On 2020-12-16 10:36, Yong Wu wrote:
Currently gather->end is "unsigned long" which may be overflow in arch32 in the corner case: 0xfff00000 + 0x100000(iova + size). Although it doesn't affect the size(end - start), it affects the checking "gather->end < end"
This won't help the same situation at the top of a 64-bit address space, though, and now that we have TTBR1 support for AArch64 format that is definitely a thing. Better to just encode the end address as the actual end address, i.e. iova + size - 1. We don't lose anything other than the ability to encode zero-sized invalidations that don't make sense anyway.
Robin.
Fixes: a7d20dc19d9e ("iommu: Introduce struct iommu_iotlb_gather for batching TLB flushes") Signed-off-by: Yong Wu <[email protected]> --- include/linux/iommu.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 794d4085edd3..6e907a95d981 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -178,7 +178,7 @@ enum iommu_dev_features { */ struct iommu_iotlb_gather { unsigned long start; - unsigned long end; + unsigned long long end; size_t pgsize; };@@ -537,7 +537,8 @@ static inline void iommu_iotlb_gather_add_page(struct iommu_domain *domain,struct iommu_iotlb_gather *gather, unsigned long iova, size_t size) { - unsigned long start = iova, end = start + size; + unsigned long start = iova; + unsigned long long end = start + size;/** If the new page is disjoint from the current range or is mapped at
