]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
x86/amd-iommu: Change alloc_pte to support 64 bit address space
authorJoerg Roedel <joerg.roedel@amd.com>
Wed, 2 Sep 2009 14:48:40 +0000 (16:48 +0200)
committerJoerg Roedel <joerg.roedel@amd.com>
Thu, 3 Sep 2009 14:03:46 +0000 (16:03 +0200)
This patch changes the alloc_pte function to be able to map
pages into the whole 64 bit address space supported by AMD
IOMMU hardware from the old limit of 2**39 bytes.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
arch/x86/kernel/amd_iommu.c

index fc97b51f0287f3cfc8bd187eee00eb45ca064ed1..3be2b61fc31f842d4c9694805c1c3c45cb956995 100644 (file)
@@ -55,7 +55,7 @@ struct iommu_cmd {
 static int dma_ops_unity_map(struct dma_ops_domain *dma_dom,
                             struct unity_map_entry *e);
 static struct dma_ops_domain *find_protection_domain(u16 devid);
-static u64* alloc_pte(struct protection_domain *dom,
+static u64 *alloc_pte(struct protection_domain *domain,
                      unsigned long address, u64
                      **pte_page, gfp_t gfp);
 static void dma_ops_reserve_addresses(struct dma_ops_domain *dom,
@@ -1351,39 +1351,35 @@ static bool increase_address_space(struct protection_domain *domain,
        return true;
 }
 
-/*
- * If the pte_page is not yet allocated this function is called
- */
-static u64* alloc_pte(struct protection_domain *dom,
+static u64 *alloc_pte(struct protection_domain *domain,
                      unsigned long address, u64 **pte_page, gfp_t gfp)
 {
        u64 *pte, *page;
+       int level;
 
-       pte = &dom->pt_root[IOMMU_PTE_L2_INDEX(address)];
+       while (address > PM_LEVEL_SIZE(domain->mode))
+               increase_address_space(domain, gfp);
 
-       if (!IOMMU_PTE_PRESENT(*pte)) {
-               page = (u64 *)get_zeroed_page(gfp);
-               if (!page)
-                       return NULL;
-               *pte = IOMMU_L2_PDE(virt_to_phys(page));
-       }
+       level =  domain->mode - 1;
+       pte   = &domain->pt_root[PM_LEVEL_INDEX(level, address)];
 
-       pte = IOMMU_PTE_PAGE(*pte);
-       pte = &pte[IOMMU_PTE_L1_INDEX(address)];
+       while (level > 0) {
+               if (!IOMMU_PTE_PRESENT(*pte)) {
+                       page = (u64 *)get_zeroed_page(gfp);
+                       if (!page)
+                               return NULL;
+                       *pte = PM_LEVEL_PDE(level, virt_to_phys(page));
+               }
 
-       if (!IOMMU_PTE_PRESENT(*pte)) {
-               page = (u64 *)get_zeroed_page(gfp);
-               if (!page)
-                       return NULL;
-               *pte = IOMMU_L1_PDE(virt_to_phys(page));
-       }
+               level -= 1;
 
-       pte = IOMMU_PTE_PAGE(*pte);
+               pte = IOMMU_PTE_PAGE(*pte);
 
-       if (pte_page)
-               *pte_page = pte;
+               if (pte_page && level == 0)
+                       *pte_page = pte;
 
-       pte = &pte[IOMMU_PTE_L0_INDEX(address)];
+               pte = &pte[PM_LEVEL_INDEX(level, address)];
+       }
 
        return pte;
 }