On 14/12/17 11:03, Jean-Philippe Brucker wrote:
Kasan reports a double free when finalise_stage_fn fails: the io_pgtable
ops are freed by arm_smmu_domain_finalise and then again by
arm_smmu_domain_free. Prevent this by leaving pgtbl_ops empty on failure.
It might be a tiny bit more consistent to simply clear the domain ops
again upon failure (cf. how arm_smmu_attach_dev() treats
smmu_domain->smmu), but I'm sure nobody really cares that much:
Reviewed-by: Robin Murphy <robin.mur...@arm.com>
Fixes: 48ec83bcbcf5 ("iommu/arm-smmu: Add initial driver support for ARM SMMUv3
devices")
Signed-off-by: Jean-Philippe Brucker <jean-philippe.bruc...@arm.com>
---
drivers/iommu/arm-smmu-v3.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c
index f122071688fd..db4281d0e269 100644
--- a/drivers/iommu/arm-smmu-v3.c
+++ b/drivers/iommu/arm-smmu-v3.c
@@ -1698,13 +1698,15 @@ static int arm_smmu_domain_finalise(struct iommu_domain
*domain)
domain->pgsize_bitmap = pgtbl_cfg.pgsize_bitmap;
domain->geometry.aperture_end = (1UL << ias) - 1;
domain->geometry.force_aperture = true;
- smmu_domain->pgtbl_ops = pgtbl_ops;
ret = finalise_stage_fn(smmu_domain, &pgtbl_cfg);
- if (ret < 0)
+ if (ret < 0) {
free_io_pgtable_ops(pgtbl_ops);
+ return ret;
+ }
- return ret;
+ smmu_domain->pgtbl_ops = pgtbl_ops;
+ return 0;
}
static __le64 *arm_smmu_get_step_for_sid(struct arm_smmu_device *smmu, u32 sid)
_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu