Skip to contentSkip to conversation

iommu: Check the result of iommu_group_get() for NULL

PATCH Applied Jordan Crouse 3 participants 3 messages Dec 4, 2017 – Jan 9, 2018 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iommu: Check the result of iommu_group_get() for NULL

The result of iommu_group_get() was being blindly used in both attach and detach which results in a dereference when trying to work with an unknown device.

---
 drivers/iommu/iommu.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/iommu/iommu.c +5 -0
index 3de5c0b..69fef99 100644
@@ -1303,6 +1303,9 @@ int iommu_attach_device(struct iommu_domain *domain, struct device *dev)
 	int ret;
 
 	group = iommu_group_get(dev);
+	if (!group)
+		return -ENODEV;
+
 	/*
 	 * Lock the group to make sure the device-count doesn't
 	 * change while we are attaching
@@ -1341,6 +1344,8 @@ void iommu_detach_device(struct iommu_domain *domain, struct device *dev)
 	struct iommu_group *group;
 
 	group = iommu_group_get(dev);
+	if (!group)
+		return;
 
 	mutex_lock(&group->mutex);
 	if (iommu_group_device_count(group) != 1) {
Signature · 1 line
1.9.1

A message that is not in the archive (sent off-list or not yet received): 1512408130-24833-1-git-send-email-jcrouse-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org

#2 lore raw
Quoted 8 lines · On Mon, 4 Dec 2017 10:22:10 -0700 Jordan Crouse…

The result of iommu_group_get() was being blindly used in both attach and detach which results in a dereference when trying to work with an unknown device.

---
 drivers/iommu/iommu.c | 5 +++++
 1 file changed, 5 insertions(+)

Applied to v4.16-iommu/misc. Thanks,

Alex

#3 lore raw
In reply to a message not in the archive

Hi Jordan,

Quoted 3 lines · Jordan Crouse

The result of iommu_group_get() was being blindly used in both attach and detach which results in a dereference when trying to work with an unknown device.

I missed this one before it got applied, but FWIW that behaviour was actually somewhat intentional (see 05f80300dc8b). Callers really shouldn't be trying to attach random unknown devices to IOMMU domains anyway, but more crucially, now that iommu_{attach,detach}_group() work everywhere I think the long-term plan is still to remove iommu_{attach,detach}_device() completely.

I have no great objection to this patch myself, but do consider it a heads-up that any callers it fixes are liable to be ridden roughshod over by further API changes in future.

Robin.

Quoted 29 lines
---
  drivers/iommu/iommu.c | 5 +++++
  1 file changed, 5 insertions(+)
drivers/iommu/iommu.c +5 -0
index 3de5c0b..69fef99 100644
@@ -1303,6 +1303,9 @@ int iommu_attach_device(struct iommu_domain *domain, struct device *dev)
  	int ret;
 
  	group = iommu_group_get(dev);
+	if (!group)
+		return -ENODEV;
+
  	/*
  	 * Lock the group to make sure the device-count doesn't
  	 * change while we are attaching
@@ -1341,6 +1344,8 @@ void iommu_detach_device(struct iommu_domain *domain, struct device *dev)
  	struct iommu_group *group;
 
  	group = iommu_group_get(dev);
+	if (!group)
+		return;
 
  	mutex_lock(&group->mutex);
  	if (iommu_group_device_count(group) != 1) {