMU drivers calls arm_iommu_attach_device() which
+                * in-turn might call back into IOMMU core code, where it tries
+                * to take group->mutex, resulting in a deadlock.
+                */
+               for_each_group_device(group, gdev)
+                       iommu_group_do_probe_finalize(gdev->dev);
        }
 
        return 0;
@@ -2995,8 +2988,12 @@ static ssize_t iommu_group_store_type(struct iommu_group *group,
        mutex_unlock(&group->mutex);
 
        /* Make sure dma_ops is appropriatley set */
-       if (!ret)
-               __iommu_group_dma_finalize(group);
+       if (!ret) {
+               struct group_device *gdev;
+
+               for_each_group_device(group, gdev)
+                       iommu_group_do_probe_finalize(gdev->dev);
+       }
 
        return ret ?: count;
 }