On 08/08/2017 04:21 AM, Will Deacon wrote:
[+ Joerg]

On Tue, Aug 08, 2017 at 11:37:40AM +0100, Robin Murphy wrote:
On 08/08/17 11:26, Artem Savkov wrote:
Commit c54451a "iommu/arm-smmu: Fix the error path in arm_smmu_add_device"
removed fwspec assignment in legacy_binding path as redundant which is
wrong. It needs to be updated after fwspec initialisation in
arm_smmu_register_legacy_master() as it is dereferenced later. Without
this there is a NULL-pointer dereference panic during boot on some hosts.

Reviewed-by: Robin Murphy <robin.mur...@arm.com>

Thanks for fixing it up, and sorry for failing to document the
unfortunately subtle logic in the first place!

Well, I was the one that messed it up:

Acked-by: Will Deacon <will.dea...@arm.com>

Joerg, can you pick this up as a fix for 4.13, please?

I hit the Oops as well. This patch fixes it for me on a Cavium CN88xx system, so:

Acked-by: David Daney <david.da...@cavium.com>

Thanks for working on this.


Signed-off-by: Artem Savkov <asav...@redhat.com>
  drivers/iommu/arm-smmu.c | 7 +++++++
  1 file changed, 7 insertions(+)

diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c
index b97188a..2d80fa8 100644
--- a/drivers/iommu/arm-smmu.c
+++ b/drivers/iommu/arm-smmu.c
@@ -1519,6 +1519,13 @@ static int arm_smmu_add_device(struct device *dev)
if (using_legacy_binding) {
                ret = arm_smmu_register_legacy_master(dev, &smmu);
+               /*
+                * If dev->iommu_fwspec is initally NULL, 
+                * will allocate/initialise a new one. Thus we need to update 
fwspec for
+                * later use.
+                */
+               fwspec = dev->iommu_fwspec;
                if (ret)
                        goto out_free;
        } else if (fwspec && fwspec->ops == &arm_smmu_ops) {

linux-arm-kernel mailing list

iommu mailing list

Reply via email to