Merge tag 'iommu-updates-v4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu

Pull IOMMU updates from Joerg Roedel:

 - big-endian support and preparation for defered probing for the Exynos
   IOMMU driver

 - simplifications in iommu-group id handling

 - support for Mediatek generation one IOMMU hardware

 - conversion of the AMD IOMMU driver to use the generic IOVA allocator.
   This driver now also benefits from the recent scalability
   improvements in the IOVA code.

 - preparations to use generic DMA mapping code in the Rockchip IOMMU
   driver

 - device tree adaption and conversion to use generic page-table code
   for the MSM IOMMU driver

 - an iova_to_phys optimization in the ARM-SMMU driver to greatly
   improve page-table teardown performance with VFIO

 - various other small fixes and conversions

* tag 'iommu-updates-v4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (59 commits)
  iommu/amd: Initialize dma-ops domains with 3-level page-table
  iommu/amd: Update Alias-DTE in update_device_table()
  iommu/vt-d: Return error code in domain_context_mapping_one()
  iommu/amd: Use container_of to get dma_ops_domain
  iommu/amd: Flush iova queue before releasing dma_ops_domain
  iommu/amd: Handle IOMMU_DOMAIN_DMA in ops->domain_free call-back
  iommu/amd: Use dev_data->domain in get_domain()
  iommu/amd: Optimize map_sg and unmap_sg
  iommu/amd: Introduce dir2prot() helper
  iommu/amd: Implement timeout to flush unmap queues
  iommu/amd: Implement flush queue
  iommu/amd: Allow NULL pointer parameter for domain_flush_complete()
  iommu/amd: Set up data structures for flush queue
  iommu/amd: Remove align-parameter from __map_single()
  iommu/amd: Remove other remains of old address allocator
  iommu/amd: Make use of the generic IOVA allocator
  iommu/amd: Remove special mapping code for dma_ops path
  iommu/amd: Pass gfp-flags to iommu_map_page()
  iommu/amd: Implement apply_dm_region call-back
  iommu/amd: Create a list of reserved iova addresses
  ...

20 files changed, 2132 insertions, 1650 deletions

diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index ad0860383cb3..d432ca828472 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -89,8 +89,8 @@ config MSM_IOMMU
 	bool "MSM IOMMU Support"
 	depends on ARM
 	depends on ARCH_MSM8X60 || ARCH_MSM8960 || COMPILE_TEST
-	depends on BROKEN
 	select IOMMU_API
+	select IOMMU_IO_PGTABLE_ARMV7S
 	help
 	  Support for the IOMMUs found on certain Qualcomm SOCs.
 	  These IOMMUs allow virtualization of the address space used by most
@@ -111,6 +111,7 @@ config AMD_IOMMU
 	select PCI_PRI
 	select PCI_PASID
 	select IOMMU_API
+	select IOMMU_IOVA
 	depends on X86_64 && PCI && ACPI
 	---help---
 	  With this option you can enable support for AMD IOMMU hardware in
@@ -343,4 +344,22 @@ config MTK_IOMMU
 
 	  If unsure, say N here.
 
+config MTK_IOMMU_V1
+	bool "MTK IOMMU Version 1 (M4U gen1) Support"
+	depends on ARM
+	depends on ARCH_MEDIATEK || COMPILE_TEST
+	select ARM_DMA_USE_IOMMU
+	select IOMMU_API
+	select MEMORY
+	select MTK_SMI
+	select COMMON_CLK_MT2701_MMSYS
+	select COMMON_CLK_MT2701_IMGSYS
+	select COMMON_CLK_MT2701_VDECSYS
+	help
+	  Support for the M4U on certain Mediatek SoCs. M4U generation 1 HW is
+	  Multimedia Memory Managememt Unit. This option enables remapping of
+	  DMA memory accesses for the multimedia subsystem.
+
+	  if unsure, say N here.
+
 endif # IOMMU_SUPPORT
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index c6edb31bf8c6..195f7b997d8e 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -7,7 +7,7 @@ obj-$(CONFIG_IOMMU_IO_PGTABLE_ARMV7S) += io-pgtable-arm-v7s.o
 obj-$(CONFIG_IOMMU_IO_PGTABLE_LPAE) += io-pgtable-arm.o
 obj-$(CONFIG_IOMMU_IOVA) += iova.o
 obj-$(CONFIG_OF_IOMMU)	+= of_iommu.o
-obj-$(CONFIG_MSM_IOMMU) += msm_iommu.o msm_iommu_dev.o
+obj-$(CONFIG_MSM_IOMMU) += msm_iommu.o
 obj-$(CONFIG_AMD_IOMMU) += amd_iommu.o amd_iommu_init.o
 obj-$(CONFIG_AMD_IOMMU_V2) += amd_iommu_v2.o
 obj-$(CONFIG_ARM_SMMU) += arm-smmu.o
@@ -18,6 +18,7 @@ obj-$(CONFIG_INTEL_IOMMU_SVM) += intel-svm.o
 obj-$(CONFIG_IPMMU_VMSA) += ipmmu-vmsa.o
 obj-$(CONFIG_IRQ_REMAP) += intel_irq_remapping.o irq_remapping.o
 obj-$(CONFIG_MTK_IOMMU) += mtk_iommu.o
+obj-$(CONFIG_MTK_IOMMU_V1) += mtk_iommu_v1.o
 obj-$(CONFIG_OMAP_IOMMU) += omap-iommu.o
 obj-$(CONFIG_OMAP_IOMMU_DEBUG) += omap-iommu-debug.o
 obj-$(CONFIG_ROCKCHIP_IOMMU) += rockchip-iommu.o
diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index 634f636393d5..33c177ba93be 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -21,6 +21,7 @@
 #include <linux/pci.h>
 #include <linux/acpi.h>
 #include <linux/amba/bus.h>
+#include <linux/platform_device.h>
 #include <linux/pci-ats.h>
 #include <linux/bitmap.h>
 #include <linux/slab.h>
@@ -38,6 +39,7 @@
 #include <linux/dma-contiguous.h>
 #include <linux/irqdomain.h>
 #include <linux/percpu.h>
+#include <linux/iova.h>
 #include <asm/irq_remapping.h>
 #include <asm/io_apic.h>
 #include <asm/apic.h>
@@ -56,6 +58,17 @@
 
 #define LOOP_TIMEOUT	100000
 
+/* IO virtual address start page frame number */
+#define IOVA_START_PFN		(1)
+#define IOVA_PFN(addr)		((addr) >> PAGE_SHIFT)
+#define DMA_32BIT_PFN		IOVA_PFN(DMA_BIT_MASK(32))
+
+/* Reserved IOVA ranges */
+#define MSI_RANGE_START		(0xfee00000)
+#define MSI_RANGE_END		(0xfeefffff)
+#define HT_RANGE_START		(0xfd00000000ULL)
+#define HT_RANGE_END		(0xffffffffffULL)
+
 /*
  * This bitmap is used to advertise the page sizes our hardware support
  * to the IOMMU core, which will then use this information to split
@@ -76,6 +89,25 @@ LIST_HEAD(ioapic_map);
 LIST_HEAD(hpet_map);
 LIST_HEAD(acpihid_map);
 
+#define FLUSH_QUEUE_SIZE 256
+
+struct flush_queue_entry {
+	unsigned long iova_pfn;
+	unsigned long pages;
+	struct dma_ops_domain *dma_dom;
+};
+
+struct flush_queue {
+	spinlock_t lock;
+	unsigned next;
+	struct flush_queue_entry *entries;
+};
+
+DEFINE_PER_CPU(struct flush_queue, flush_queue);
+
+static atomic_t queue_timer_on;
+static struct timer_list queue_timer;
+
 /*
  * Domain for untranslated devices - only allocated
  * if iommu=pt passed on kernel cmd line.
@@ -121,44 +153,19 @@ static int protection_domain_init(struct protection_domain *domain);
 static void detach_device(struct device *dev);
 
 /*
- * For dynamic growth the aperture size is split into ranges of 128MB of
- * DMA address space each. This struct represents one such range.
- */
-struct aperture_range {
-
-	spinlock_t bitmap_lock;
-
-	/* address allocation bitmap */
-	unsigned long *bitmap;
-	unsigned long offset;
-	unsigned long next_bit;
-
-	/*
-	 * Array of PTE pages for the aperture. In this array we save all the
-	 * leaf pages of the domain page table used for the aperture. This way
-	 * we don't need to walk the page table to find a specific PTE. We can
-	 * just calculate its address in constant time.
-	 */
-	u64 *pte_pages[64];
-};
-
-/*
  * Data container for a dma_ops specific protection domain
  */
 struct dma_ops_domain {
 	/* generic protection domain information */
 	struct protection_domain domain;
 
-	/* size of the aperture for the mappings */
-	unsigned long aperture_size;
-
-	/* aperture index we start searching for free addresses */
-	u32 __percpu *next_index;
-
-	/* address space relevant data */
-	struct aperture_range *aperture[APERTURE_MAX_RANGES];
+	/* IOVA RB-Tree */
+	struct iova_domain iovad;
 };
 
+static struct iova_domain reserved_iova_ranges;
+static struct lock_class_key reserved_rbtree_key;
+
 /****************************************************************************
  *
  * Helper functions
@@ -224,6 +231,12 @@ static struct protection_domain *to_pdomain(struct iommu_domain *dom)
 	return container_of(dom, struct protection_domain, domain);
 }
 
+static struct dma_ops_domain* to_dma_ops_domain(struct protection_domain *domain)
+{
+	BUG_ON(domain->flags != PD_DMA_OPS_MASK);
+	return container_of(domain, struct dma_ops_domain, domain);
+}
+
 static struct iommu_dev_data *alloc_dev_data(u16 devid)
 {
 	struct iommu_dev_data *dev_data;
@@ -391,43 +404,6 @@ static bool pdev_pri_erratum(struct pci_dev *pdev, u32 erratum)
 }
 
 /*
- * This function actually applies the mapping to the page table of the
- * dma_ops domain.
- */
-static void alloc_unity_mapping(struct dma_ops_domain *dma_dom,
-				struct unity_map_entry *e)
-{
-	u64 addr;
-
-	for (addr = e->address_start; addr < e->address_end;
-	     addr += PAGE_SIZE) {
-		if (addr < dma_dom->aperture_size)
-			__set_bit(addr >> PAGE_SHIFT,
-				  dma_dom->aperture[0]->bitmap);
-	}
-}
-
-/*
- * Inits the unity mappings required for a specific device
- */
-static void init_unity_mappings_for_device(struct device *dev,
-					   struct dma_ops_domain *dma_dom)
-{
-	struct unity_map_entry *e;
-	int devid;
-
-	devid = get_device_id(dev);
-	if (devid < 0)
-		return;
-
-	list_for_each_entry(e, &amd_iommu_unity_map, list) {
-		if (!(devid >= e->devid_start && devid <= e->devid_end))
-			continue;
-		alloc_unity_mapping(dma_dom, e);
-	}
-}
-
-/*
  * This function checks if the driver got a valid device from the caller to
  * avoid dereferencing invalid pointers.
  */
@@ -454,22 +430,12 @@ static bool check_device(struct device *dev)
 
 static void init_iommu_group(struct device *dev)
 {
-	struct dma_ops_domain *dma_domain;
-	struct iommu_domain *domain;
 	struct iommu_group *group;
 
 	group = iommu_group_get_for_dev(dev);
 	if (IS_ERR(group))
 		return;
 
-	domain = iommu_group_default_domain(group);
-	if (!domain)
-		goto out;
-
-	dma_domain = to_pdomain(domain)->priv;
-
-	init_unity_mappings_for_device(dev, dma_domain);
-out:
 	iommu_group_put(group);
 }
 
@@ -1220,7 +1186,7 @@ static void domain_flush_complete(struct protection_domain *domain)
 	int i;
 
 	for (i = 0; i < amd_iommus_present; ++i) {
-		if (!domain->dev_iommu[i])
+		if (domain && !domain->dev_iommu[i])
 			continue;
 
 		/*
@@ -1397,8 +1363,9 @@ static u64 *fetch_pte(struct protection_domain *domain,
 static int iommu_map_page(struct protection_domain *dom,
 			  unsigned long bus_addr,
 			  unsigned long phys_addr,
+			  unsigned long page_size,
 			  int prot,
-			  unsigned long page_size)
+			  gfp_t gfp)
 {
 	u64 __pte, *pte;
 	int i, count;
@@ -1410,7 +1377,7 @@ static int iommu_map_page(struct protection_domain *dom,
 		return -EINVAL;
 
 	count = PAGE_SIZE_PTE_COUNT(page_size);
-	pte   = alloc_pte(dom, bus_addr, page_size, NULL, GFP_KERNEL);
+	pte   = alloc_pte(dom, bus_addr, page_size, NULL, gfp);
 
 	if (!pte)
 		return -ENOMEM;
@@ -1474,320 +1441,37 @@ static unsigned long iommu_unmap_page(struct protection_domain *dom,
 /****************************************************************************
  *
  * The next functions belong to the address allocator for the dma_ops
- * interface functions. They work like the allocators in the other IOMMU
- * drivers. Its basically a bitmap which marks the allocated pages in
- * the aperture. Maybe it could be enhanced in the future to a more
- * efficient allocator.
+ * interface functions.
  *
  ****************************************************************************/
 
-/*
- * The address allocator core functions.
- *
- * called with domain->lock held
- */
 
-/*
- * Used to reserve address ranges in the aperture (e.g. for exclusion
- * ranges.
- */
-static void dma_ops_reserve_addresses(struct dma_ops_domain *dom,
-				      unsigned long start_page,
-				      unsigned int pages)
+static unsigned long dma_ops_alloc_iova(struct device *dev,
+					struct dma_ops_domain *dma_dom,
+					unsigned int pages, u64 dma_mask)
 {
-	unsigned int i, last_page = dom->aperture_size >> PAGE_SHIFT;
-
-	if (start_page + pages > last_page)
-		pages = last_page - start_page;
-
-	for (i = start_page; i < start_page + pages; ++i) {
-		int index = i / APERTURE_RANGE_PAGES;
-		int page  = i % APERTURE_RANGE_PAGES;
-		__set_bit(page, dom->aperture[index]->bitmap);
-	}
-}
+	unsigned long pfn = 0;
 
-/*
- * This function is used to add a new aperture range to an existing
- * aperture in case of dma_ops domain allocation or address allocation
- * failure.
- */
-static int alloc_new_range(struct dma_ops_domain *dma_dom,
-			   bool populate, gfp_t gfp)
-{
-	int index = dma_dom->aperture_size >> APERTURE_RANGE_SHIFT;
-	unsigned long i, old_size, pte_pgsize;
-	struct aperture_range *range;
-	struct amd_iommu *iommu;
-	unsigned long flags;
+	pages = __roundup_pow_of_two(pages);
 
-#ifdef CONFIG_IOMMU_STRESS
-	populate = false;
-#endif
+	if (dma_mask > DMA_BIT_MASK(32))
+		pfn = alloc_iova_fast(&dma_dom->iovad, pages,
+				      IOVA_PFN(DMA_BIT_MASK(32)));
 
-	if (index >= APERTURE_MAX_RANGES)
-		return -ENOMEM;
-
-	range = kzalloc(sizeof(struct aperture_range), gfp);
-	if (!range)
-		return -ENOMEM;
-
-	range->bitmap = (void *)get_zeroed_page(gfp);
-	if (!range->bitmap)
-		goto out_free;
-
-	range->offset = dma_dom->aperture_size;
-
-	spin_lock_init(&range->bitmap_lock);
-
-	if (populate) {
-		unsigned long address = dma_dom->aperture_size;
-		int i, num_ptes = APERTURE_RANGE_PAGES / 512;
-		u64 *pte, *pte_page;
-
-		for (i = 0; i < num_ptes; ++i) {
-			pte = alloc_pte(&dma_dom->domain, address, PAGE_SIZE,
-					&pte_page, gfp);
-			if (!pte)
-				goto out_free;
-
-			range->pte_pages[i] = pte_page;
-
-			address += APERTURE_RANGE_SIZE / 64;
-		}
-	}
+	if (!pfn)
+		pfn = alloc_iova_fast(&dma_dom->iovad, pages, IOVA_PFN(dma_mask));
 
-	spin_lock_irqsave(&dma_dom->domain.lock, flags);
-
-	/* First take the bitmap_lock and then publish the range */
-	spin_lock(&range->bitmap_lock);
-
-	old_size                 = dma_dom->aperture_size;
-	dma_dom->aperture[index] = range;
-	dma_dom->aperture_size  += APERTURE_RANGE_SIZE;
-
-	/* Reserve address range used for MSI messages */
-	if (old_size < MSI_ADDR_BASE_LO &&
-	    dma_dom->aperture_size > MSI_ADDR_BASE_LO) {
-		unsigned long spage;
-		int pages;
-
-		pages = iommu_num_pages(MSI_ADDR_BASE_LO, 0x10000, PAGE_SIZE);
-		spage = MSI_ADDR_BASE_LO >> PAGE_SHIFT;
-
-		dma_ops_reserve_addresses(dma_dom, spage, pages);
-	}
-
-	/* Initialize the exclusion range if necessary */
-	for_each_iommu(iommu) {
-		if (iommu->exclusion_start &&
-		    iommu->exclusion_start >= dma_dom->aperture[index]->offset
-		    && iommu->exclusion_start < dma_dom->aperture_size) {
-			unsigned long startpage;
-			int pages = iommu_num_pages(iommu->exclusion_start,
-						    iommu->exclusion_length,
-						    PAGE_SIZE);
-			startpage = iommu->exclusion_start >> PAGE_SHIFT;
-			dma_ops_reserve_addresses(dma_dom, startpage, pages);
-		}
-	}
-
-	/*
-	 * Check for areas already mapped as present in the new aperture
-	 * range and mark those pages as reserved in the allocator. Such
-	 * mappings may already exist as a result of requested unity
-	 * mappings for devices.
-	 */
-	for (i = dma_dom->aperture[index]->offset;
-	     i < dma_dom->aperture_size;
-	     i += pte_pgsize) {
-		u64 *pte = fetch_pte(&dma_dom->domain, i, &pte_pgsize);
-		if (!pte || !IOMMU_PTE_PRESENT(*pte))
-			continue;
-
-		dma_ops_reserve_addresses(dma_dom, i >> PAGE_SHIFT,
-					  pte_pgsize >> 12);
-	}
-
-	update_domain(&dma_dom->domain);
-
-	spin_unlock(&range->bitmap_lock);
-
-	spin_unlock_irqrestore(&dma_dom->domain.lock, flags);
-
-	return 0;
-
-out_free:
-	update_domain(&dma_dom->domain);
-
-	free_page((unsigned long)range->bitmap);
-
-	kfree(range);
-
-	return -ENOMEM;
+	return (pfn << PAGE_SHIFT);
 }
 
-static dma_addr_t dma_ops_aperture_alloc(struct dma_ops_domain *dom,
-					 struct aperture_range *range,
-					 unsigned long pages,
-					 unsigned long dma_mask,
-					 unsigned long boundary_size,
-					 unsigned long align_mask,
-					 bool trylock)
+static void dma_ops_free_iova(struct dma_ops_domain *dma_dom,
+			      unsigned long address,
+			      unsigned int pages)
 {
-	unsigned long offset, limit, flags;
-	dma_addr_t address;
-	bool flush = false;
-
-	offset = range->offset >> PAGE_SHIFT;
-	limit  = iommu_device_max_index(APERTURE_RANGE_PAGES, offset,
-					dma_mask >> PAGE_SHIFT);
-
-	if (trylock) {
-		if (!spin_trylock_irqsave(&range->bitmap_lock, flags))
-			return -1;
-	} else {
-		spin_lock_irqsave(&range->bitmap_lock, flags);
-	}
-
-	address = iommu_area_alloc(range->bitmap, limit, range->next_bit,
-				   pages, offset, boundary_size, align_mask);
-	if (address == -1) {
-		/* Nothing found, retry one time */
-		address = iommu_area_alloc(range->bitmap, limit,
-					   0, pages, offset, boundary_size,
-					   align_mask);
-		flush = true;
-	}
-
-	if (address != -1)
-		range->next_bit = address + pages;
-
-	spin_unlock_irqrestore(&range->bitmap_lock, flags);
-
-	if (flush) {
-		domain_flush_tlb(&dom->domain);
-		domain_flush_complete(&dom->domain);
-	}
-
-	return address;
-}
-
-static unsigned long dma_ops_area_alloc(struct device *dev,
-					struct dma_ops_domain *dom,
-					unsigned int pages,
-					unsigned long align_mask,
-					u64 dma_mask)
-{
-	unsigned long boundary_size, mask;
-	unsigned long address = -1;
-	bool first = true;
-	u32 start, i;
-
-	preempt_disable();
-
-	mask = dma_get_seg_boundary(dev);
-
-again:
-	start = this_cpu_read(*dom->next_index);
-
-	/* Sanity check - is it really necessary? */
-	if (unlikely(start > APERTURE_MAX_RANGES)) {
-		start = 0;
-		this_cpu_write(*dom->next_index, 0);
-	}
-
-	boundary_size = mask + 1 ? ALIGN(mask + 1, PAGE_SIZE) >> PAGE_SHIFT :
-				   1UL << (BITS_PER_LONG - PAGE_SHIFT);
-
-	for (i = 0; i < APERTURE_MAX_RANGES; ++i) {
-		struct aperture_range *range;
-		int index;
-
-		index = (start + i) % APERTURE_MAX_RANGES;
-
-		range = dom->aperture[index];
-
-		if (!range || range->offset >= dma_mask)
-			continue;
-
-		address = dma_ops_aperture_alloc(dom, range, pages,
-						 dma_mask, boundary_size,
-						 align_mask, first);
-		if (address != -1) {
-			address = range->offset + (address << PAGE_SHIFT);
-			this_cpu_write(*dom->next_index, index);
-			break;
-		}
-	}
-
-	if (address == -1 && first) {
-		first = false;
-		goto again;
-	}
-
-	preempt_enable();
-
-	return address;
-}
-
-static unsigned long dma_ops_alloc_addresses(struct device *dev,
-					     struct dma_ops_domain *dom,
-					     unsigned int pages,
-					     unsigned long align_mask,
-					     u64 dma_mask)
-{
-	unsigned long address = -1;
-
-	while (address == -1) {
-		address = dma_ops_area_alloc(dev, dom, pages,
-					     align_mask, dma_mask);
-
-		if (address == -1 && alloc_new_range(dom, false, GFP_ATOMIC))
-			break;
-	}
-
-	if (unlikely(address == -1))
-		address = DMA_ERROR_CODE;
-
-	WARN_ON((address + (PAGE_SIZE*pages)) > dom->aperture_size);
-
-	return address;
-}
-
-/*
- * The address free function.
- *
- * called with domain->lock held
- */
-static void dma_ops_free_addresses(struct dma_ops_domain *dom,
-				   unsigned long address,
-				   unsigned int pages)
-{
-	unsigned i = address >> APERTURE_RANGE_SHIFT;
-	struct aperture_range *range = dom->aperture[i];
-	unsigned long flags;
-
-	BUG_ON(i >= APERTURE_MAX_RANGES || range == NULL);
-
-#ifdef CONFIG_IOMMU_STRESS
-	if (i < 4)
-		return;
-#endif
-
-	if (amd_iommu_unmap_flush) {
-		domain_flush_tlb(&dom->domain);
-		domain_flush_complete(&dom->domain);
-	}
-
-	address = (address % APERTURE_RANGE_SIZE) >> PAGE_SHIFT;
-
-	spin_lock_irqsave(&range->bitmap_lock, flags);
-	if (address + pages > range->next_bit)
-		range->next_bit = address + pages;
-	bitmap_clear(range->bitmap, address, pages);
-	spin_unlock_irqrestore(&range->bitmap_lock, flags);
+	pages = __roundup_pow_of_two(pages);
+	address >>= PAGE_SHIFT;
 
+	free_iova_fast(&dma_dom->iovad, address, pages);
 }
 
 /****************************************************************************
@@ -1961,44 +1645,18 @@ static void free_gcr3_table(struct protection_domain *domain)
  */
 static void dma_ops_domain_free(struct dma_ops_domain *dom)
 {
-	int i;
-
 	if (!dom)
 		return;
 
-	free_percpu(dom->next_index);
-
 	del_domain_from_list(&dom->domain);
 
-	free_pagetable(&dom->domain);
+	put_iova_domain(&dom->iovad);
 
-	for (i = 0; i < APERTURE_MAX_RANGES; ++i) {
-		if (!dom->aperture[i])
-			continue;
-		free_page((unsigned long)dom->aperture[i]->bitmap);
-		kfree(dom->aperture[i]);
-	}
+	free_pagetable(&dom->domain);
 
 	kfree(dom);
 }
 
-static int dma_ops_domain_alloc_apertures(struct dma_ops_domain *dma_dom,
-					  int max_apertures)
-{
-	int ret, i, apertures;
-
-	apertures = dma_dom->aperture_size >> APERTURE_RANGE_SHIFT;
-	ret       = 0;
-
-	for (i = apertures; i < max_apertures; ++i) {
-		ret = alloc_new_range(dma_dom, false, GFP_KERNEL);
-		if (ret)
-			break;
-	}
-
-	return ret;
-}
-
 /*
  * Allocates a new protection domain usable for the dma_ops functions.
  * It also initializes the page table and the address allocator data
@@ -2007,7 +1665,6 @@ static int dma_ops_domain_alloc_apertures(struct dma_ops_domain *dma_dom,
 static struct dma_ops_domain *dma_ops_domain_alloc(void)
 {
 	struct dma_ops_domain *dma_dom;
-	int cpu;
 
 	dma_dom = kzalloc(sizeof(struct dma_ops_domain), GFP_KERNEL);
 	if (!dma_dom)
@@ -2016,30 +1673,19 @@ static struct dma_ops_domain *dma_ops_domain_alloc(void)
 	if (protection_domain_init(&dma_dom->domain))
 		goto free_dma_dom;
 
-	dma_dom->next_index = alloc_percpu(u32);
-	if (!dma_dom->next_index)
-		goto free_dma_dom;
-
-	dma_dom->domain.mode = PAGE_MODE_2_LEVEL;
+	dma_dom->domain.mode = PAGE_MODE_3_LEVEL;
 	dma_dom->domain.pt_root = (void *)get_zeroed_page(GFP_KERNEL);
 	dma_dom->domain.flags = PD_DMA_OPS_MASK;
-	dma_dom->domain.priv = dma_dom;
 	if (!dma_dom->domain.pt_root)
 		goto free_dma_dom;
 
-	add_domain_to_list(&dma_dom->domain);
+	init_iova_domain(&dma_dom->iovad, PAGE_SIZE,
+			 IOVA_START_PFN, DMA_32BIT_PFN);
 
-	if (alloc_new_range(dma_dom, true, GFP_KERNEL))
-		goto free_dma_dom;
-
-	/*
-	 * mark the first page as allocated so we never return 0 as
-	 * a valid dma-address. So we can use 0 as error value
-	 */
-	dma_dom->aperture[0]->bitmap[0] = 1;
+	/* Initialize reserved ranges */
+	copy_reserved_iova(&reserved_iova_ranges, &dma_dom->iovad);
 
-	for_each_possible_cpu(cpu)
-		*per_cpu_ptr(dma_dom->next_index, cpu) = 0;
+	add_domain_to_list(&dma_dom->domain);
 
 	return dma_dom;
 
@@ -2482,6 +2128,92 @@ static struct iommu_group *amd_iommu_device_group(struct device *dev)
  *
  *****************************************************************************/
 
+static void __queue_flush(struct flush_queue *queue)
+{
+	struct protection_domain *domain;
+	unsigned long flags;
+	int idx;
+
+	/* First flush TLB of all known domains */
+	spin_lock_irqsave(&amd_iommu_pd_lock, flags);
+	list_for_each_entry(domain, &amd_iommu_pd_list, list)
+		domain_flush_tlb(domain);
+	spin_unlock_irqrestore(&amd_iommu_pd_lock, flags);
+
+	/* Wait until flushes have completed */
+	domain_flush_complete(NULL);
+
+	for (idx = 0; idx < queue->next; ++idx) {
+		struct flush_queue_entry *entry;
+
+		entry = queue->entries + idx;
+
+		free_iova_fast(&entry->dma_dom->iovad,
+				entry->iova_pfn,
+				entry->pages);
+
+		/* Not really necessary, just to make sure we catch any bugs */
+		entry->dma_dom = NULL;
+	}
+
+	queue->next = 0;
+}
+
+static void queue_flush_all(void)
+{
+	int cpu;
+
+	for_each_possible_cpu(cpu) {
+		struct flush_queue *queue;
+		unsigned long flags;
+
+		queue = per_cpu_ptr(&flush_queue, cpu);
+		spin_lock_irqsave(&queue->lock, flags);
+		if (queue->next > 0)
+			__queue_flush(queue);
+		spin_unlock_irqrestore(&queue->lock, flags);
+	}
+}
+
+static void queue_flush_timeout(unsigned long unsused)
+{
+	atomic_set(&queue_timer_on, 0);
+	queue_flush_all();
+}
+
+static void queue_add(struct dma_ops_domain *dma_dom,
+		      unsigned long address, unsigned long pages)
+{
+	struct flush_queue_entry *entry;
+	struct flush_queue *queue;
+	unsigned long flags;
+	int idx;
+
+	pages     = __roundup_pow_of_two(pages);
+	address >>= PAGE_SHIFT;
+
+	queue = get_cpu_ptr(&flush_queue);
+	spin_lock_irqsave(&queue->lock, flags);
+
+	if (queue->next == FLUSH_QUEUE_SIZE)
+		__queue_flush(queue);
+
+	idx   = queue->next++;
+	entry = queue->entries + idx;
+
+	entry->iova_pfn = address;
+	entry->pages    = pages;
+	entry->dma_dom  = dma_dom;
+
+	spin_unlock_irqrestore(&queue->lock, flags);
+
+	if (atomic_cmpxchg(&queue_timer_on, 0, 1) == 0)
+		mod_timer(&queue_timer, jiffies + msecs_to_jiffies(10));
+
+	put_cpu_ptr(&flush_queue);
+}
+
+
 /*
  * In the dma_ops path we only have the struct device. This function
  * finds the corresponding IOMMU, the protection domain and the
@@ -2492,16 +2224,11 @@ static struct iommu_group *amd_iommu_device_group(struct device *dev)
 static struct protection_domain *get_domain(struct device *dev)
 {
 	struct protection_domain *domain;
-	struct iommu_domain *io_domain;
 
 	if (!check_device(dev))
 		return ERR_PTR(-EINVAL);
 
-	io_domain = iommu_get_domain_for_dev(dev);
-	if (!io_domain)
-		return NULL;
-
-	domain = to_pdomain(io_domain);
+	domain = get_dev_data(dev)->domain;
 	if (!dma_ops_domain(domain))
 		return ERR_PTR(-EBUSY);
 
@@ -2512,8 +2239,15 @@ static void update_device_table(struct protection_domain *domain)
 {
 	struct iommu_dev_data *dev_data;
 
-	list_for_each_entry(dev_data, &domain->dev_list, list)
+	list_for_each_entry(dev_data, &domain->dev_list, list) {
 		set_dte_entry(dev_data->devid, domain, dev_data->ats.enabled);
+
+		if (dev_data->devid == dev_data->alias)
+			continue;
+
+		/* There is an alias, update device table entry for it */
+		set_dte_entry(dev_data->alias, domain, dev_data->ats.enabled);
+	}
 }
 
 static void update_domain(struct protection_domain *domain)
@@ -2529,94 +2263,17 @@ static void update_domain(struct protection_domain *domain)
 	domain->updated = false;
 }
 
-/*
- * This function fetches the PTE for a given address in the aperture
- */
-static u64* dma_ops_get_pte(struct dma_ops_domain *dom,
-			    unsigned long address)
-{
-	struct aperture_range *aperture;
-	u64 *pte, *pte_page;
-
-	aperture = dom->aperture[APERTURE_RANGE_INDEX(address)];
-	if (!aperture)
-		return NULL;
-
-	pte = aperture->pte_pages[APERTURE_PAGE_INDEX(address)];
-	if (!pte) {
-		pte = alloc_pte(&dom->domain, address, PAGE_SIZE, &pte_page,
-				GFP_ATOMIC);
-		aperture->pte_pages[APERTURE_PAGE_INDEX(address)] = pte_page;
-	} else
-		pte += PM_LEVEL_INDEX(0, address);
-
-	update_domain(&dom->domain);
-
-	return pte;
-}
-
-/*
- * This is the generic map function. It maps one 4kb page at paddr to
- * the given address in the DMA address space for the domain.
- */
-static dma_addr_t dma_ops_domain_map(struct dma_ops_domain *dom,
-				     unsigned long address,
-				     phys_addr_t paddr,
-				     int direction)
+static int dir2prot(enum dma_data_direction direction)
 {
-	u64 *pte, __pte;
-
-	WARN_ON(address > dom->aperture_size);
-
-	paddr &= PAGE_MASK;
-
-	pte  = dma_ops_get_pte(dom, address);
-	if (!pte)
-		return DMA_ERROR_CODE;
-
-	__pte = paddr | IOMMU_PTE_P | IOMMU_PTE_FC;
-
 	if (direction == DMA_TO_DEVICE)
-		__pte |= IOMMU_PTE_IR;
+		return IOMMU_PROT_IR;
 	else if (direction == DMA_FROM_DEVICE)
-		__pte |= IOMMU_PTE_IW;
+		return IOMMU_PROT_IW;
 	else if (direction == DMA_BIDIRECTIONAL)
-		__pte |= IOMMU_PTE_IR | IOMMU_PTE_IW;
-
-	WARN_ON_ONCE(*pte);
-
-	*pte = __pte;
-
-	return (dma_addr_t)address;
-}
-
-/*
- * The generic unmapping function for on page in the DMA address space.
- */
-static void dma_ops_domain_unmap(struct dma_ops_domain *dom,
-				 unsigned long address)
-{
-	struct aperture_range *aperture;
-	u64 *pte;
-
-	if (address >= dom->aperture_size)
-		return;
-
-	aperture = dom->aperture[APERTURE_RANGE_INDEX(address)];
-	if (!aperture)
-		return;
-
-	pte  = aperture->pte_pages[APERTURE_PAGE_INDEX(address)];
-	if (!pte)
-		return;
-
-	pte += PM_LEVEL_INDEX(0, address);
-
-	WARN_ON_ONCE(!*pte);
-
-	*pte = 0ULL;
+		return IOMMU_PROT_IW | IOMMU_PROT_IR;
+	else
+		return 0;
 }
-
 /*
  * This function contains common code for mapping of a physically
  * contiguous memory region into DMA address space. It is used by all
@@ -2627,32 +2284,29 @@ static dma_addr_t __map_single(struct device *dev,
 			       struct dma_ops_domain *dma_dom,
 			       phys_addr_t paddr,
 			       size_t size,
-			       int dir,
-			       bool align,
+			       enum dma_data_direction direction,
 			       u64 dma_mask)
 {
 	dma_addr_t offset = paddr & ~PAGE_MASK;
 	dma_addr_t address, start, ret;
 	unsigned int pages;
-	unsigned long align_mask = 0;
+	int prot = 0;
 	int i;
 
 	pages = iommu_num_pages(paddr, size, PAGE_SIZE);
 	paddr &= PAGE_MASK;
 
-	if (align)
-		align_mask = (1UL << get_order(size)) - 1;
-
-	addr