iommu/amd: Put list_add/del(dev_data) back under the domain->lock

The list domain->dev_list is protected by the domain->lock spinlock.
Any iteration, addition or removal must be under the lock.

Move the list_del() up into the critical section. pdom_is_sva_capable(),
and destroy_gcr3_table() do not interact with the list element.

Wrap the list_add() in a lock, it would make more sense if this was under
the same critical section as adjusting the refcounts earlier, but that
requires more complications.

Fixes: d6b47de ("iommu/amd: Reduce domain lock scope in attach device path")
Signed-off-by: Jason Gunthorpe <[email protected]>
Reviewed-by: Vasant Hegde <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Joerg Roedel <[email protected]>

Author: jgunthorpe

drivers/iommu/amd/iommu.c

@@ -2073,6 +2073,7 @@ static int attach_device(struct device *dev,
 	struct iommu_dev_data *dev_data = dev_iommu_priv_get(dev);
 	struct amd_iommu *iommu = get_amd_iommu_from_dev_data(dev_data);
 	struct pci_dev *pdev;
+	unsigned long flags;
 	int ret = 0;
 
 	mutex_lock(&dev_data->mutex);
@@ -2113,7 +2114,9 @@ static int attach_device(struct device *dev,
 
 	/* Update data structures */
 	dev_data->domain = domain;
+	spin_lock_irqsave(&domain->lock, flags);
 	list_add(&dev_data->list, &domain->dev_list);
+	spin_unlock_irqrestore(&domain->lock, flags);
 
 	/* Update device table */
 	dev_update_dte(dev_data, true);
@@ -2160,6 +2163,7 @@ static void detach_device(struct device *dev)
 	/* Flush IOTLB and wait for the flushes to finish */
 	spin_lock_irqsave(&domain->lock, flags);
 	amd_iommu_domain_flush_all(domain);
+	list_del(&dev_data->list);
 	spin_unlock_irqrestore(&domain->lock, flags);
 
 	/* Clear GCR3 table */
@@ -2168,7 +2172,6 @@ static void detach_device(struct device *dev)
 
 	/* Update data structures */
 	dev_data->domain = NULL;
-	list_del(&dev_data->list);
 
 	/* decrease reference counters - needs to happen after the flushes */
 	pdom_detach_iommu(iommu, domain);