[PATCH v3] iommu/vt-d: Fix a bug in intel_iommu_iova_to_phys() for huge page

From: Yonghyun Hwang
Date: Wed Feb 26 2020 - 15:30:46 EST


intel_iommu_iova_to_phys() has a bug when it translates an IOVA for a huge
page onto its corresponding physical address. This commit fixes the bug by
accomodating the level of page entry for the IOVA and adds IOVA's lower
address to the physical address.

Cc: <stable@xxxxxxxxxxxxxxx>
Acked-by: Lu Baolu <baolu.lu@xxxxxxxxxxxxxxx>
Reviewed-by: Moritz Fischer <mdf@xxxxxxxxxx>
Signed-off-by: Yonghyun Hwang <yonghyun@xxxxxxxxxx>
---

Changes from v2:
- a new condition is added to check whether the pte is present.

Changes from v1:
- level cannot be 0. So, the condition, "if (level > 1)", is removed, which results in a simple code.
- a macro, BIT_MASK, is used to have a bit mask

---
drivers/iommu/intel-iommu.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index 932267f49f9a..0837e0063973 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -5553,8 +5553,10 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain,
u64 phys = 0;

pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level);
- if (pte)
- phys = dma_pte_addr(pte);
+ if (pte && dma_pte_present(pte))
+ phys = dma_pte_addr(pte) +
+ (iova & (BIT_MASK(level_to_offset_bits(level) +
+ VTD_PAGE_SHIFT) - 1));

return phys;
}
--
2.25.0.265.gbab2e86ba0-goog