Merge branch 'linus' into perf/urgent

Merge reason: we need to fix a non-trivial merge conflict.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

8 files changed, 1645 insertions, 67 deletions

diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index 6bea6962f8ee..3bd9fff5c589 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -142,4 +142,24 @@ config OMAP_IOMMU_DEBUG
 
          Say N unless you know you need this.
 
+config TEGRA_IOMMU_GART
+	bool "Tegra GART IOMMU Support"
+	depends on ARCH_TEGRA_2x_SOC
+	select IOMMU_API
+	help
+	  Enables support for remapping discontiguous physical memory
+	  shared with the operating system into contiguous I/O virtual
+	  space through the GART (Graphics Address Relocation Table)
+	  hardware included on Tegra SoCs.
+
+config TEGRA_IOMMU_SMMU
+	bool "Tegra SMMU IOMMU Support"
+	depends on ARCH_TEGRA_3x_SOC
+	select IOMMU_API
+	help
+	  Enables support for remapping discontiguous physical memory
+	  shared with the operating system into contiguous I/O virtual
+	  space through the SMMU (System Memory Management Unit)
+	  hardware included on Tegra SoCs.
+
 endif # IOMMU_SUPPORT
diff --git a/drivers/iommu/Makefile b/drivers/iommu/Makefile
index 0e36b4934aff..7ad7a3bc1242 100644
--- a/drivers/iommu/Makefile
+++ b/drivers/iommu/Makefile
@@ -8,3 +8,5 @@ obj-$(CONFIG_IRQ_REMAP) += intr_remapping.o
 obj-$(CONFIG_OMAP_IOMMU) += omap-iommu.o
 obj-$(CONFIG_OMAP_IOVMM) += omap-iovmm.o
 obj-$(CONFIG_OMAP_IOMMU_DEBUG) += omap-iommu-debug.o
+obj-$(CONFIG_TEGRA_IOMMU_GART) += tegra-gart.o
+obj-$(CONFIG_TEGRA_IOMMU_SMMU) += tegra-smmu.o
diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index f75e0608be5b..ae2ec929e52f 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -2804,7 +2804,7 @@ static int amd_iommu_dma_supported(struct device *dev, u64 mask)
  * we don't need to preallocate the protection domains anymore.
  * For now we have to.
  */
-static void prealloc_protection_domains(void)
+static void __init prealloc_protection_domains(void)
 {
 	struct iommu_dev_data *dev_data;
 	struct dma_ops_domain *dma_dom;
diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index a35e98ad9725..c56790375e0f 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -196,6 +196,8 @@ static u32 rlookup_table_size;	/* size if the rlookup table */
  */
 extern void iommu_flush_all_caches(struct amd_iommu *iommu);
 
+static int amd_iommu_enable_interrupts(void);
+
 static inline void update_last_devid(u16 devid)
 {
 	if (devid > amd_iommu_last_bdf)
@@ -358,8 +360,6 @@ static void iommu_disable(struct amd_iommu *iommu)
  */
 static u8 * __init iommu_map_mmio_space(u64 address)
 {
-	u8 *ret;
-
 	if (!request_mem_region(address, MMIO_REGION_LENGTH, "amd_iommu")) {
 		pr_err("AMD-Vi: Can not reserve memory region %llx for mmio\n",
 			address);
@@ -367,13 +367,7 @@ static u8 * __init iommu_map_mmio_space(u64 address)
 		return NULL;
 	}
 
-	ret = ioremap_nocache(address, MMIO_REGION_LENGTH);
-	if (ret != NULL)
-		return ret;
-
-	release_mem_region(address, MMIO_REGION_LENGTH);
-
-	return NULL;
+	return ioremap_nocache(address, MMIO_REGION_LENGTH);
 }
 
 static void __init iommu_unmap_mmio_space(struct amd_iommu *iommu)
@@ -1131,8 +1125,9 @@ static int iommu_setup_msi(struct amd_iommu *iommu)
 {
 	int r;
 
-	if (pci_enable_msi(iommu->dev))
-		return 1;
+	r = pci_enable_msi(iommu->dev);
+	if (r)
+		return r;
 
 	r = request_threaded_irq(iommu->dev->irq,
 				 amd_iommu_int_handler,
@@ -1142,27 +1137,36 @@ static int iommu_setup_msi(struct amd_iommu *iommu)
 
 	if (r) {
 		pci_disable_msi(iommu->dev);
-		return 1;
+		return r;
 	}
 
 	iommu->int_enabled = true;
-	iommu_feature_enable(iommu, CONTROL_EVT_INT_EN);
-
-	if (iommu->ppr_log != NULL)
-		iommu_feature_enable(iommu, CONTROL_PPFINT_EN);
 
 	return 0;
 }
 
 static int iommu_init_msi(struct amd_iommu *iommu)
 {
+	int ret;
+
 	if (iommu->int_enabled)
-		return 0;
+		goto enable_faults;
 
 	if (pci_find_capability(iommu->dev, PCI_CAP_ID_MSI))
-		return iommu_setup_msi(iommu);
+		ret = iommu_setup_msi(iommu);
+	else
+		ret = -ENODEV;
 
-	return 1;
+	if (ret)
+		return ret;
+
+enable_faults:
+	iommu_feature_enable(iommu, CONTROL_EVT_INT_EN);
+
+	if (iommu->ppr_log != NULL)
+		iommu_feature_enable(iommu, CONTROL_PPFINT_EN);
+
+	return 0;
 }
 
 /****************************************************************************
@@ -1381,7 +1385,6 @@ static void enable_iommus(void)
 		iommu_enable_ppr_log(iommu);
 		iommu_enable_gt(iommu);
 		iommu_set_exclusion_range(iommu);
-		iommu_init_msi(iommu);
 		iommu_enable(iommu);
 		iommu_flush_all_caches(iommu);
 	}
@@ -1409,6 +1412,8 @@ static void amd_iommu_resume(void)
 
 	/* re-load the hardware */
 	enable_iommus();
+
+	amd_iommu_enable_interrupts();
 }
 
 static int amd_iommu_suspend(void)
@@ -1424,10 +1429,40 @@ static struct syscore_ops amd_iommu_syscore_ops = {
 	.resume = amd_iommu_resume,
 };
 
+static void __init free_on_init_error(void)
+{
+	amd_iommu_uninit_devices();
+
+	free_pages((unsigned long)amd_iommu_pd_alloc_bitmap,
+		   get_order(MAX_DOMAIN_ID/8));
+
+	free_pages((unsigned long)amd_iommu_rlookup_table,
+		   get_order(rlookup_table_size));
+
+	free_pages((unsigned long)amd_iommu_alias_table,
+		   get_order(alias_table_size));
+
+	free_pages((unsigned long)amd_iommu_dev_table,
+		   get_order(dev_table_size));
+
+	free_iommu_all();
+
+	free_unity_maps();
+
+#ifdef CONFIG_GART_IOMMU
+	/*
+	 * We failed to initialize the AMD IOMMU - try fallback to GART
+	 * if possible.
+	 */
+	gart_iommu_init();
+
+#endif
+}
+
 /*
- * This is the core init function for AMD IOMMU hardware in the system.
- * This function is called from the generic x86 DMA layer initialization
- * code.
+ * This is the hardware init function for AMD IOMMU in the system.
+ * This function is called either from amd_iommu_init or from the interrupt
+ * remapping setup code.
  *
  * This function basically parses the ACPI table for AMD IOMMU (IVRS)
  * three times:
@@ -1446,16 +1481,21 @@ static struct syscore_ops amd_iommu_syscore_ops = {
  *		remapping requirements parsed out of the ACPI table in
  *		this last pass.
  *
- * After that the hardware is initialized and ready to go. In the last
- * step we do some Linux specific things like registering the driver in
- * the dma_ops interface and initializing the suspend/resume support
- * functions. Finally it prints some information about AMD IOMMUs and
- * the driver state and enables the hardware.
+ * After everything is set up the IOMMUs are enabled and the necessary
+ * hotplug and suspend notifiers are registered.
  */
-static int __init amd_iommu_init(void)
+int __init amd_iommu_init_hardware(void)
 {
 	int i, ret = 0;
 
+	if (!amd_iommu_detected)
+		return -ENODEV;
+
+	if (amd_iommu_dev_table != NULL) {
+		/* Hardware already initialized */
+		return 0;
+	}
+
 	/*
 	 * First parse ACPI tables to find the largest Bus/Dev/Func
 	 * we need to handle. Upon this information the shared data
@@ -1472,9 +1512,8 @@ static int __init amd_iommu_init(void)
 	alias_table_size   = tbl_size(ALIAS_TABLE_ENTRY_SIZE);
 	rlookup_table_size = tbl_size(RLOOKUP_TABLE_ENTRY_SIZE);
 
-	ret = -ENOMEM;
-
 	/* Device table - directly used by all IOMMUs */
+	ret = -ENOMEM;
 	amd_iommu_dev_table = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO,
 				      get_order(dev_table_size));
 	if (amd_iommu_dev_table == NULL)
@@ -1546,20 +1585,65 @@ static int __init amd_iommu_init(void)
 
 	enable_iommus();
 
+	amd_iommu_init_notifier();
+
+	register_syscore_ops(&amd_iommu_syscore_ops);
+
+out:
+	return ret;
+
+free:
+	free_on_init_error();
+
+	return ret;
+}
+
+static int amd_iommu_enable_interrupts(void)
+{
+	struct amd_iommu *iommu;
+	int ret = 0;
+
+	for_each_iommu(iommu) {
+		ret = iommu_init_msi(iommu);
+		if (ret)
+			goto out;
+	}
+
+out:
+	return ret;
+}
+
+/*
+ * This is the core init function for AMD IOMMU hardware in the system.
+ * This function is called from the generic x86 DMA layer initialization
+ * code.
+ *
+ * The function calls amd_iommu_init_hardware() to setup and enable the
+ * IOMMU hardware if this has not happened yet. After that the driver
+ * registers for the DMA-API and for the IOMMU-API as necessary.
+ */
+static int __init amd_iommu_init(void)
+{
+	int ret = 0;
+
+	ret = amd_iommu_init_hardware();
+	if (ret)
+		goto out;
+
+	ret = amd_iommu_enable_interrupts();
+	if (ret)
+		goto free;
+
 	if (iommu_pass_through)
 		ret = amd_iommu_init_passthrough();
 	else
 		ret = amd_iommu_init_dma_ops();
 
 	if (ret)
-		goto free_disable;
+		goto free;
 
 	amd_iommu_init_api();
 
-	amd_iommu_init_notifier();
-
-	register_syscore_ops(&amd_iommu_syscore_ops);
-
 	if (iommu_pass_through)
 		goto out;
 
@@ -1569,39 +1653,14 @@ static int __init amd_iommu_init(void)
 		printk(KERN_INFO "AMD-Vi: Lazy IO/TLB flushing enabled\n");
 
 	x86_platform.iommu_shutdown = disable_iommus;
+
 out:
 	return ret;
 
-free_disable:
-	disable_iommus();
-
 free:
-	amd_iommu_uninit_devices();
-
-	free_pages((unsigned long)amd_iommu_pd_alloc_bitmap,
-		   get_order(MAX_DOMAIN_ID/8));
-
-	free_pages((unsigned long)amd_iommu_rlookup_table,
-		   get_order(rlookup_table_size));
-
-	free_pages((unsigned long)amd_iommu_alias_table,
-		   get_order(alias_table_size));
-
-	free_pages((unsigned long)amd_iommu_dev_table,
-		   get_order(dev_table_size));
-
-	free_iommu_all();
-
-	free_unity_maps();
-
-#ifdef CONFIG_GART_IOMMU
-	/*
-	 * We failed to initialize the AMD IOMMU - try fallback to GART
-	 * if possible.
-	 */
-	gart_iommu_init();
+	disable_iommus();
 
-#endif
+	free_on_init_error();
 
 	goto out;
 }
diff --git a/drivers/iommu/amd_iommu_v2.c b/drivers/iommu/amd_iommu_v2.c
index 8add9f125d3e..036fe9bf157e 100644
--- a/drivers/iommu/amd_iommu_v2.c
+++ b/drivers/iommu/amd_iommu_v2.c
@@ -921,7 +921,16 @@ static int __init amd_iommu_v2_init(void)
 	size_t state_table_size;
 	int ret;
 
-	pr_info("AMD IOMMUv2 driver by Joerg Roedel <joerg.roedel@amd.com>");
+	pr_info("AMD IOMMUv2 driver by Joerg Roedel <joerg.roedel@amd.com>\n");
+
+	if (!amd_iommu_v2_supported()) {
+		pr_info("AMD IOMMUv2 functionality not available on this sytem\n");
+		/*
+		 * Load anyway to provide the symbols to other modules
+		 * which may use AMD IOMMUv2 optionally.
+		 */
+		return 0;
+	}
 
 	spin_lock_init(&state_lock);
 
@@ -961,6 +970,9 @@ static void __exit amd_iommu_v2_exit(void)
 	size_t state_table_size;
 	int i;
 
+	if (!amd_iommu_v2_supported())
+		return;
+
 	profile_event_unregister(PROFILE_TASK_EXIT, &profile_nb);
 	amd_iommu_unregister_ppr_notifier(&ppr_nb);
 
diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index a08a53448b72..132f93b05154 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -1247,7 +1247,7 @@ static int iommu_init_domains(struct intel_iommu *iommu)
 	unsigned long nlongs;
 
 	ndomains = cap_ndoms(iommu->cap);
-	pr_debug("IOMMU %d: Number of Domains supportd <%ld>\n", iommu->seq_id,
+	pr_debug("IOMMU %d: Number of Domains supported <%ld>\n", iommu->seq_id,
 			ndomains);
 	nlongs = BITS_TO_LONGS(ndomains);
 
diff --git a/drivers/iommu/tegra-gart.c b/drivers/iommu/tegra-gart.c
new file mode 100644
index 000000000000..779306ee7b16
--- /dev/null
+++ b/drivers/iommu/tegra-gart.c
@@ -0,0 +1,451 @@
+/*
+ * IOMMU API for GART in Tegra20
+ *
+ * Copyright (c) 2010-2012, NVIDIA CORPORATION.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#define pr_fmt(fmt)	"%s(): " fmt, __func__
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/list.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/iommu.h>
+
+#include <asm/cacheflush.h>
+
+/* bitmap of the page sizes currently supported */
+#define GART_IOMMU_PGSIZES	(SZ_4K)
+
+#define GART_CONFIG		0x24
+#define GART_ENTRY_ADDR		0x28
+#define GART_ENTRY_DATA		0x2c
+#define GART_ENTRY_PHYS_ADDR_VALID	(1 << 31)
+
+#define GART_PAGE_SHIFT		12
+#define GART_PAGE_SIZE		(1 << GART_PAGE_SHIFT)
+#define GART_PAGE_MASK						\
+	(~(GART_PAGE_SIZE - 1) & ~GART_ENTRY_PHYS_ADDR_VALID)
+
+struct gart_client {
+	struct device		*dev;
+	struct list_head	list;
+};
+
+struct gart_device {
+	void __iomem		*regs;
+	u32			*savedata;
+	u32			page_count;	/* total remappable size */
+	dma_addr_t		iovmm_base;	/* offset to vmm_area */
+	spinlock_t		pte_lock;	/* for pagetable */
+	struct list_head	client;
+	spinlock_t		client_lock;	/* for client list */
+	struct device		*dev;
+};
+
+static struct gart_device *gart_handle; /* unique for a system */
+
+#define GART_PTE(_pfn)						\
+	(GART_ENTRY_PHYS_ADDR_VALID | ((_pfn) << PAGE_SHIFT))
+
+/*
+ * Any interaction between any block on PPSB and a block on APB or AHB
+ * must have these read-back to ensure the APB/AHB bus transaction is
+ * complete before initiating activity on the PPSB block.
+ */
+#define FLUSH_GART_REGS(gart)	((void)readl((gart)->regs + GART_CONFIG))
+
+#define for_each_gart_pte(gart, iova)					\
+	for (iova = gart->iovmm_base;					\
+	     iova < gart->iovmm_base + GART_PAGE_SIZE * gart->page_count; \
+	     iova += GART_PAGE_SIZE)
+
+static inline void gart_set_pte(struct gart_device *gart,
+				unsigned long offs, u32 pte)
+{
+	writel(offs, gart->regs + GART_ENTRY_ADDR);
+	writel(pte, gart->regs + GART_ENTRY_DATA);
+
+	dev_dbg(gart->dev, "%s %08lx:%08x\n",
+		 pte ? "map" : "unmap", offs, pte & GART_PAGE_MASK);
+}
+
+static inline unsigned long gart_read_pte(struct gart_device *gart,
+					  unsigned long offs)
+{
+	unsigned long pte;
+
+	writel(offs, gart->regs + GART_ENTRY_ADDR);
+	pte = readl(gart->regs + GART_ENTRY_DATA);
+
+	return pte;
+}
+
+static void do_gart_setup(struct gart_device *gart, const u32 *data)
+{
+	unsigned long iova;
+
+	for_each_gart_pte(gart, iova)
+		gart_set_pte(gart, iova, data ? *(data++) : 0);
+
+	writel(1, gart->regs + GART_CONFIG);
+	FLUSH_GART_REGS(gart);
+}
+
+#ifdef DEBUG
+static void gart_dump_table(struct gart_device *gart)
+{
+	unsigned long iova;
+	unsigned long flags;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	for_each_gart_pte(gart, iova) {
+		unsigned long pte;
+
+		pte = gart_read_pte(gart, iova);
+
+		dev_dbg(gart->dev, "%s %08lx:%08lx\n",
+			(GART_ENTRY_PHYS_ADDR_VALID & pte) ? "v" : " ",
+			iova, pte & GART_PAGE_MASK);
+	}
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+}
+#else
+static inline void gart_dump_table(struct gart_device *gart)
+{
+}
+#endif
+
+static inline bool gart_iova_range_valid(struct gart_device *gart,
+					 unsigned long iova, size_t bytes)
+{
+	unsigned long iova_start, iova_end, gart_start, gart_end;
+
+	iova_start = iova;
+	iova_end = iova_start + bytes - 1;
+	gart_start = gart->iovmm_base;
+	gart_end = gart_start + gart->page_count * GART_PAGE_SIZE - 1;
+
+	if (iova_start < gart_start)
+		return false;
+	if (iova_end > gart_end)
+		return false;
+	return true;
+}
+
+static int gart_iommu_attach_dev(struct iommu_domain *domain,
+				 struct device *dev)
+{
+	struct gart_device *gart;
+	struct gart_client *client, *c;
+	int err = 0;
+
+	gart = dev_get_drvdata(dev->parent);
+	if (!gart)
+		return -EINVAL;
+	domain->priv = gart;
+
+	client = devm_kzalloc(gart->dev, sizeof(*c), GFP_KERNEL);
+	if (!client)
+		return -ENOMEM;
+	client->dev = dev;
+
+	spin_lock(&gart->client_lock);
+	list_for_each_entry(c, &gart->client, list) {
+		if (c->dev == dev) {
+			dev_err(gart->dev,
+				"%s is already attached\n", dev_name(dev));
+			err = -EINVAL;
+			goto fail;
+		}
+	}
+	list_add(&client->list, &gart->client);
+	spin_unlock(&gart->client_lock);
+	dev_dbg(gart->dev, "Attached %s\n", dev_name(dev));
+	return 0;
+
+fail:
+	devm_kfree(gart->dev, client);
+	spin_unlock(&gart->client_lock);
+	return err;
+}
+
+static void gart_iommu_detach_dev(struct iommu_domain *domain,
+				  struct device *dev)
+{
+	struct gart_device *gart = domain->priv;
+	struct gart_client *c;
+
+	spin_lock(&gart->client_lock);
+
+	list_for_each_entry(c, &gart->client, list) {
+		if (c->dev == dev) {
+			list_del(&c->list);
+			devm_kfree(gart->dev, c);
+			dev_dbg(gart->dev, "Detached %s\n", dev_name(dev));
+			goto out;
+		}
+	}
+	dev_err(gart->dev, "Couldn't find\n");
+out:
+	spin_unlock(&gart->client_lock);
+}
+
+static int gart_iommu_domain_init(struct iommu_domain *domain)
+{
+	return 0;
+}
+
+static void gart_iommu_domain_destroy(struct iommu_domain *domain)
+{
+	struct gart_device *gart = domain->priv;
+
+	if (!gart)
+		return;
+
+	spin_lock(&gart->client_lock);
+	if (!list_empty(&gart->client)) {
+		struct gart_client *c;
+
+		list_for_each_entry(c, &gart->client, list)
+			gart_iommu_detach_dev(domain, c->dev);
+	}
+	spin_unlock(&gart->client_lock);
+	domain->priv = NULL;
+}
+
+static int gart_iommu_map(struct iommu_domain *domain, unsigned long iova,
+			  phys_addr_t pa, size_t bytes, int prot)
+{
+	struct gart_device *gart = domain->priv;
+	unsigned long flags;
+	unsigned long pfn;
+
+	if (!gart_iova_range_valid(gart, iova, bytes))
+		return -EINVAL;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	pfn = __phys_to_pfn(pa);
+	if (!pfn_valid(pfn)) {
+		dev_err(gart->dev, "Invalid page: %08x\n", pa);
+		spin_unlock_irqrestore(&gart->pte_lock, flags);
+		return -EINVAL;
+	}
+	gart_set_pte(gart, iova, GART_PTE(pfn));
+	FLUSH_GART_REGS(gart);
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+	return 0;
+}
+
+static size_t gart_iommu_unmap(struct iommu_domain *domain, unsigned long iova,
+			       size_t bytes)
+{
+	struct gart_device *gart = domain->priv;
+	unsigned long flags;
+
+	if (!gart_iova_range_valid(gart, iova, bytes))
+		return 0;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	gart_set_pte(gart, iova, 0);
+	FLUSH_GART_REGS(gart);
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+	return 0;
+}
+
+static phys_addr_t gart_iommu_iova_to_phys(struct iommu_domain *domain,
+					   unsigned long iova)
+{
+	struct gart_device *gart = domain->priv;
+	unsigned long pte;
+	phys_addr_t pa;
+	unsigned long flags;
+
+	if (!gart_iova_range_valid(gart, iova, 0))
+		return -EINVAL;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	pte = gart_read_pte(gart, iova);
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+
+	pa = (pte & GART_PAGE_MASK);
+	if (!pfn_valid(__phys_to_pfn(pa))) {
+		dev_err(gart->dev, "No entry for %08lx:%08x\n", iova, pa);
+		gart_dump_table(gart);
+		return -EINVAL;
+	}
+	return pa;
+}
+
+static int gart_iommu_domain_has_cap(struct iommu_domain *domain,
+				     unsigned long cap)
+{
+	return 0;
+}
+
+static struct iommu_ops gart_iommu_ops = {
+	.domain_init	= gart_iommu_domain_init,
+	.domain_destroy	= gart_iommu_domain_destroy,
+	.attach_dev	= gart_iommu_attach_dev,
+	.detach_dev	= gart_iommu_detach_dev,
+	.map		= gart_iommu_map,
+	.unmap		= gart_iommu_unmap,
+	.iova_to_phys	= gart_iommu_iova_to_phys,
+	.domain_has_cap	= gart_iommu_domain_has_cap,
+	.pgsize_bitmap	= GART_IOMMU_PGSIZES,
+};
+
+static int tegra_gart_suspend(struct device *dev)
+{
+	struct gart_device *gart = dev_get_drvdata(dev);
+	unsigned long iova;
+	u32 *data = gart->savedata;
+	unsigned long flags;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	for_each_gart_pte(gart, iova)
+		*(data++) = gart_read_pte(gart, iova);
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+	return 0;
+}
+
+static int tegra_gart_resume(struct device *dev)
+{
+	struct gart_device *gart = dev_get_drvdata(dev);
+	unsigned long flags;
+
+	spin_lock_irqsave(&gart->pte_lock, flags);
+	do_gart_setup(gart, gart->savedata);
+	spin_unlock_irqrestore(&gart->pte_lock, flags);
+	return 0;
+}
+
+static int tegra_gart_probe(struct platform_device *pdev)
+{
+	struct gart_device *gart;
+	struct resource *res, *res_remap;
+	void __iomem *gart_regs;
+	int err;
+	struct device *dev = &pdev->dev;
+
+	if (gart_handle)
+		return -EIO;
+
+	BUILD_BUG_ON(PAGE_SHIFT != GART_PAGE_SHIFT);
+
+	/* the GART memory aperture is required */
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	res_remap = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (!res || !res_remap) {
+		dev_err(dev, "GART memory aperture expected\n");
+		return -ENXIO;
+	}
+
+	gart = devm_kzalloc(dev, sizeof(*gart), GFP_KERNEL);
+	if (!gart) {
+		dev_err(dev, "failed to allocate gart_device\n");
+		return -ENOMEM;
+	}
+
+	gart_regs = devm_ioremap(dev, res->start, resource_size(res));
+	if (!gart_regs) {
+		dev_err(dev, "failed to remap GART registers\n");
+		err = -ENXIO;
+		goto fail;
+	}
+
+	gart->dev = &pdev->dev;
+	spin_lock_init(&gart->pte_lock);
+	spin_lock_init(&gart->client_lock);
+	INIT_LIST_HEAD(&gart->client);
+	gart->regs = gart_regs;
+	gart->iovmm_base = (dma_addr_t)res_remap->start;
+	gart->page_count = (resource_size(res_remap) >> GART_PAGE_SHIFT);
+
+	gart->savedata = vmalloc(sizeof(u32) * gart->page_count);
+	if (!gart->savedata) {
+		dev_err(dev, "failed to allocate context save area\n");
+		err = -ENOMEM;
+		goto fail;
+	}
+
+	platform_set_drvdata(pdev, gart);
+	do_gart_setup(gart, NULL);
+
+	gart_handle = gart;
+	return 0;
+
+fail:
+	if (gart_regs)
+		devm_iounmap(dev, gart_regs);
+	if (gart && gart->savedata)
+		vfree(gart->savedata);
+	devm_kfree(dev, gart);
+	return err;
+}
+
+static int tegra_gart_remove(struct platform_device *pdev)
+{
+	struct gart_device *gart = platform_get_drvdata(pdev);
+	struct device *dev = gart->dev;
+
+	writel(0, gart->regs + GART_CONFIG);
+	if (gart->savedata)
+		vfree(gart->savedata);
+	if (gart->regs)
+		devm_iounmap(dev, gart->regs);
+	devm_kfree(dev, gart);
+	gart_handle = NULL;
+	return 0;
+}
+
+const struct dev_pm_ops tegra_gart_pm_ops = {
+	.suspend	= tegra_gart_suspend,
+	.resume		= tegra_gart_resume,
+};
+
+static struct platform_driver tegra_gart_driver = {
+	.probe		= tegra_gart_probe,
+	.remove		= tegra_gart_remove,
+	.driver = {
+		.owner	= THIS_MODULE,
+		.name	= "tegra-gart",
+		.pm	= &tegra_gart_pm_ops,
+	},
+};
+
+static int __devinit tegra_gart_init(void)
+{
+	bus_set_iommu(&platform_bus_type, &gart_iommu_ops);
+	return platform_driver_register(&tegra_gart_driver);
+}
+
+static void __exit tegra_gart_exit(void)
+{
+	platform_driver_unregister(&tegra_gart_driver);
+}
+
+subsys_initcall(tegra_gart_init);
+module_exit(tegra_gart_exit);
+
+MODULE_DESCRIPTION("IOMMU API for GART in Tegra20");
+MODULE_AUTHOR("Hiroshi DOYU <hdoyu@nvidia.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/iommu/tegra-smmu.c b/drivers/iommu/tegra-smmu.c
new file mode 100644
index 000000000000..eb93c821f592
--- /dev/null
+++ b/drivers/iommu/tegra-smmu.c
@@ -0,0 +1,1034 @@
+/*
+ * IOMMU API for SMMU in Tegra30
+ *
+ * Copyright (c) 2011-2012, NVIDIA CORPORATION.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#define pr_fmt(fmt)	"%s(): " fmt, __func__
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+#include <linux/device.h>
+#include <linux/sched.h>
+#include <linux/iommu.h>
+#include <linux/io.h>
+
+#include <asm/page.h>
+#include <asm/cacheflush.h>
+
+#include <mach/iomap.h>
+#include <mach/smmu.h>
+
+/* bitmap of the page sizes currently supported */
+#define SMMU_IOMMU_PGSIZES	(SZ_4K)
+
+#define SMMU_CONFIG				0x10
+#define SMMU_CONFIG_DISABLE			0
+#define SMMU_CONFIG_ENABLE			1
+
+#define SMMU_TLB_CONFIG				0x14
+#define SMMU_TLB_CONFIG_STATS__MASK		(1 << 31)
+#define SMMU_TLB_CONFIG_STATS__ENABLE		(1 << 31)
+#define SMMU_TLB_CONFIG_HIT_UNDER_MISS__ENABLE	(1 << 29)
+#define SMMU_TLB_CONFIG_ACTIVE_LINES__VALUE	0x10
+#define SMMU_TLB_CONFIG_RESET_VAL		0x20000010
+
+#define SMMU_PTC_CONFIG				0x18
+#define SMMU_PTC_CONFIG_STATS__MASK		(1 << 31)
+#define SMMU_PTC_CONFIG_STATS__ENABLE		(1 << 31)
+#define SMMU_PTC_CONFIG_CACHE__ENABLE		(1 << 29)
+#define SMMU_PTC_CONFIG_INDEX_MAP__PATTERN	0x3f
+#define SMMU_PTC_CONFIG_RESET_VAL		0x2000003f
+
+#define SMMU_PTB_ASID				0x1c
+#define SMMU_PTB_ASID_CURRENT_SHIFT		0
+
+#define SMMU_PTB_DATA				0x20
+#define SMMU_PTB_DATA_RESET_VAL			0
+#define SMMU_PTB_DATA_ASID_NONSECURE_SHIFT	29
+#define SMMU_PTB_DATA_ASID_WRITABLE_SHIFT	30
+#define SMMU_PTB_DATA_ASID_READABLE_SHIFT	31
+
+#define SMMU_TLB_FLUSH				0x30
+#define SMMU_TLB_FLUSH_VA_MATCH_ALL		0
+#define SMMU_TLB_FLUSH_VA_MATCH_SECTION		2
+#define SMMU_TLB_FLUSH_VA_MATCH_GROUP		3
+#define SMMU_TLB_FLUSH_ASID_SHIFT		29
+#define SMMU_TLB_FLUSH_ASID_MATCH_DISABLE	0
+#define SMMU_TLB_FLUSH_ASID_MATCH_ENABLE	1
+#define SMMU_TLB_FLUSH_ASID_MATCH_SHIFT		31
+
+#define SMMU_PTC_FLUSH				0x34
+#define SMMU_PTC_FLUSH_TYPE_ALL			0
+#define SMMU_PTC_FLUSH_TYPE_ADR			1
+#define SMMU_PTC_FLUSH_ADR_SHIFT		4
+
+#define SMMU_ASID_SECURITY			0x38
+
+#define SMMU_STATS_TLB_HIT_COUNT		0x1f0
+#define SMMU_STATS_TLB_MISS_COUNT		0x1f4
+#define SMMU_STATS_PTC_HIT_COUNT		0x1f8
+#define SMMU_STATS_PTC_MISS_COUNT		0x1fc
+
+#define SMMU_TRANSLATION_ENABLE_0		0x228
+#define SMMU_TRANSLATION_ENABLE_1		0x22c
+#define SMMU_TRANSLATION_ENABLE_2		0x230
+
+#define SMMU_AFI_ASID	0x238   /* PCIE */
+#define SMMU_AVPC_ASID	0x23c   /* AVP */
+#define SMMU_DC_ASID	0x240   /* Display controller */
+#define SMMU_DCB_ASID	0x244   /* Display controller B */
+#define SMMU_EPP_ASID	0x248   /* Encoder pre-processor */
+#define SMMU_G2_ASID	0x24c   /* 2D engine */
+#define SMMU_HC_ASID	0x250   /* Host1x */
+#define SMMU_HDA_ASID	0x254   /* High-def audio */
+#define SMMU_ISP_ASID	0x258   /* Image signal processor */
+#define SMMU_MPE_ASID	0x264   /* MPEG encoder */
+#define SMMU_NV_ASID	0x268   /* (3D) */
+#define SMMU_NV2_ASID	0x26c   /* (3D) */
+#define SMMU_PPCS_ASID	0x270   /* AHB */
+#define SMMU_SATA_ASID	0x278   /* SATA */
+#define SMMU_VDE_ASID	0x27c   /* Video decoder */
+#define SMMU_VI_ASID	0x280   /* Video input */
+
+#define SMMU_PDE_NEXT_SHIFT		28
+
+/* AHB Arbiter Registers */
+#define AHB_XBAR_CTRL				0xe0
+#define AHB_XBAR_CTRL_SMMU_INIT_DONE_DONE	1
+#define AHB_XBAR_CTRL_SMMU_INIT_DONE_SHIFT	17
+
+#define SMMU_NUM_ASIDS				4
+#define SMMU_TLB_FLUSH_VA_SECTION__MASK		0xffc00000
+#define SMMU_TLB_FLUSH_VA_SECTION__SHIFT	12 /* right shift */
+#define SMMU_TLB_FLUSH_VA_GROUP__MASK		0xffffc000
+#define SMMU_TLB_FLUSH_VA_GROUP__SHIFT		12 /* right shift */
+#define SMMU_TLB_FLUSH_VA(iova, which)	\
+	((((iova) & SMMU_TLB_FLUSH_VA_##which##__MASK) >> \
+		SMMU_TLB_FLUSH_VA_##which##__SHIFT) |	\
+	SMMU_TLB_FLUSH_VA_MATCH_##which)
+#define SMMU_PTB_ASID_CUR(n)	\
+		((n) << SMMU_PTB_ASID_CURRENT_SHIFT)
+#define SMMU_TLB_FLUSH_ASID_MATCH_disable		\
+		(SMMU_TLB_FLUSH_ASID_MATCH_DISABLE <<	\
+			SMMU_TLB_FLUSH_ASID_MATCH_SHIFT)
+#define SMMU_TLB_FLUSH_ASID_MATCH__ENABLE		\
+		(SMMU_TLB_FLUSH_ASID_MATCH_ENABLE <<	\
+			SMMU_TLB_FLUSH_ASID_MATCH_SHIFT)
+
+#define SMMU_PAGE_SHIFT 12
+#define SMMU_PAGE_SIZE	(1 << SMMU_PAGE_SHIFT)
+
+#define SMMU_PDIR_COUNT	1024
+#define SMMU_PDIR_SIZE	(sizeof(unsigned long) * SMMU_PDIR_COUNT)
+#define SMMU_PTBL_COUNT	1024
+#define SMMU_PTBL_SIZE	(sizeof(unsigned long) * SMMU_PTBL_COUNT)
+#define SMMU_PDIR_SHIFT	12
+#define SMMU_PDE_SHIFT	12
+#define SMMU_PTE_SHIFT	12
+#define SMMU_PFN_MASK	0x000fffff
+
+#define SMMU_ADDR_TO_PFN(addr)	((addr) >> 12)
+#define SMMU_ADDR_TO_PDN(addr)	((addr) >> 22)
+#define SMMU_PDN_TO_ADDR(addr)	((pdn) << 22)
+
+#define _READABLE	(1 << SMMU_PTB_DATA_ASID_READABLE_SHIFT)
+#define _WRITABLE	(1 << SMMU_PTB_DATA_ASID_WRITABLE_SHIFT)
+#define _NONSECURE	(1 << SMMU_PTB_DATA_ASID_NONSECURE_SHIFT)
+#define _PDE_NEXT	(1 << SMMU_PDE_NEXT_SHIFT)
+#define _MASK_ATTR	(_READABLE | _WRITABLE | _NONSECURE)
+
+#define _PDIR_ATTR	(_READABLE | _WRITABLE | _NONSECURE)
+
+#define _PDE_ATTR	(_READABLE | _WRITABLE | _NONSECURE)
+#define _PDE_ATTR_N	(_PDE_ATTR | _PDE_NEXT)
+#define _PDE_VACANT(pdn)	(((pdn) << 10) | _PDE_ATTR)
+
+#define _PTE_ATTR	(_READABLE | _WRITABLE | _NONSECURE)
+#define _PTE_VACANT(addr)	(((addr) >> SMMU_PAGE_SHIFT) | _PTE_ATTR)
+
+#define SMMU_MK_PDIR(page, attr)	\
+		((page_to_phys(page) >> SMMU_PDIR_SHIFT) | (attr))
+#define SMMU_MK_PDE(page, attr)		\
+		(unsigned long)((page_to_phys(page) >> SMMU_PDE_SHIFT) | (attr))
+#define SMMU_EX_PTBL_PAGE(pde)		\
+		pfn_to_page((unsigned long)(pde) & SMMU_PFN_MASK)
+#define SMMU_PFN_TO_PTE(pfn, attr)	(unsigned long)((pfn) | (attr))
+
+#define SMMU_ASID_ENABLE(asid)	((asid) | (1 << 31))
+#define SMMU_ASID_DISABLE	0
+#define SMMU_ASID_ASID(n)	((n) & ~SMMU_ASID_ENABLE(0))
+
+#define smmu_client_enable_hwgrp(c, m)	smmu_client_set_hwgrp(c, m, 1)
+#define smmu_client_disable_hwgrp(c)	smmu_client_set_hwgrp(c, 0, 0)
+#define __smmu_client_enable_hwgrp(c, m) __smmu_client_set_hwgrp(c, m, 1)
+#define __smmu_client_disable_hwgrp(c)	__smmu_client_set_hwgrp(c, 0, 0)
+
+#define HWGRP_INIT(client) [HWGRP_##client] = SMMU_##client##_ASID
+
+static const u32 smmu_hwgrp_asid_reg[] = {
+	HWGRP_INIT(AFI),
+	HWGRP_INIT(AVPC),
+	HWGRP_INIT(DC),
+	HWGRP_INIT(DCB),
+	HWGRP_INIT(EPP),
+	HWGRP_INIT(G2),
+	HWGRP_INIT(HC),
+	HWGRP_INIT(HDA),
+	HWGRP_INIT(ISP),
+	HWGRP_INIT(MPE),
+	HWGRP_INIT(NV),
+	HWGRP_INIT(NV2),
+	HWGRP_INIT(PPCS),
+	HWGRP_INIT(SATA),
+	HWGRP_INIT(VDE),
+	HWGRP_INIT(VI),
+};
+#define HWGRP_ASID_REG(x) (smmu_hwgrp_asid_reg[x])
+
+/*
+ * Per client for address space
+ */
+struct smmu_client {
+	struct device		*dev;
+	struct list_head	list;
+	struct smmu_as		*as;
+	u32			hwgrp;
+};
+
+/*
+ * Per address space
+ */
+struct smmu_as {
+	struct smmu_device	*smmu;	/* back pointer to container */
+	unsigned int		asid;
+	spinlock_t		lock;	/* for pagetable */
+	struct page		*pdir_page;
+	unsigned long		pdir_attr;
+	unsigned long		pde_attr;
+	unsigned long		pte_attr;
+	unsigned int		*pte_count;
+
+	struct list_head	client;
+	spinlock_t		client_lock; /* for client list */
+};
+
+/*
+ * Per SMMU device - IOMMU device
+ */
+struct smmu_device {
+	void __iomem	*regs, *regs_ahbarb;
+	unsigned long	iovmm_base;	/* remappable base address */
+	unsigned long	page_count;	/* total remappable size */
+	spinlock_t	lock;
+	char		*name;
+	struct device	*dev;
+	int		num_as;
+	struct smmu_as	*as;		/* Run-time allocated array */
+	struct page *avp_vector_page;	/* dummy page shared by all AS's */
+
+	/*
+	 * Register image savers for suspend/resume
+	 */
+	unsigned long translation_enable_0;
+	unsigned long translation_enable_1;
+	unsigned long translation_enable_2;
+	unsigned long asid_security;
+};
+
+static struct smmu_device *smmu_handle; /* unique for a system */
+
+/*
+ *	SMMU/AHB register accessors
+ */
+static inline u32 smmu_read(struct smmu_device *smmu, size_t offs)
+{
+	return readl(smmu->regs + offs);
+}
+static inline void smmu_write(struct smmu_device *smmu, u32 val, size_t offs)
+{
+	writel(val, smmu->regs + offs);
+}
+
+static inline u32 ahb_read(struct smmu_device *smmu, size_t offs)
+{
+	return readl(smmu->regs_ahbarb + offs);
+}
+static inline void ahb_write(struct smmu_device *smmu, u32 val, size_t offs)
+{
+	writel(val, smmu->regs_ahbarb + offs);
+}
+
+#define VA_PAGE_TO_PA(va, page)	\
+	(page_to_phys(page) + ((unsigned long)(va) & ~PAGE_MASK))
+
+#define FLUSH_CPU_DCACHE(va, page, size)	\
+	do {	\
+		unsigned long _pa_ = VA_PAGE_TO_PA(va, page);		\
+		__cpuc_flush_dcache_area((void *)(va), (size_t)(size));	\
+		outer_flush_range(_pa_, _pa_+(size_t)(size));		\
+	} while (0)
+
+/*
+ * Any interaction between any block on PPSB and a block on APB or AHB
+ * must have these read-back barriers to ensure the APB/AHB bus
+ * transaction is complete before initiating activity on the PPSB
+ * block.
+ */
+#define FLUSH_SMMU_REGS(smmu)	smmu_read(smmu, SMMU_CONFIG)
+
+#define smmu_client_hwgrp(c) (u32)((c)->dev->platform_data)
+
+static int __smmu_client_set_hwgrp(struct smmu_client *c,
+				   unsigned long map, int on)
+{
+	int i;
+	struct smmu_as *as = c->as;
+	u32 val, offs, mask = SMMU_ASID_ENABLE(as->asid);
+	struct smmu_device *smmu = as->smmu;
+
+	WARN_ON(!on && map);
+	if (on && !map)
+		return -EINVAL;
+	if (!on)
+		map = smmu_client_hwgrp(c);
+
+	for_each_set_bit(i, &map, HWGRP_COUNT) {
+		offs = HWGRP_ASID_REG(i);
+		val = smmu_read(smmu, offs);
+		if (on) {
+			if (WARN_ON(val & mask))
+				goto err_hw_busy;
+			val |= mask;
+		} else {
+			WARN_ON((val & mask) == mask);
+			val &= ~mask;
+		}
+		smmu_write(smmu, val, offs);
+	}
+	FLUSH_SMMU_REGS(smmu);
+	c->hwgrp = map;
+	return 0;
+
+err_hw_busy:
+	for_each_set_bit(i, &map, HWGRP_COUNT) {
+		offs = HWGRP_ASID_REG(i);
+		val = smmu_read(smmu, offs);
+		val &= ~mask;
+		smmu_write(smmu, val, offs);
+	}
+	return -EBUSY;
+}
+
+static int smmu_client_set_hwgrp(struct smmu_client *c, u32 map, int on)
+{
+	u32 val;
+	unsigned long flags;
+	struct smmu_as *as = c->as;
+	struct smmu_device *smmu = as->smmu;
+
+	spin_lock_irqsave(&smmu->lock, flags);
+	val = __smmu_client_set_hwgrp(c, map, on);
+	spin_unlock_irqrestore(&smmu->lock, flags);
+	return val;
+}
+
+/*
+ * Flush all TLB entries and all PTC entries
+ * Caller must lock smmu
+ */
+static void smmu_flush_regs(struct smmu_device *smmu, int enable)
+{
+	u32 val;
+
+	smmu_write(smmu, SMMU_PTC_FLUSH_TYPE_ALL, SMMU_PTC_FLUSH);
+	FLUSH_SMMU_REGS(smmu);
+	val = SMMU_TLB_FLUSH_VA_MATCH_ALL |
+		SMMU_TLB_FLUSH_ASID_MATCH_disable;
+	smmu_write(smmu, val, SMMU_TLB_FLUSH);
+
+	if (enable)
+		smmu_write(smmu, SMMU_CONFIG_ENABLE, SMMU_CONFIG);
+	FLUSH_SMMU_REGS(smmu);
+}
+
+static void smmu_setup_regs(struct smmu_device *smmu)
+{
+	int i;
+	u32 val;
+
+	for (i = 0; i < smmu->num_as; i++) {
+		struct smmu_as *as = &smmu->as[i];
+		struct smmu_client *c;
+
+		smmu_write(smmu, SMMU_PTB_ASID_CUR(as->asid), SMMU_PTB_ASID);
+		val = as->pdir_page ?
+			SMMU_MK_PDIR(as->pdir_page, as->pdir_attr) :
+			SMMU_PTB_DATA_RESET_VAL;
+		smmu_write(smmu, val, SMMU_PTB_DATA);
+
+		list_for_each_entry(c, &as->client, list)
+			__smmu_client_set_hwgrp(c, c->hwgrp, 1);
+	}
+
+	smmu_write(smmu, smmu->translation_enable_0, SMMU_TRANSLATION_ENABLE_0);
+	smmu_write(smmu, smmu->translation_enable_1, SMMU_TRANSLATION_ENABLE_1);
+	smmu_write(smmu, smmu->translation_enable_2, SMMU_TRANSLATION_ENABLE_2);
+	smmu_write(smmu, smmu->asid_security, SMMU_ASID_SECURITY);
+	smmu_write(smmu, SMMU_TLB_CONFIG_RESET_VAL, SMMU_TLB_CONFIG);
+	smmu_write(smmu, SMMU_PTC_CONFIG_RESET_VAL, SMMU_PTC_CONFIG);
+
+	smmu_flush_regs(smmu, 1);
+
+	val = ahb_read(smmu, AHB_XBAR_CTRL);
+	val |= AHB_XBAR_CTRL_SMMU_INIT_DONE_DONE <<
+		AHB_XBAR_CTRL_SMMU_INIT_DONE_SHIFT;
+	ahb_write(smmu, val, AHB_XBAR_CTRL);
+}
+
+static void flush_ptc_and_tlb(struct smmu_device *smmu,
+		      struct smmu_as *as, dma_addr_t iova,
+		      unsigned long *pte, struct page *page, int is_pde)
+{
+	u32 val;
+	unsigned long tlb_flush_va = is_pde
+		?  SMMU_TLB_FLUSH_VA(iova, SECTION)
+		:  SMMU_TLB_FLUSH_VA(iova, GROUP);
+
+	val = SMMU_PTC_FLUSH_TYPE_ADR | VA_PAGE_TO_PA(pte, page);
+	smmu_write(smmu, val, SMMU_PTC_FLUSH);
+	FLUSH_SMMU_REGS(smmu);
+	val = tlb_flush_va |
+		SMMU_TLB_FLUSH_ASID_MATCH__ENABLE |
+		(as->asid << SMMU_TLB_FLUSH_ASID_SHIFT);
+	smmu_write(smmu, val, SMMU_TLB_FLUSH);
+	FLUSH_SMMU_REGS(smmu);
+}
+
+static void free_ptbl(struct smmu_as *as, dma_addr_t iova)
+{
+	unsigned long pdn = SMMU_ADDR_TO_PDN(iova);
+	unsigned long *pdir = (unsigned long *)page_address(as->pdir_page);
+
+	if (pdir[pdn] != _PDE_VACANT(pdn)) {
+		dev_dbg(as->smmu->dev, "pdn: %lx\n", pdn);
+
+		ClearPageReserved(SMMU_EX_PTBL_PAGE(pdir[pdn]));
+		__free_page(SMMU_EX_PTBL_PAGE(pdir[pdn]));
+		pdir[pdn] = _PDE_VACANT(pdn);
+		FLUSH_CPU_DCACHE(&pdir[pdn], as->pdir_page, sizeof pdir[pdn]);
+		flush_ptc_and_tlb(as->smmu, as, iova, &pdir[pdn],
+				  as->pdir_page, 1);
+	}
+}
+
+static void free_pdir(struct smmu_as *as)
+{
+	unsigned addr;
+	int count;
+	struct device *dev = as->smmu->dev;
+
+	if (!as->pdir_page)
+		return;
+
+	addr = as->smmu->iovmm_base;
+	count = as->smmu->page_count;
+	while (count-- > 0) {
+		free_ptbl(as, addr);
+		addr += SMMU_PAGE_SIZE * SMMU_PTBL_COUNT;
+	}
+	ClearPageReserved(as->pdir_page);
+	__free_page(as->pdir_page);
+	as->pdir_page = NULL;
+	devm_kfree(dev, as->pte_count);
+	as->pte_count = NULL;
+}
+
+/*
+ * Maps PTBL for given iova and returns the PTE address
+ * Caller must unmap the mapped PTBL returned in *ptbl_page_p
+ */
+static unsigned long *locate_pte(struct smmu_as *as,
+				 dma_addr_t iova, bool allocate,
+				 struct page **ptbl_page_p,
+				 unsigned int **count)
+{
+	unsigned long ptn = SMMU_ADDR_TO_PFN(iova);
+	unsigned long pdn = SMMU_ADDR_TO_PDN(iova);
+	unsigned long *pdir = page_address(as->pdir_page);
+	unsigned long *ptbl;
+
+	if (pdir[pdn] != _PDE_VACANT(pdn)) {
+		/* Mapped entry table already exists */
+		*ptbl_page_p = SMMU_EX_PTBL_PAGE(pdir[pdn]);
+		ptbl = page_address(*ptbl_page_p);
+	} else if (!allocate) {
+		return NULL;
+	} else {
+		int pn;
+		unsigned long addr = SMMU_PDN_TO_ADDR(pdn);
+
+		/* Vacant - allocate a new page table */
+		dev_dbg(as->smmu->dev, "New PTBL pdn: %lx\n", pdn);
+
+		*ptbl_page_p = alloc_page(GFP_ATOMIC);
+		if (!*ptbl_page_p) {
+			dev_err(as->smmu->dev,
+				"failed to allocate smmu_device page table\n");
+			return NULL;
+		}
+		SetPageReserved(*ptbl_page_p);
+		ptbl = (unsigned long *)page_address(*ptbl_page_p);
+		for (pn = 0; pn < SMMU_PTBL_COUNT;
+		     pn++, addr += SMMU_PAGE_SIZE) {
+			ptbl[pn] = _PTE_VACANT(addr);
+		}
+		FLUSH_CPU_DCACHE(ptbl, *ptbl_page_p, SMMU_PTBL_SIZE);
+		pdir[pdn] = SMMU_MK_PDE(*ptbl_page_p,
+					as->pde_attr | _PDE_NEXT);
+		FLUSH_CPU_DCACHE(&pdir[pdn], as->pdir_page, sizeof pdir[pdn]);
+		flush_ptc_and_tlb(as->smmu, as, iova, &pdir[pdn],
+				  as->pdir_page, 1);
+	}
+	*count = &as->pte_count[pdn];
+
+	return &ptbl[ptn % SMMU_PTBL_COUNT];
+}
+
+#ifdef CONFIG_SMMU_SIG_DEBUG
+static void put_signature(struct smmu_as *as,
+			  dma_addr_t iova, unsigned long pfn)
+{
+	struct page *page;
+	unsigned long *vaddr;
+
+	page = pfn_to_page(pfn);
+	vaddr = page_address(page);
+	if (!vaddr)
+		return;
+
+	vaddr[0] = iova;
+	vaddr[1] = pfn << PAGE_SHIFT;
+	FLUSH_CPU_DCACHE(vaddr, page, sizeof(vaddr[0]) * 2);
+}
+#else
+static inline void put_signature(struct smmu_as *as,
+				 unsigned long addr, unsigned long pfn)
+{
+}
+#endif
+
+/*
+ * Caller must lock/unlock as
+ */
+static int alloc_pdir(struct smmu_as *as)
+{
+	unsigned long *pdir;
+	int pdn;
+	u32 val;
+	struct smmu_device *smmu = as->smmu;
+
+	if (as->pdir_page)
+		return 0;
+
+	as->pte_count = devm_kzalloc(smmu->dev,
+		     sizeof(as->pte_count[0]) * SMMU_PDIR_COUNT, GFP_KERNEL);
+	if (!as->pte_count) {
+		dev_err(smmu->dev,
+			"failed to allocate smmu_device PTE cunters\n");
+		return -ENOMEM;
+	}
+	as->pdir_page = alloc_page(GFP_KERNEL | __GFP_DMA);
+	if (!as->pdir_page) {
+		dev_err(smmu->dev,
+			"failed to allocate smmu_device page directory\n");
+		devm_kfree(smmu->dev, as->pte_count);
+		as->pte_count = NULL;
+		return -ENOMEM;
+	}
+	SetPageReserved(as->pdir_page);
+	pdir = page_address(as->pdir_page);
+
+	for (pdn = 0; pdn < SMMU_PDIR_COUNT; pdn++)
+		pdir[pdn] = _PDE_VACANT(pdn);
+	FLUSH_CPU_DCACHE(pdir, as->pdir_page, SMMU_PDIR_SIZE);
+	val = SMMU_PTC_FLUSH_TYPE_ADR | VA_PAGE_TO_PA(pdir, as->pdir_page);
+	smmu_write(smmu, val, SMMU_PTC_FLUSH);
+	FLUSH_SMMU_REGS(as->smmu);
+	val = SMMU_TLB_FLUSH_VA_MATCH_ALL |
+		SMMU_TLB_FLUSH_ASID_MATCH__ENABLE |
+		(as->asid << SMMU_TLB_FLUSH_ASID_SHIFT);
+	smmu_write(smmu, val, SMMU_TLB_FLUSH);
+	FLUSH_SMMU_REGS(as->smmu);
+
+	return 0;
+}
+
+static void __smmu_iommu_unmap(struct smmu_as *as, dma_addr_t iova)
+{
+	unsigned long *pte;
+	struct page *page;
+	unsigned int *count;
+
+	pte = locate_pte(as, iova, false, &page, &count);
+	if (WARN_ON(!pte))
+		return;
+
+	if (WARN_ON(*pte == _PTE_VACANT(iova)))
+		return;
+
+	*pte = _PTE_VACANT(iova);
+	FLUSH_CPU_DCACHE(pte, page, sizeof(*pte));
+	flush_ptc_and_tlb(as->smmu, as, iova, pte, page, 0);
+	if (!--(*count)) {
+		free_ptbl(as, iova);
+		smmu_flush_regs(as->smmu, 0);
+	}
+}
+
+static void __smmu_iommu_map_pfn(struct smmu_as *as, dma_addr_t iova,
+				 unsigned long pfn)
+{
+	struct smmu_device *smmu = as->smmu;
+	unsigned long *pte;
+	unsigned int *count;
+	struct page *page;
+
+	pte = locate_pte(as, iova, true, &page, &count);
+	if (WARN_ON(!pte))
+		return;
+
+	if (*pte == _PTE_VACANT(iova))
+		(*count)++;
+	*pte = SMMU_PFN_TO_PTE(pfn, as->pte_attr);
+	if (unlikely((*pte == _PTE_VACANT(iova))))
+		(*count)--;
+	FLUSH_CPU_DCACHE(pte, page, sizeof(*pte));
+	flush_ptc_and_tlb(smmu, as, iova, pte, page, 0);
+	put_signature(as, iova, pfn);
+}
+
+static int smmu_iommu_map(struct iommu_domain *domain, unsigned long iova,
+			  phys_addr_t pa, size_t bytes, int prot)
+{
+	struct smmu_as *as = domain->priv;
+	unsigned long pfn = __phys_to_pfn(pa);
+	unsigned long flags;
+
+	dev_dbg(as->smmu->dev, "[%d] %08lx:%08x\n", as->asid, iova, pa);
+
+	if (!pfn_valid(pfn))
+		return -ENOMEM;
+
+	spin_lock_irqsave(&as->lock, flags);
+	__smmu_iommu_map_pfn(as, iova, pfn);
+	spin_unlock_irqrestore(&as->lock, flags);
+	return 0;
+}
+
+static size_t smmu_iommu_unmap(struct iommu_domain *domain, unsigned long iova,
+			       size_t bytes)
+{
+	struct smmu_as *as = domain->priv;
+	unsigned long flags;
+
+	dev_dbg(as->smmu->dev, "[%d] %08lx\n", as->asid, iova);
+
+	spin_lock_irqsave(&as->lock, flags);
+	__smmu_iommu_unmap(as, iova);
+	spin_unlock_irqrestore(&as->lock, flags);
+	return SMMU_PAGE_SIZE;
+}
+
+static phys_addr_t smmu_iommu_iova_to_phys(struct iommu_domain *domain,
+					   unsigned long iova)
+{
+	struct smmu_as *as = domain->priv;
+	unsigned long *pte;
+	unsigned int *count;
+	struct page *page;
+	unsigned long pfn;
+	unsigned long flags;
+
+	spin_lock_irqsave(&as->lock, flags);
+
+	pte = locate_pte(as, iova, true, &page, &count);
+	pfn = *pte & SMMU_PFN_MASK;
+	WARN_ON(!pfn_valid(pfn));
+	dev_dbg(as->smmu->dev,
+		"iova:%08lx pfn:%08lx asid:%d\n", iova, pfn, as->asid);
+
+	spin_unlock_irqrestore(&as->lock, flags);
+	return PFN_PHYS(pfn);
+}
+
+static int smmu_iommu_domain_has_cap(struct iommu_domain *domain,
+				     unsigned long cap)
+{
+	return 0;
+}
+
+static int smmu_iommu_attach_dev(struct iommu_domain *domain,
+				 struct device *dev)
+{
+	struct smmu_as *as = domain->priv;
+	struct smmu_device *smmu = as->smmu;
+	struct smmu_client *client, *c;
+	u32 map;
+	int err;
+
+	client = devm_kzalloc(smmu->dev, sizeof(*c), GFP_KERNEL);
+	if (!client)
+		return -ENOMEM;
+	client->dev = dev;
+	client->as = as;
+	map = (unsigned long)dev->platform_data;
+	if (!map)
+		return -EINVAL;
+
+	err = smmu_client_enable_hwgrp(client, map);
+	if (err)
+		goto err_hwgrp;
+
+	spin_lock(&as->client_lock);
+	list_for_each_entry(c, &as->client, list) {
+		if (c->dev == dev) {
+			dev_err(smmu->dev,
+				"%s is already attached\n", dev_name(c->dev));
+			err = -EINVAL;
+			goto err_client;
+		}
+	}
+	list_add(&client->list, &as->client);
+	spin_unlock(&as->client_lock);
+
+	/*
+	 * Reserve "page zero" for AVP vectors using a common dummy
+	 * page.
+	 */
+	if (map & HWG_AVPC) {
+		struct page *page;
+
+		page = as->smmu->avp_vector_page;
+		__smmu_iommu_map_pfn(as, 0, page_to_pfn(page));
+
+		pr_info("Reserve \"page zero\" for AVP vectors using a common dummy\n");
+	}
+
+	dev_dbg(smmu->dev, "%s is attached\n", dev_name(c->dev));
+	return 0;
+
+err_client:
+	smmu_client_disable_hwgrp(client);
+	spin_unlock(&as->client_lock);
+err_hwgrp:
+	devm_kfree(smmu->dev, client);
+	return err;
+}
+
+static void smmu_iommu_detach_dev(struct iommu_domain *domain,
+				  struct device *dev)
+{
+	struct smmu_as *as = domain->priv;
+	struct smmu_device *smmu = as->smmu;
+	struct smmu_client *c;
+
+	spin_lock(&as->client_lock);
+
+	list_for_each_entry(c, &as->client, list) {
+		if (c->dev == dev) {
+			smmu_client_disable_hwgrp(c);
+			list_del(&c->list);
+			devm_kfree(smmu->dev, c);
+			c->as = NULL;
+			dev_dbg(smmu->dev,
+				"%s is detached\n", dev_name(c->dev));
+			goto out;
+		}
+	}
+	dev_err(smmu->dev, "Couldn't find %s\n", dev_name(c->dev));
+out:
+	spin_unlock(&as->client_lock);
+}
+
+static int smmu_iommu_domain_init(struct iommu_domain *domain)
+{
+	int i;
+	unsigned long flags;
+	struct smmu_as *as;
+	struct smmu_device *smmu = smmu_handle;
+
+	/* Look for a free AS with lock held */
+	for  (i = 0; i < smmu->num_as; i++) {
+		struct smmu_as *tmp = &smmu->as[i];
+
+		spin_lock_irqsave(&tmp->lock, flags);
+		if (!tmp->pdir_page) {
+			as = tmp;
+			goto found;
+		}
+		spin_unlock_irqrestore(&tmp->lock, flags);
+	}
+	dev_err(smmu->dev, "no free AS\n");
+	return -ENODEV;
+
+found:
+	if (alloc_pdir(as) < 0)
+		goto err_alloc_pdir;
+
+	spin_lock(&smmu->lock);
+
+	/* Update PDIR register */
+	smmu_write(smmu, SMMU_PTB_ASID_CUR(as->asid), SMMU_PTB_ASID);
+	smmu_write(smmu,
+		   SMMU_MK_PDIR(as->pdir_page, as->pdir_attr), SMMU_PTB_DATA);
+	FLUSH_SMMU_REGS(smmu);
+
+	spin_unlock(&smmu->lock);
+
+	spin_unlock_irqrestore(&as->lock, flags);
+	domain->priv = as;
+
+	dev_dbg(smmu->dev, "smmu_as@%p\n", as);
+	return 0;
+
+err_alloc_pdir:
+	spin_unlock_irqrestore(&as->lock, flags);
+	return -ENODEV;
+}
+
+static void smmu_iommu_domain_destroy(struct iommu_domain *domain)
+{
+	struct smmu_as *as = domain->priv;
+	struct smmu_device *smmu = as->smmu;
+	unsigned long flags;
+
+	spin_lock_irqsave(&as->lock, flags);
+
+	if (as->pdir_page) {
+		spin_lock(&smmu->lock);
+		smmu_write(smmu, SMMU_PTB_ASID_CUR(as->asid), SMMU_PTB_ASID);
+		smmu_write(smmu, SMMU_PTB_DATA_RESET_VAL, SMMU_PTB_DATA);
+		FLUSH_SMMU_REGS(smmu);
+		spin_unlock(&smmu->lock);
+
+		free_pdir(as);
+	}
+
+	if (!list_empty(&as->client)) {
+		struct smmu_client *c;
+
+		list_for_each_entry(c, &as->client, list)
+			smmu_iommu_detach_dev(domain, c->dev);
+	}
+
+	spin_unlock_irqrestore(&as->lock, flags);
+
+	domain->priv = NULL;
+	dev_dbg(smmu->dev, "smmu_as@%p\n", as);
+}
+
+static struct iommu_ops smmu_iommu_ops = {
+	.domain_init	= smmu_iommu_domain_init,
+	.domain_destroy	= smmu_iommu_domain_destroy,
+	.attach_dev	= smmu_iommu_attach_dev,
+	.detach_dev	= smmu_iommu_detach_dev,
+	.map		= smmu_iommu_map,
+	.unmap		= smmu_iommu_unmap,
+	.iova_to_phys	= smmu_iommu_iova_to_phys,
+	.domain_has_cap	= smmu_iommu_domain_has_cap,
+	.pgsize_bitmap	= SMMU_IOMMU_PGSIZES,
+};
+
+static int tegra_smmu_suspend(struct device *dev)
+{
+	struct smmu_device *smmu = dev_get_drvdata(dev);
+
+	smmu->translation_enable_0 = smmu_read(smmu, SMMU_TRANSLATION_ENABLE_0);
+	smmu->translation_enable_1 = smmu_read(smmu, SMMU_TRANSLATION_ENABLE_1);
+	smmu->translation_enable_2 = smmu_read(smmu, SMMU_TRANSLATION_ENABLE_2);
+	smmu->asid_security = smmu_read(smmu, SMMU_ASID_SECURITY);
+	return 0;
+}
+
+static int tegra_smmu_resume(struct device *dev)
+{
+	struct smmu_device *smmu = dev_get_drvdata(dev);
+	unsigned long flags;
+
+	spin_lock_irqsave(&smmu->lock, flags);
+	smmu_setup_regs(smmu);
+	spin_unlock_irqrestore(&smmu->lock, flags);
+	return 0;
+}
+
+static int tegra_smmu_probe(struct platform_device *pdev)
+{
+	struct smmu_device *smmu;
+	struct resource *regs, *regs2, *window;
+	struct device *dev = &pdev->dev;
+	int i, err = 0;
+
+	if (smmu_handle)
+		return -EIO;
+
+	BUILD_BUG_ON(PAGE_SHIFT != SMMU_PAGE_SHIFT);
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	regs2 = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	window = platform_get_resource(pdev, IORESOURCE_MEM, 2);
+	if (!regs || !regs2 || !window) {
+		dev_err(dev, "No SMMU resources\n");
+		return -ENODEV;
+	}
+
+	smmu = devm_kzalloc(dev, sizeof(*smmu), GFP_KERNEL);
+	if (!smmu) {
+		dev_err(dev, "failed to allocate smmu_device\n");
+		return -ENOMEM;
+	}
+
+	smmu->dev = dev;
+	smmu->num_as = SMMU_NUM_ASIDS;
+	smmu->iovmm_base = (unsigned long)window->start;
+	smmu->page_count = resource_size(window) >> SMMU_PAGE_SHIFT;
+	smmu->regs = devm_ioremap(dev, regs->start, resource_size(regs));
+	smmu->regs_ahbarb = devm_ioremap(dev, regs2->start,
+					 resource_size(regs2));
+	if (!smmu->regs || !smmu->regs_ahbarb) {
+		dev_err(dev, "failed to remap SMMU registers\n");
+		err = -ENXIO;
+		goto fail;
+	}
+
+	smmu->translation_enable_0 = ~0;
+	smmu->translation_enable_1 = ~0;
+	smmu->translation_enable_2 = ~0;
+	smmu->asid_security = 0;
+
+	smmu->as = devm_kzalloc(dev,
+			sizeof(smmu->as[0]) * smmu->num_as, GFP_KERNEL);
+	if (!smmu->as) {
+		dev_err(dev, "failed to allocate smmu_as\n");
+		err = -ENOMEM;
+		goto fail;
+	}
+
+	for (i = 0; i < smmu->num_as; i++) {
+		struct smmu_as *as = &smmu->as[i];
+
+		as->smmu = smmu;
+		as->asid = i;
+		as->pdir_attr = _PDIR_ATTR;
+		as->pde_attr = _PDE_ATTR;
+		as->pte_attr = _PTE_ATTR;
+
+		spin_lock_init(&as->lock);
+		INIT_LIST_HEAD(&as->client);
+	}
+	spin_lock_init(&smmu->lock);
+	smmu_setup_regs(smmu);
+	platform_set_drvdata(pdev, smmu);
+
+	smmu->avp_vector_page = alloc_page(GFP_KERNEL);
+	if (!smmu->avp_vector_page)
+		goto fail;
+
+	smmu_handle = smmu;
+	return 0;
+
+fail:
+	if (smmu->avp_vector_page)
+		__free_page(smmu->avp_vector_page);
+	if (smmu->regs)
+		devm_iounmap(dev, smmu->regs);
+	if (smmu->regs_ahbarb)
+		devm_iounmap(dev, smmu->regs_ahbarb);
+	if (smmu && smmu->as) {
+		for (i = 0; i < smmu->num_as; i++) {
+			if (smmu->as[i].pdir_page) {
+				ClearPageReserved(smmu->as[i].pdir_page);
+				__free_page(smmu->as[i].pdir_page);
+			}
+		}
+		devm_kfree(dev, smmu->as);
+	}
+	devm_kfree(dev, smmu);
+	return err;
+}
+
+static int tegra_smmu_remove(struct platform_device *pdev)
+{
+	struct smmu_device *smmu = platform_get_drvdata(pdev);
+	struct device *dev = smmu->dev;
+
+	smmu_write(smmu, SMMU_CONFIG_DISABLE, SMMU_CONFIG);
+	platform_set_drvdata(pdev, NULL);
+	if (smmu->as) {
+		int i;
+
+		for (i = 0; i < smmu->num_as; i++)
+			free_pdir(&smmu->as[i]);
+		devm_kfree(dev, smmu->as);
+	}
+	if (smmu->avp_vector_page)
+		__free_page(smmu->avp_vector_page);
+	if (smmu->regs)
+		devm_iounmap(dev, smmu->regs);
+	if (smmu->regs_ahbarb)
+		devm_iounmap(dev, smmu->regs_ahbarb);
+	devm_kfree(dev, smmu);
+	smmu_handle = NULL;
+	return 0;
+}
+
+const struct dev_pm_ops tegra_smmu_pm_ops = {
+	.suspend	= tegra_smmu_suspend,
+	.resume		= tegra_smmu_resume,
+};
+
+static struct platform_driver tegra_smmu_driver = {
+	.probe		= tegra_smmu_probe,
+	.remove		= tegra_smmu_remove,
+	.driver = {
+		.owner	= THIS_MODULE,
+		.name	= "tegra-smmu",
+		.pm	= &tegra_smmu_pm_ops,
+	},
+};
+
+static int __devinit tegra_smmu_init(void)
+{
+	bus_set_iommu(&platform_bus_type, &smmu_iommu_ops);
+	return platform_driver_register(&tegra_smmu_driver);
+}
+
+static void __exit tegra_smmu_exit(void)
+{
+	platform_driver_unregister(&tegra_smmu_driver);
+}
+
+subsys_initcall(tegra_smmu_init);
+module_exit(tegra_smmu_exit);
+
+MODULE_DESCRIPTION("IOMMU API for SMMU in Tegra30");
+MODULE_AUTHOR("Hiroshi DOYU <hdoyu@nvidia.com>");
+MODULE_LICENSE("GPL v2");