iommu: Remove iommu_group_do_dma_first_attach() from iommu_group_add_device()

This function is only used to construct the groups, it should not be
operating the iommu driver.

External callers in VFIO and POWER do not have any iommu drivers on the
devices so group->domain will be NULL.

The only internal caller is from iommu_probe_device() which already calls
iommu_group_do_dma_first_attach(), meaning we are calling it twice in the
only case it matters.

Since iommu_probe_device() is the logical place to sort out the group's
domain, remove the call from iommu_group_add_device().

Reviewed-by: Kevin Tian <kevin.tian@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/7-v5-1b99ae392328+44574-iommu_err_unwind_jgg@nvidia.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>

Author: jgunthorpe

drivers/iommu/iommu.c

@@ -1080,25 +1080,13 @@ int iommu_group_add_device(struct iommu_group *group, struct device *dev)
 
 	mutex_lock(&group->mutex);
 	list_add_tail(&device->list, &group->devices);
-	if (group->domain)
-		ret = iommu_group_do_dma_first_attach(dev, group->domain);
 	mutex_unlock(&group->mutex);
-	if (ret)
-		goto err_put_group;
-
 	trace_add_device_to_group(group->id, dev);
 
 	dev_info(dev, "Adding to iommu group %d\n", group->id);
 
 	return 0;
 
-err_put_group:
-	mutex_lock(&group->mutex);
-	list_del(&device->list);
-	mutex_unlock(&group->mutex);
-	dev->iommu_group = NULL;
-	kobject_put(group->devices_kobj);
-	sysfs_remove_link(group->devices_kobj, device->name);
 err_free_name:
 	kfree(device->name);
 err_remove_link: