iommu/vt-d: Preset A/D bits for user space DMA usage
authorLu Baolu <baolu.lu@linux.intel.com>
Wed, 18 Aug 2021 13:48:48 +0000 (21:48 +0800)
committerJoerg Roedel <jroedel@suse.de>
Thu, 19 Aug 2021 08:41:08 +0000 (10:41 +0200)
We preset the access and dirty bits for IOVA over first level usage only
for the kernel DMA (i.e., when domain type is IOMMU_DOMAIN_DMA). We should
also preset the FL A/D for user space DMA usage. The idea is that even the
user space A/D bit memory write is unnecessary. We should avoid it to
minimize the overhead.

Suggested-by: Sanjay Kumar <sanjay.k.kumar@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/20210720013856.4143880-1-baolu.lu@linux.intel.com
Link: https://lore.kernel.org/r/20210818134852.1847070-6-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
drivers/iommu/intel/iommu.c

index 8c9a9ed..8d4d49e 100644 (file)
@@ -2334,13 +2334,9 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
        attr = prot & (DMA_PTE_READ | DMA_PTE_WRITE | DMA_PTE_SNP);
        attr |= DMA_FL_PTE_PRESENT;
        if (domain_use_first_level(domain)) {
-               attr |= DMA_FL_PTE_XD | DMA_FL_PTE_US;
-
-               if (iommu_is_dma_domain(&domain->domain)) {
-                       attr |= DMA_FL_PTE_ACCESS;
-                       if (prot & DMA_PTE_WRITE)
-                               attr |= DMA_FL_PTE_DIRTY;
-               }
+               attr |= DMA_FL_PTE_XD | DMA_FL_PTE_US | DMA_FL_PTE_ACCESS;
+               if (prot & DMA_PTE_WRITE)
+                       attr |= DMA_FL_PTE_DIRTY;
        }
 
        pteval = ((phys_addr_t)phys_pfn << VTD_PAGE_SHIFT) | attr;