memory: Add NVIDIA Tegra memory controller support

The memory controller on NVIDIA Tegra exposes various knobs that can be
used to tune the behaviour of the clients attached to it.

Currently this driver sets up the latency allowance registers to the HW
defaults. Eventually an API should be exported by this driver (via a
custom API or a generic subsystem) to allow clients to register latency
requirements.

This driver also registers an IOMMU (SMMU) that's implemented by the
memory controller. It is supported on Tegra30, Tegra114 and Tegra124
currently. Tegra20 has a GART instead.

The Tegra SMMU operates on memory clients and SWGROUPs. A memory client
is a unidirectional, special-purpose DMA master. A SWGROUP represents a
set of memory clients that form a logical functional unit corresponding
to a single device. Typically a device has two clients: one client for
read transactions and one client for write transactions, but there are
also devices that have only read clients, but many of them (such as the
display controllers).

Because there is no 1:1 relationship between memory clients and devices
the driver keeps a table of memory clients and the SWGROUPs that they
belong to per SoC. Note that this is an exception and due to the fact
that the SMMU is tightly integrated with the rest of the Tegra SoC. The
use of these tables is discouraged in drivers for generic IOMMU devices
such as the ARM SMMU because the same IOMMU could be used in any number
of SoCs and keeping such tables for each SoC would not scale.

Acked-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Thierry Reding <treding@nvidia.com>

1 files changed, 523 insertions, 1087 deletions

diff --git a/drivers/iommu/tegra-smmu.c b/drivers/iommu/tegra-smmu.c
index 73e845a66925..6e134c7c227f 100644
--- a/drivers/iommu/tegra-smmu.c
+++ b/drivers/iommu/tegra-smmu.c
@@ -1,1296 +1,732 @@
 /*
- * IOMMU API for SMMU in Tegra30
+ * Copyright (C) 2011-2014 NVIDIA CORPORATION.  All rights reserved.
  *
- * Copyright (c) 2011-2013, 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.
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
  */
 
-#define pr_fmt(fmt)	"%s(): " fmt, __func__
-
 #include <linux/err.h>
-#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 <linux/kernel.h>
 #include <linux/of.h>
-#include <linux/of_iommu.h>
-#include <linux/debugfs.h>
-#include <linux/seq_file.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
 
 #include <soc/tegra/ahb.h>
+#include <soc/tegra/mc.h>
 
-#include <asm/page.h>
-#include <asm/cacheflush.h>
-
-enum smmu_hwgrp {
-	HWGRP_AFI,
-	HWGRP_AVPC,
-	HWGRP_DC,
-	HWGRP_DCB,
-	HWGRP_EPP,
-	HWGRP_G2,
-	HWGRP_HC,
-	HWGRP_HDA,
-	HWGRP_ISP,
-	HWGRP_MPE,
-	HWGRP_NV,
-	HWGRP_NV2,
-	HWGRP_PPCS,
-	HWGRP_SATA,
-	HWGRP_VDE,
-	HWGRP_VI,
-
-	HWGRP_COUNT,
-
-	HWGRP_END = ~0,
-};
+struct tegra_smmu {
+	void __iomem *regs;
+	struct device *dev;
 
-#define HWG_AFI		(1 << HWGRP_AFI)
-#define HWG_AVPC	(1 << HWGRP_AVPC)
-#define HWG_DC		(1 << HWGRP_DC)
-#define HWG_DCB		(1 << HWGRP_DCB)
-#define HWG_EPP		(1 << HWGRP_EPP)
-#define HWG_G2		(1 << HWGRP_G2)
-#define HWG_HC		(1 << HWGRP_HC)
-#define HWG_HDA		(1 << HWGRP_HDA)
-#define HWG_ISP		(1 << HWGRP_ISP)
-#define HWG_MPE		(1 << HWGRP_MPE)
-#define HWG_NV		(1 << HWGRP_NV)
-#define HWG_NV2		(1 << HWGRP_NV2)
-#define HWG_PPCS	(1 << HWGRP_PPCS)
-#define HWG_SATA	(1 << HWGRP_SATA)
-#define HWG_VDE		(1 << HWGRP_VDE)
-#define HWG_VI		(1 << HWGRP_VI)
-
-/* 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
-
-/* REVISIT: To support multiple MCs */
-enum {
-	_MC = 0,
-};
+	struct tegra_mc *mc;
+	const struct tegra_smmu_soc *soc;
 
-enum {
-	_TLB = 0,
-	_PTC,
-};
+	unsigned long *asids;
+	struct mutex lock;
 
-#define SMMU_CACHE_CONFIG_BASE			0x14
-#define __SMMU_CACHE_CONFIG(mc, cache)		(SMMU_CACHE_CONFIG_BASE + 4 * cache)
-#define SMMU_CACHE_CONFIG(cache)		__SMMU_CACHE_CONFIG(_MC, cache)
-
-#define SMMU_CACHE_CONFIG_STATS_SHIFT		31
-#define SMMU_CACHE_CONFIG_STATS_ENABLE		(1 << SMMU_CACHE_CONFIG_STATS_SHIFT)
-#define SMMU_CACHE_CONFIG_STATS_TEST_SHIFT	30
-#define SMMU_CACHE_CONFIG_STATS_TEST		(1 << SMMU_CACHE_CONFIG_STATS_TEST_SHIFT)
-
-#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_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_CACHE_COUNT_BASE		0x1f0
-
-#define SMMU_STATS_CACHE_COUNT(mc, cache, hitmiss)		\
-	(SMMU_STATS_CACHE_COUNT_BASE + 8 * cache + 4 * hitmiss)
-
-#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
-
-#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_PAGE_MASK	((1 << SMMU_PAGE_SHIFT) - 1)
-
-#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(pdn)	((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 NUM_SMMU_REG_BANKS	3
-
-#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),
+	struct list_head list;
 };
-#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;
+struct tegra_smmu_as {
+	struct iommu_domain *domain;
+	struct tegra_smmu *smmu;
+	unsigned int use_count;
+	struct page *count;
+	struct page *pd;
+	unsigned id;
+	u32 attr;
 };
 
-/*
- * 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 */
-};
+static inline void smmu_writel(struct tegra_smmu *smmu, u32 value,
+			       unsigned long offset)
+{
+	writel(value, smmu->regs + offset);
+}
 
-struct smmu_debugfs_info {
-	struct smmu_device *smmu;
-	int mc;
-	int cache;
-};
+static inline u32 smmu_readl(struct tegra_smmu *smmu, unsigned long offset)
+{
+	return readl(smmu->regs + offset);
+}
 
-/*
- * Per SMMU device - IOMMU device
- */
-struct smmu_device {
-	void __iomem	*regbase;	/* register offset base */
-	void __iomem	**regs;		/* register block start address array */
-	void __iomem	**rege;		/* register block end address array */
-	int		nregs;		/* number of register blocks */
-
-	unsigned long	iovmm_base;	/* remappable base address */
-	unsigned long	page_count;	/* total remappable size */
-	spinlock_t	lock;
-	char		*name;
-	struct device	*dev;
-	struct page *avp_vector_page;	/* dummy page shared by all AS's */
+#define SMMU_CONFIG 0x010
+#define  SMMU_CONFIG_ENABLE (1 << 0)
 
-	/*
-	 * 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;
+#define SMMU_TLB_CONFIG 0x14
+#define  SMMU_TLB_CONFIG_HIT_UNDER_MISS (1 << 29)
+#define  SMMU_TLB_CONFIG_ROUND_ROBIN_ARBITRATION (1 << 28)
+#define  SMMU_TLB_CONFIG_ACTIVE_LINES(x) ((x) & 0x3f)
 
-	struct dentry *debugfs_root;
-	struct smmu_debugfs_info *debugfs_info;
+#define SMMU_PTC_CONFIG 0x18
+#define  SMMU_PTC_CONFIG_ENABLE (1 << 29)
+#define  SMMU_PTC_CONFIG_REQ_LIMIT(x) (((x) & 0x0f) << 24)
+#define  SMMU_PTC_CONFIG_INDEX_MAP(x) ((x) & 0x3f)
 
-	struct device_node *ahb;
+#define SMMU_PTB_ASID 0x01c
+#define  SMMU_PTB_ASID_VALUE(x) ((x) & 0x7f)
 
-	int		num_as;
-	struct smmu_as	as[0];		/* Run-time allocated array */
-};
+#define SMMU_PTB_DATA 0x020
+#define  SMMU_PTB_DATA_VALUE(page, attr) (page_to_phys(page) >> 12 | (attr))
 
-static struct smmu_device *smmu_handle; /* unique for a system */
+#define SMMU_MK_PDE(page, attr) (page_to_phys(page) >> SMMU_PTE_SHIFT | (attr))
 
-/*
- *	SMMU register accessors
- */
-static bool inline smmu_valid_reg(struct smmu_device *smmu,
-				  void __iomem *addr)
-{
-	int i;
+#define SMMU_TLB_FLUSH 0x030
+#define  SMMU_TLB_FLUSH_VA_MATCH_ALL     (0 << 0)
+#define  SMMU_TLB_FLUSH_VA_MATCH_SECTION (2 << 0)
+#define  SMMU_TLB_FLUSH_VA_MATCH_GROUP   (3 << 0)
+#define  SMMU_TLB_FLUSH_ASID(x)          (((x) & 0x7f) << 24)
+#define  SMMU_TLB_FLUSH_VA_SECTION(addr) ((((addr) & 0xffc00000) >> 12) | \
+					  SMMU_TLB_FLUSH_VA_MATCH_SECTION)
+#define  SMMU_TLB_FLUSH_VA_GROUP(addr)   ((((addr) & 0xffffc000) >> 12) | \
+					  SMMU_TLB_FLUSH_VA_MATCH_GROUP)
+#define  SMMU_TLB_FLUSH_ASID_MATCH       (1 << 31)
 
-	for (i = 0; i < smmu->nregs; i++) {
-		if (addr < smmu->regs[i])
-			break;
-		if (addr <= smmu->rege[i])
-			return true;
-	}
+#define SMMU_PTC_FLUSH 0x034
+#define  SMMU_PTC_FLUSH_TYPE_ALL (0 << 0)
+#define  SMMU_PTC_FLUSH_TYPE_ADR (1 << 0)
 
-	return false;
-}
+#define SMMU_PTC_FLUSH_HI 0x9b8
+#define  SMMU_PTC_FLUSH_HI_MASK 0x3
 
-static inline u32 smmu_read(struct smmu_device *smmu, size_t offs)
-{
-	void __iomem *addr = smmu->regbase + offs;
+/* per-SWGROUP SMMU_*_ASID register */
+#define SMMU_ASID_ENABLE (1 << 31)
+#define SMMU_ASID_MASK 0x7f
+#define SMMU_ASID_VALUE(x) ((x) & SMMU_ASID_MASK)
 
-	BUG_ON(!smmu_valid_reg(smmu, addr));
+/* page table definitions */
+#define SMMU_NUM_PDE 1024
+#define SMMU_NUM_PTE 1024
 
-	return readl(addr);
-}
+#define SMMU_SIZE_PD (SMMU_NUM_PDE * 4)
+#define SMMU_SIZE_PT (SMMU_NUM_PTE * 4)
 
-static inline void smmu_write(struct smmu_device *smmu, u32 val, size_t offs)
-{
-	void __iomem *addr = smmu->regbase + offs;
+#define SMMU_PDE_SHIFT 22
+#define SMMU_PTE_SHIFT 12
 
-	BUG_ON(!smmu_valid_reg(smmu, addr));
+#define SMMU_PFN_MASK 0x000fffff
 
-	writel(val, addr);
-}
+#define SMMU_PD_READABLE	(1 << 31)
+#define SMMU_PD_WRITABLE	(1 << 30)
+#define SMMU_PD_NONSECURE	(1 << 29)
 
-#define VA_PAGE_TO_PA(va, page)	\
-	(page_to_phys(page) + ((unsigned long)(va) & ~PAGE_MASK))
+#define SMMU_PDE_READABLE	(1 << 31)
+#define SMMU_PDE_WRITABLE	(1 << 30)
+#define SMMU_PDE_NONSECURE	(1 << 29)
+#define SMMU_PDE_NEXT		(1 << 28)
 
-#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)
+#define SMMU_PTE_READABLE	(1 << 31)
+#define SMMU_PTE_WRITABLE	(1 << 30)
+#define SMMU_PTE_NONSECURE	(1 << 29)
 
-/*
- * 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_PDE_ATTR		(SMMU_PDE_READABLE | SMMU_PDE_WRITABLE | \
+				 SMMU_PDE_NONSECURE)
+#define SMMU_PTE_ATTR		(SMMU_PTE_READABLE | SMMU_PTE_WRITABLE | \
+				 SMMU_PTE_NONSECURE)
 
-#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)
+static inline void smmu_flush_ptc(struct tegra_smmu *smmu, struct page *page,
+				  unsigned long offset)
 {
-	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;
+	phys_addr_t phys = page ? page_to_phys(page) : 0;
+	u32 value;
+
+	if (page) {
+		offset &= ~(smmu->mc->soc->atom_size - 1);
+
+		if (smmu->mc->soc->num_address_bits > 32) {
+#ifdef CONFIG_PHYS_ADDR_T_64BIT
+			value = (phys >> 32) & SMMU_PTC_FLUSH_HI_MASK;
+#else
+			value = 0;
+#endif
+			smmu_writel(smmu, value, SMMU_PTC_FLUSH_HI);
 		}
-		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);
+		value = (phys + offset) | SMMU_PTC_FLUSH_TYPE_ADR;
+	} else {
+		value = SMMU_PTC_FLUSH_TYPE_ALL;
 	}
-	return -EBUSY;
+
+	smmu_writel(smmu, value, SMMU_PTC_FLUSH);
 }
 
-static int smmu_client_set_hwgrp(struct smmu_client *c, u32 map, int on)
+static inline void smmu_flush_tlb(struct tegra_smmu *smmu)
 {
-	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;
+	smmu_writel(smmu, SMMU_TLB_FLUSH_VA_MATCH_ALL, SMMU_TLB_FLUSH);
 }
 
-/*
- * Flush all TLB entries and all PTC entries
- * Caller must lock smmu
- */
-static void smmu_flush_regs(struct smmu_device *smmu, int enable)
+static inline void smmu_flush_tlb_asid(struct tegra_smmu *smmu,
+				       unsigned long asid)
 {
-	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);
+	u32 value;
 
-	if (enable)
-		smmu_write(smmu, SMMU_CONFIG_ENABLE, SMMU_CONFIG);
-	FLUSH_SMMU_REGS(smmu);
+	value = SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_ASID(asid) |
+		SMMU_TLB_FLUSH_VA_MATCH_ALL;
+	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
 }
 
-static int smmu_setup_regs(struct smmu_device *smmu)
+static inline void smmu_flush_tlb_section(struct tegra_smmu *smmu,
+					  unsigned long asid,
+					  unsigned long iova)
 {
-	int i;
-	u32 val;
+	u32 value;
 
-	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_CACHE_CONFIG(_TLB));
-	smmu_write(smmu, SMMU_PTC_CONFIG_RESET_VAL, SMMU_CACHE_CONFIG(_PTC));
-
-	smmu_flush_regs(smmu, 1);
-
-	return tegra_ahb_enable_smmu(smmu->ahb);
+	value = SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_ASID(asid) |
+		SMMU_TLB_FLUSH_VA_SECTION(iova);
+	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
 }
 
-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)
+static inline void smmu_flush_tlb_group(struct tegra_smmu *smmu,
+					unsigned long asid,
+					unsigned long iova)
 {
-	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);
-}
+	u32 value;
 
-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);
-	}
+	value = SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_ASID(asid) |
+		SMMU_TLB_FLUSH_VA_GROUP(iova);
+	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
 }
 
-static void free_pdir(struct smmu_as *as)
+static inline void smmu_flush(struct tegra_smmu *smmu)
 {
-	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;
+	smmu_readl(smmu, SMMU_CONFIG);
 }
 
-/*
- * 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)
+static int tegra_smmu_alloc_asid(struct tegra_smmu *smmu, unsigned int *idp)
 {
-	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);
+	unsigned long id;
 
-		/* Vacant - allocate a new page table */
-		dev_dbg(as->smmu->dev, "New PTBL pdn: %lx\n", pdn);
+	mutex_lock(&smmu->lock);
 
-		*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);
+	id = find_first_zero_bit(smmu->asids, smmu->soc->num_asids);
+	if (id >= smmu->soc->num_asids) {
+		mutex_unlock(&smmu->lock);
+		return -ENOSPC;
 	}
-	*count = &as->pte_count[pdn];
 
-	return &ptbl[ptn % SMMU_PTBL_COUNT];
+	set_bit(id, smmu->asids);
+	*idp = id;
+
+	mutex_unlock(&smmu->lock);
+	return 0;
 }
 
-#ifdef CONFIG_SMMU_SIG_DEBUG
-static void put_signature(struct smmu_as *as,
-			  dma_addr_t iova, unsigned long pfn)
+static void tegra_smmu_free_asid(struct tegra_smmu *smmu, unsigned int id)
 {
-	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);
+	mutex_lock(&smmu->lock);
+	clear_bit(id, smmu->asids);
+	mutex_unlock(&smmu->lock);
 }
-#else
-static inline void put_signature(struct smmu_as *as,
-				 unsigned long addr, unsigned long pfn)
+
+static bool tegra_smmu_capable(enum iommu_cap cap)
 {
+	return false;
 }
-#endif
 
-/*
- * Caller must not hold as->lock
- */
-static int alloc_pdir(struct smmu_as *as)
+static int tegra_smmu_domain_init(struct iommu_domain *domain)
 {
-	unsigned long *pdir, flags;
-	int pdn, err = 0;
-	u32 val;
-	struct smmu_device *smmu = as->smmu;
-	struct page *page;
-	unsigned int *cnt;
+	struct tegra_smmu_as *as;
+	unsigned int i;
+	uint32_t *pd;
 
-	/*
-	 * do the allocation, then grab as->lock
-	 */
-	cnt = devm_kzalloc(smmu->dev,
-			   sizeof(cnt[0]) * SMMU_PDIR_COUNT,
-			   GFP_KERNEL);
-	page = alloc_page(GFP_KERNEL | __GFP_DMA);
+	as = kzalloc(sizeof(*as), GFP_KERNEL);
+	if (!as)
+		return -ENOMEM;
 
-	spin_lock_irqsave(&as->lock, flags);
+	as->attr = SMMU_PD_READABLE | SMMU_PD_WRITABLE | SMMU_PD_NONSECURE;
+	as->domain = domain;
 
-	if (as->pdir_page) {
-		/* We raced, free the redundant */
-		err = -EAGAIN;
-		goto err_out;
+	as->pd = alloc_page(GFP_KERNEL | __GFP_DMA);
+	if (!as->pd) {
+		kfree(as);
+		return -ENOMEM;
 	}
 
-	if (!page || !cnt) {
-		dev_err(smmu->dev, "failed to allocate at %s\n", __func__);
-		err = -ENOMEM;
-		goto err_out;
+	as->count = alloc_page(GFP_KERNEL);
+	if (!as->count) {
+		__free_page(as->pd);
+		kfree(as);
+		return -ENOMEM;
 	}
 
-	as->pdir_page = page;
-	as->pte_count = cnt;
+	/* clear PDEs */
+	pd = page_address(as->pd);
+	SetPageReserved(as->pd);
 
-	SetPageReserved(as->pdir_page);
-	pdir = page_address(as->pdir_page);
+	for (i = 0; i < SMMU_NUM_PDE; i++)
+		pd[i] = 0;
 
-	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);
+	/* clear PDE usage counters */
+	pd = page_address(as->count);
+	SetPageReserved(as->count);
 
-	spin_unlock_irqrestore(&as->lock, flags);
-
-	return 0;
+	for (i = 0; i < SMMU_NUM_PDE; i++)
+		pd[i] = 0;
 
-err_out:
-	spin_unlock_irqrestore(&as->lock, flags);
+	domain->priv = as;
 
-	devm_kfree(smmu->dev, cnt);
-	if (page)
-		__free_page(page);
-	return err;
+	return 0;
 }
 
-static void __smmu_iommu_unmap(struct smmu_as *as, dma_addr_t iova)
+static void tegra_smmu_domain_destroy(struct iommu_domain *domain)
 {
-	unsigned long *pte;
-	struct page *page;
-	unsigned int *count;
+	struct tegra_smmu_as *as = domain->priv;
 
-	pte = locate_pte(as, iova, false, &page, &count);
-	if (WARN_ON(!pte))
-		return;
+	/* TODO: free page directory and page tables */
+	ClearPageReserved(as->pd);
 
-	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);
+	kfree(as);
 }
 
-static void __smmu_iommu_map_pfn(struct smmu_as *as, dma_addr_t iova,
-				 unsigned long pfn)
+static const struct tegra_smmu_swgroup *
+tegra_smmu_find_swgroup(struct tegra_smmu *smmu, unsigned int swgroup)
 {
-	struct smmu_device *smmu = as->smmu;
-	unsigned long *pte;
-	unsigned int *count;
-	struct page *page;
+	const struct tegra_smmu_swgroup *group = NULL;
+	unsigned int i;
 
-	pte = locate_pte(as, iova, true, &page, &count);
-	if (WARN_ON(!pte))
-		return;
+	for (i = 0; i < smmu->soc->num_swgroups; i++) {
+		if (smmu->soc->swgroups[i].swgroup == swgroup) {
+			group = &smmu->soc->swgroups[i];
+			break;
+		}
+	}
 
-	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);
+	return group;
 }
 
-static int smmu_iommu_map(struct iommu_domain *domain, unsigned long iova,
-			  phys_addr_t pa, size_t bytes, int prot)
+static void tegra_smmu_enable(struct tegra_smmu *smmu, unsigned int swgroup,
+			      unsigned int asid)
 {
-	struct smmu_as *as = domain->priv;
-	unsigned long pfn = __phys_to_pfn(pa);
-	unsigned long flags;
+	const struct tegra_smmu_swgroup *group;
+	unsigned int i;
+	u32 value;
 
-	dev_dbg(as->smmu->dev, "[%d] %08lx:%pa\n", as->asid, iova, &pa);
+	for (i = 0; i < smmu->soc->num_clients; i++) {
+		const struct tegra_mc_client *client = &smmu->soc->clients[i];
 
-	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;
+		if (client->swgroup != swgroup)
+			continue;
 
-	dev_dbg(as->smmu->dev, "[%d] %08lx\n", as->asid, iova);
+		value = smmu_readl(smmu, client->smmu.reg);
+		value |= BIT(client->smmu.bit);
+		smmu_writel(smmu, value, client->smmu.reg);
+	}
 
-	spin_lock_irqsave(&as->lock, flags);
-	__smmu_iommu_unmap(as, iova);
-	spin_unlock_irqrestore(&as->lock, flags);
-	return SMMU_PAGE_SIZE;
+	group = tegra_smmu_find_swgroup(smmu, swgroup);
+	if (group) {
+		value = smmu_readl(smmu, group->reg);
+		value &= ~SMMU_ASID_MASK;
+		value |= SMMU_ASID_VALUE(asid);
+		value |= SMMU_ASID_ENABLE;
+		smmu_writel(smmu, value, group->reg);
+	}
 }
 
-static phys_addr_t smmu_iommu_iova_to_phys(struct iommu_domain *domain,
-					   dma_addr_t iova)
+static void tegra_smmu_disable(struct tegra_smmu *smmu, unsigned int swgroup,
+			       unsigned int asid)
 {
-	struct smmu_as *as = domain->priv;
-	unsigned long *pte;
-	unsigned int *count;
-	struct page *page;
-	unsigned long pfn;
-	unsigned long flags;
+	const struct tegra_smmu_swgroup *group;
+	unsigned int i;
+	u32 value;
 
-	spin_lock_irqsave(&as->lock, flags);
+	group = tegra_smmu_find_swgroup(smmu, swgroup);
+	if (group) {
+		value = smmu_readl(smmu, group->reg);
+		value &= ~SMMU_ASID_MASK;
+		value |= SMMU_ASID_VALUE(asid);
+		value &= ~SMMU_ASID_ENABLE;
+		smmu_writel(smmu, value, group->reg);
+	}
 
-	pte = locate_pte(as, iova, true, &page, &count);
-	pfn = *pte & SMMU_PFN_MASK;
-	WARN_ON(!pfn_valid(pfn));
-	dev_dbg(as->smmu->dev,
-		"iova:%08llx pfn:%08lx asid:%d\n", (unsigned long long)iova,
-		 pfn, as->asid);
+	for (i = 0; i < smmu->soc->num_clients; i++) {
+		const struct tegra_mc_client *client = &smmu->soc->clients[i];
 
-	spin_unlock_irqrestore(&as->lock, flags);
-	return PFN_PHYS(pfn);
-}
+		if (client->swgroup != swgroup)
+			continue;
 
-static bool smmu_iommu_capable(enum iommu_cap cap)
-{
-	return false;
+		value = smmu_readl(smmu, client->smmu.reg);
+		value &= ~BIT(client->smmu.bit);
+		smmu_writel(smmu, value, client->smmu.reg);
+	}
 }
 
-static int smmu_iommu_attach_dev(struct iommu_domain *domain,
-				 struct device *dev)
+static int tegra_smmu_as_prepare(struct tegra_smmu *smmu,
+				 struct tegra_smmu_as *as)
 {
-	struct smmu_as *as = domain->priv;
-	struct smmu_device *smmu = as->smmu;
-	struct smmu_client *client, *c;
-	u32 map;
+	u32 value;
 	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;
-		}
+	if (as->use_count > 0) {
+		as->use_count++;
+		return 0;
 	}
-	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;
+	err = tegra_smmu_alloc_asid(smmu, &as->id);
+	if (err < 0)
+		return err;
 
-		page = as->smmu->avp_vector_page;
-		__smmu_iommu_map_pfn(as, 0, page_to_pfn(page));
+	smmu->soc->ops->flush_dcache(as->pd, 0, SMMU_SIZE_PD);
+	smmu_flush_ptc(smmu, as->pd, 0);
+	smmu_flush_tlb_asid(smmu, as->id);
 
-		pr_info("Reserve \"page zero\" for AVP vectors using a common dummy\n");
-	}
+	smmu_writel(smmu, as->id & 0x7f, SMMU_PTB_ASID);
+	value = SMMU_PTB_DATA_VALUE(as->pd, as->attr);
+	smmu_writel(smmu, value, SMMU_PTB_DATA);
+	smmu_flush(smmu);
 
-	dev_dbg(smmu->dev, "%s is attached\n", dev_name(dev));
-	return 0;
+	as->smmu = smmu;
+	as->use_count++;
 
-err_client:
-	smmu_client_disable_hwgrp(client);
-	spin_unlock(&as->client_lock);
-err_hwgrp:
-	devm_kfree(smmu->dev, client);
-	return err;
+	return 0;
 }
 
-static void smmu_iommu_detach_dev(struct iommu_domain *domain,
-				  struct device *dev)
+static void tegra_smmu_as_unprepare(struct tegra_smmu *smmu,
+				    struct tegra_smmu_as *as)
 {
-	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(dev));
-out:
-	spin_unlock(&as->client_lock);
+	if (--as->use_count > 0)
+		return;
+
+	tegra_smmu_free_asid(smmu, as->id);
+	as->smmu = NULL;
 }
 
-static int smmu_iommu_domain_init(struct iommu_domain *domain)
+static int tegra_smmu_attach_dev(struct iommu_domain *domain,
+				 struct device *dev)
 {
-	int i, err = -EAGAIN;
-	unsigned long flags;
-	struct smmu_as *as;
-	struct smmu_device *smmu = smmu_handle;
+	struct tegra_smmu *smmu = dev->archdata.iommu;
+	struct tegra_smmu_as *as = domain->priv;
+	struct device_node *np = dev->of_node;
+	struct of_phandle_args args;
+	unsigned int index = 0;
+	int err = 0;
 
-	/* Look for a free AS with lock held */
-	for  (i = 0; i < smmu->num_as; i++) {
-		as = &smmu->as[i];
+	while (!of_parse_phandle_with_args(np, "iommus", "#iommu-cells", index,
+					   &args)) {
+		unsigned int swgroup = args.args[0];
 
-		if (as->pdir_page)
+		if (args.np != smmu->dev->of_node) {
+			of_node_put(args.np);
 			continue;
+		}
 
-		err = alloc_pdir(as);
-		if (!err)
-			goto found;
+		of_node_put(args.np);
 
-		if (err != -EAGAIN)
-			break;
+		err = tegra_smmu_as_prepare(smmu, as);
+		if (err < 0)
+			return err;
+
+		tegra_smmu_enable(smmu, swgroup, as->id);
+		index++;
 	}
-	if (i == smmu->num_as)
-		dev_err(smmu->dev,  "no free AS\n");
-	return err;
 
-found:
-	spin_lock_irqsave(&smmu->lock, flags);
+	if (index == 0)
+		return -ENODEV;
 
-	/* 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);
+	return 0;
+}
 
-	spin_unlock_irqrestore(&smmu->lock, flags);
+static void tegra_smmu_detach_dev(struct iommu_domain *domain, struct device *dev)
+{
+	struct tegra_smmu_as *as = domain->priv;
+	struct device_node *np = dev->of_node;
+	struct tegra_smmu *smmu = as->smmu;
+	struct of_phandle_args args;
+	unsigned int index = 0;
 
-	domain->priv = as;
+	while (!of_parse_phandle_with_args(np, "iommus", "#iommu-cells", index,
+					   &args)) {
+		unsigned int swgroup = args.args[0];
 
-	domain->geometry.aperture_start = smmu->iovmm_base;
-	domain->geometry.aperture_end   = smmu->iovmm_base +
-		smmu->page_count * SMMU_PAGE_SIZE - 1;
-	domain->geometry.force_aperture = true;
+		if (args.np != smmu->dev->of_node) {
+			of_node_put(args.np);
+			continue;
+		}
 
-	dev_dbg(smmu->dev, "smmu_as@%p\n", as);
+		of_node_put(args.np);
 
-	return 0;
+		tegra_smmu_disable(smmu, swgroup, as->id);
+		tegra_smmu_as_unprepare(smmu, as);
+		index++;
+	}
 }
 
-static void smmu_iommu_domain_destroy(struct iommu_domain *domain)
+static u32 *as_get_pte(struct tegra_smmu_as *as, dma_addr_t iova,
+		       struct page **pagep)
 {
-	struct smmu_as *as = domain->priv;
-	struct smmu_device *smmu = as->smmu;
-	unsigned long flags;
+	u32 *pd = page_address(as->pd), *pt, *count;
+	u32 pde = (iova >> SMMU_PDE_SHIFT) & 0x3ff;
+	u32 pte = (iova >> SMMU_PTE_SHIFT) & 0x3ff;
+	struct tegra_smmu *smmu = as->smmu;
+	struct page *page;
+	unsigned int i;
+
+	if (pd[pde] == 0) {
+		page = alloc_page(GFP_KERNEL | __GFP_DMA);
+		if (!page)
+			return NULL;
 
-	spin_lock_irqsave(&as->lock, flags);
+		pt = page_address(page);
+		SetPageReserved(page);
 
-	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);
+		for (i = 0; i < SMMU_NUM_PTE; i++)
+			pt[i] = 0;
 
-		free_pdir(as);
-	}
+		smmu->soc->ops->flush_dcache(page, 0, SMMU_SIZE_PT);
 
-	if (!list_empty(&as->client)) {
-		struct smmu_client *c;
+		pd[pde] = SMMU_MK_PDE(page, SMMU_PDE_ATTR | SMMU_PDE_NEXT);
 
-		list_for_each_entry(c, &as->client, list)
-			smmu_iommu_detach_dev(domain, c->dev);
+		smmu->soc->ops->flush_dcache(as->pd, pde << 2, 4);
+		smmu_flush_ptc(smmu, as->pd, pde << 2);
+		smmu_flush_tlb_section(smmu, as->id, iova);
+		smmu_flush(smmu);
+	} else {
+		page = pfn_to_page(pd[pde] & SMMU_PFN_MASK);
+		pt = page_address(page);
 	}
 
-	spin_unlock_irqrestore(&as->lock, flags);
+	*pagep = page;
 
-	domain->priv = NULL;
-	dev_dbg(smmu->dev, "smmu_as@%p\n", as);
-}
+	/* Keep track of entries in this page table. */
+	count = page_address(as->count);
+	if (pt[pte] == 0)
+		count[pde]++;
 
-static const struct iommu_ops smmu_iommu_ops = {
-	.capable	= smmu_iommu_capable,
-	.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,
-	.map_sg		= default_iommu_map_sg,
-	.iova_to_phys	= smmu_iommu_iova_to_phys,
-	.pgsize_bitmap	= SMMU_IOMMU_PGSIZES,
-};
-
-/* Should be in the order of enum */
-static const char * const smmu_debugfs_mc[] = { "mc", };
-static const char * const smmu_debugfs_cache[] = {  "tlb", "ptc", };
+	return &pt[pte];
+}
 
-static ssize_t smmu_debugfs_stats_write(struct file *file,
-					const char __user *buffer,
-					size_t count, loff_t *pos)
+static void as_put_pte(struct tegra_smmu_as *as, dma_addr_t iova)
 {
-	struct smmu_debugfs_info *info;
-	struct smmu_device *smmu;
-	int i;
-	enum {
-		_OFF = 0,
-		_ON,
-		_RESET,
-	};
-	const char * const command[] = {
-		[_OFF]		= "off",
-		[_ON]		= "on",
-		[_RESET]	= "reset",
-	};
-	char str[] = "reset";
-	u32 val;
-	size_t offs;
+	u32 pde = (iova >> SMMU_PDE_SHIFT) & 0x3ff;
+	u32 pte = (iova >> SMMU_PTE_SHIFT) & 0x3ff;
+	u32 *count = page_address(as->count);
+	u32 *pd = page_address(as->pd), *pt;
+	struct page *page;
 
-	count = min_t(size_t, count, sizeof(str));
-	if (copy_from_user(str, buffer, count))
-		return -EINVAL;
+	page = pfn_to_page(pd[pde] & SMMU_PFN_MASK);
+	pt = page_address(page);
 
-	for (i = 0; i < ARRAY_SIZE(command); i++)
-		if (strncmp(str, command[i],
-			    strlen(command[i])) == 0)
-			break;
+	/*
+	 * When no entries in this page table are used anymore, return the
+	 * memory page to the system.
+	 */
+	if (pt[pte] != 0) {
+		if (--count[pde] == 0) {
+			ClearPageReserved(page);
+			__free_page(page);
+			pd[pde] = 0;
+		}
 
-	if (i == ARRAY_SIZE(command))
-		return -EINVAL;
-
-	info = file_inode(file)->i_private;
-	smmu = info->smmu;
-
-	offs = SMMU_CACHE_CONFIG(info->cache);
-	val = smmu_read(smmu, offs);
-	switch (i) {
-	case _OFF:
-		val &= ~SMMU_CACHE_CONFIG_STATS_ENABLE;
-		val &= ~SMMU_CACHE_CONFIG_STATS_TEST;
-		smmu_write(smmu, val, offs);
-		break;
-	case _ON:
-		val |= SMMU_CACHE_CONFIG_STATS_ENABLE;
-		val &= ~SMMU_CACHE_CONFIG_STATS_TEST;
-		smmu_write(smmu, val, offs);
-		break;
-	case _RESET:
-		val |= SMMU_CACHE_CONFIG_STATS_TEST;
-		smmu_write(smmu, val, offs);
-		val &= ~SMMU_CACHE_CONFIG_STATS_TEST;
-		smmu_write(smmu, val, offs);
-		break;
-	default:
-		BUG();
-		break;
+		pt[pte] = 0;
 	}
-
-	dev_dbg(smmu->dev, "%s() %08x, %08x @%08x\n", __func__,
-		val, smmu_read(smmu, offs), offs);
-
-	return count;
 }
 
-static int smmu_debugfs_stats_show(struct seq_file *s, void *v)
+static int tegra_smmu_map(struct iommu_domain *domain, unsigned long iova,
+			  phys_addr_t paddr, size_t size, int prot)
 {
-	struct smmu_debugfs_info *info = s->private;
-	struct smmu_device *smmu = info->smmu;
-	int i;
-	const char * const stats[] = { "hit", "miss", };
+	struct tegra_smmu_as *as = domain->priv;
+	struct tegra_smmu *smmu = as->smmu;
+	unsigned long offset;
+	struct page *page;
+	u32 *pte;
 
+	pte = as_get_pte(as, iova, &page);
+	if (!pte)
+		return -ENOMEM;
 
-	for (i = 0; i < ARRAY_SIZE(stats); i++) {
-		u32 val;
-		size_t offs;
+	*pte = __phys_to_pfn(paddr) | SMMU_PTE_ATTR;
+	offset = offset_in_page(pte);
 
-		offs = SMMU_STATS_CACHE_COUNT(info->mc, info->cache, i);
-		val = smmu_read(smmu, offs);
-		seq_printf(s, "%s:%08x ", stats[i], val);
+	smmu->soc->ops->flush_dcache(page, offset, 4);
+	smmu_flush_ptc(smmu, page, offset);
+	smmu_flush_tlb_group(smmu, as->id, iova);
+	smmu_flush(smmu);
 
-		dev_dbg(smmu->dev, "%s() %s %08x @%08x\n", __func__,
-			stats[i], val, offs);
-	}
-	seq_printf(s, "\n");
 	return 0;
 }
 
-static int smmu_debugfs_stats_open(struct inode *inode, struct file *file)
+static size_t tegra_smmu_unmap(struct iommu_domain *domain, unsigned long iova,
+			       size_t size)
 {
-	return single_open(file, smmu_debugfs_stats_show, inode->i_private);
-}
+	struct tegra_smmu_as *as = domain->priv;
+	struct tegra_smmu *smmu = as->smmu;
+	unsigned long offset;
+	struct page *page;
+	u32 *pte;
 
-static const struct file_operations smmu_debugfs_stats_fops = {
-	.open		= smmu_debugfs_stats_open,
-	.read		= seq_read,
-	.llseek		= seq_lseek,
-	.release	= single_release,
-	.write		= smmu_debugfs_stats_write,
-};
+	pte = as_get_pte(as, iova, &page);
+	if (!pte)
+		return 0;
 
-static void smmu_debugfs_delete(struct smmu_device *smmu)
-{
-	debugfs_remove_recursive(smmu->debugfs_root);
-	kfree(smmu->debugfs_info);
+	offset = offset_in_page(pte);
+	as_put_pte(as, iova);
+
+	smmu->soc->ops->flush_dcache(page, offset, 4);
+	smmu_flush_ptc(smmu, page, offset);
+	smmu_flush_tlb_group(smmu, as->id, iova);
+	smmu_flush(smmu);
+
+	return size;
 }
 
-static void smmu_debugfs_create(struct smmu_device *smmu)
+static phys_addr_t tegra_smmu_iova_to_phys(struct iommu_domain *domain,
+					   dma_addr_t iova)
 {
-	int i;
-	size_t bytes;
-	struct dentry *root;
-
-	bytes = ARRAY_SIZE(smmu_debugfs_mc) * ARRAY_SIZE(smmu_debugfs_cache) *
-		sizeof(*smmu->debugfs_info);
-	smmu->debugfs_info = kmalloc(bytes, GFP_KERNEL);
-	if (!smmu->debugfs_info)
-		return;
-
-	root = debugfs_create_dir(dev_name(smmu->dev), NULL);
-	if (!root)
-		goto err_out;
-	smmu->debugfs_root = root;
-
-	for (i = 0; i < ARRAY_SIZE(smmu_debugfs_mc); i++) {
-		int j;
-		struct dentry *mc;
-
-		mc = debugfs_create_dir(smmu_debugfs_mc[i], root);
-		if (!mc)
-			goto err_out;
-
-		for (j = 0; j < ARRAY_SIZE(smmu_debugfs_cache); j++) {
-			struct dentry *cache;
-			struct smmu_debugfs_info *info;
-
-			info = smmu->debugfs_info;
-			info += i * ARRAY_SIZE(smmu_debugfs_mc) + j;
-			info->smmu = smmu;
-			info->mc = i;
-			info->cache = j;
-
-			cache = debugfs_create_file(smmu_debugfs_cache[j],
-						    S_IWUGO | S_IRUGO, mc,
-						    (void *)info,
-						    &smmu_debugfs_stats_fops);
-			if (!cache)
-				goto err_out;
-		}
-	}
+	struct tegra_smmu_as *as = domain->priv;
+	struct page *page;
+	unsigned long pfn;
+	u32 *pte;
 
-	return;
+	pte = as_get_pte(as, iova, &page);
+	pfn = *pte & SMMU_PFN_MASK;
 
-err_out:
-	smmu_debugfs_delete(smmu);
+	return PFN_PHYS(pfn);
 }
 
-static int tegra_smmu_suspend(struct device *dev)
+static struct tegra_smmu *tegra_smmu_find(struct device_node *np)
 {
-	struct smmu_device *smmu = dev_get_drvdata(dev);
+	struct platform_device *pdev;
+	struct tegra_mc *mc;
 
-	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;
+	pdev = of_find_device_by_node(np);
+	if (!pdev)
+		return NULL;
+
+	mc = platform_get_drvdata(pdev);
+	if (!mc)
+		return NULL;
+
+	return mc->smmu;
 }
 
-static int tegra_smmu_resume(struct device *dev)
+static int tegra_smmu_add_device(struct device *dev)
 {
-	struct smmu_device *smmu = dev_get_drvdata(dev);
-	unsigned long flags;
-	int err;
+	struct device_node *np = dev->of_node;
+	struct of_phandle_args args;
+	unsigned int index = 0;
 
-	spin_lock_irqsave(&smmu->lock, flags);
-	err = smmu_setup_regs(smmu);
-	spin_unlock_irqrestore(&smmu->lock, flags);
-	return err;
+	while (of_parse_phandle_with_args(np, "iommus", "#iommu-cells", index,
+					  &args) == 0) {
+		struct tegra_smmu *smmu;
+
+		smmu = tegra_smmu_find(args.np);
+		if (smmu) {
+			/*
+			 * Only a single IOMMU master interface is currently
+			 * supported by the Linux kernel, so abort after the
+			 * first match.
+			 */
+			dev->archdata.iommu = smmu;
+			break;
+		}
+
+		index++;
+	}
+
+	return 0;
 }
 
-static int tegra_smmu_probe(struct platform_device *pdev)
+static void tegra_smmu_remove_device(struct device *dev)
 {
-	struct smmu_device *smmu;
-	struct device *dev = &pdev->dev;
-	int i, asids, err = 0;
-	dma_addr_t uninitialized_var(base);
-	size_t bytes, uninitialized_var(size);
+	dev->archdata.iommu = NULL;
+}
 
-	if (smmu_handle)
-		return -EIO;
+static const struct iommu_ops tegra_smmu_ops = {
+	.capable = tegra_smmu_capable,
+	.domain_init = tegra_smmu_domain_init,
+	.domain_destroy = tegra_smmu_domain_destroy,
+	.attach_dev = tegra_smmu_attach_dev,
+	.detach_dev = tegra_smmu_detach_dev,
+	.add_device = tegra_smmu_add_device,
+	.remove_device = tegra_smmu_remove_device,
+	.map = tegra_smmu_map,
+	.unmap = tegra_smmu_unmap,
+	.map_sg = default_iommu_map_sg,
+	.iova_to_phys = tegra_smmu_iova_to_phys,
 
-	BUILD_BUG_ON(PAGE_SHIFT != SMMU_PAGE_SHIFT);
+	.pgsize_bitmap = SZ_4K,
+};
 
-	if (of_property_read_u32(dev->of_node, "nvidia,#asids", &asids))
-		return -ENODEV;
+static void tegra_smmu_ahb_enable(void)
+{
+	static const struct of_device_id ahb_match[] = {
+		{ .compatible = "nvidia,tegra30-ahb", },
+		{ }
+	};
+	struct device_node *ahb;
 
-	bytes = sizeof(*smmu) + asids * sizeof(*smmu->as);
-	smmu = devm_kzalloc(dev, bytes, GFP_KERNEL);
-	if (!smmu) {
-		dev_err(dev, "failed to allocate smmu_device\n");
-		return -ENOMEM;
+	ahb = of_find_matching_node(NULL, ahb_match);
+	if (ahb) {
+		tegra_ahb_enable_smmu(ahb);
+		of_node_put(ahb);
 	}
+}
 
-	smmu->nregs = pdev->num_resources;
-	smmu->regs = devm_kzalloc(dev, 2 * smmu->nregs * sizeof(*smmu->regs),
-				  GFP_KERNEL);
-	smmu->rege = smmu->regs + smmu->nregs;
-	if (!smmu->regs)
-		return -ENOMEM;
-	for (i = 0; i < smmu->nregs; i++) {
-		struct resource *res;
-
-		res = platform_get_resource(pdev, IORESOURCE_MEM, i);
-		smmu->regs[i] = devm_ioremap_resource(&pdev->dev, res);
-		if (IS_ERR(smmu->regs[i]))
-			return PTR_ERR(smmu->regs[i]);
-		smmu->rege[i] = smmu->regs[i] + resource_size(res) - 1;
-	}
-	/* Same as "mc" 1st regiter block start address */
-	smmu->regbase = (void __iomem *)((u32)smmu->regs[0] & PAGE_MASK);
+struct tegra_smmu *tegra_smmu_probe(struct device *dev,
+				    const struct tegra_smmu_soc *soc,
+				    struct tegra_mc *mc)
+{
+	struct tegra_smmu *smmu;
+	size_t size;
+	u32 value;
+	int err;
 
-	err = of_get_dma_window(dev->of_node, NULL, 0, NULL, &base, &size);
-	if (err)
-		return -ENODEV;
+	/* This can happen on Tegra20 which doesn't have an SMMU */
+	if (!soc)
+		return NULL;
 
-	if (size & SMMU_PAGE_MASK)
-		return -EINVAL;
+	smmu = devm_kzalloc(dev, sizeof(*smmu), GFP_KERNEL);
+	if (!smmu)
+		return ERR_PTR(-ENOMEM);
 
-	size >>= SMMU_PAGE_SHIFT;
-	if (!size)
-		return -EINVAL;
+	/*
+	 * This is a bit of a hack. Ideally we'd want to simply return this
+	 * value. However the IOMMU registration process will attempt to add
+	 * all devices to the IOMMU when bus_set_iommu() is called. In order
+	 * not to rely on global variables to track the IOMMU instance, we
+	 * set it here so that it can be looked up from the .add_device()
+	 * callback via the IOMMU device's .drvdata field.
+	 */
+	mc->smmu = smmu;
 
-	smmu->ahb = of_parse_phandle(dev->of_node, "nvidia,ahb", 0);
-	if (!smmu->ahb)
-		return -ENODEV;
+	size = BITS_TO_LONGS(soc->num_asids) * sizeof(long);
 
-	smmu->dev = dev;
-	smmu->num_as = asids;
-	smmu->iovmm_base = base;
-	smmu->page_count = size;
-
-	smmu->translation_enable_0 = ~0;
-	smmu->translation_enable_1 = ~0;
-	smmu->translation_enable_2 = ~0;
-	smmu->asid_security = 0;
-
-	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);
-		spin_lock_init(&as->client_lock);
-		INIT_LIST_HEAD(&as->client);
-	}
-	spin_lock_init(&smmu->lock);
-	err = smmu_setup_regs(smmu);
-	if (err)
-		return err;
-	platform_set_drvdata(pdev, smmu);
+	smmu->asids = devm_kzalloc(dev, size, GFP_KERNEL);
+	if (!smmu->asids)
+		return ERR_PTR(-ENOMEM);
 
-	smmu->avp_vector_page = alloc_page(GFP_KERNEL);
-	if (!smmu->avp_vector_page)
-		return -ENOMEM;
+	mutex_init(&smmu->lock);
 
-	smmu_debugfs_create(smmu);
-	smmu_handle = smmu;
-	bus_set_iommu(&platform_bus_type, &smmu_iommu_ops);
-	return 0;
-}
+	smmu->regs = mc->regs;
+	smmu->soc = soc;
+	smmu->dev = dev;
+	smmu->mc = mc;
 
-static int tegra_smmu_remove(struct platform_device *pdev)
-{
-	struct smmu_device *smmu = platform_get_drvdata(pdev);
-	int i;
+	value = SMMU_PTC_CONFIG_ENABLE | SMMU_PTC_CONFIG_INDEX_MAP(0x3f);
 
-	smmu_debugfs_delete(smmu);
+	if (soc->supports_request_limit)
+		value |= SMMU_PTC_CONFIG_REQ_LIMIT(8);
 
-	smmu_write(smmu, SMMU_CONFIG_DISABLE, SMMU_CONFIG);
-	for (i = 0; i < smmu->num_as; i++)
-		free_pdir(&smmu->as[i]);
-	__free_page(smmu->avp_vector_page);
-	smmu_handle = NULL;
-	return 0;
-}
+	smmu_writel(smmu, value, SMMU_PTC_CONFIG);
 
-static const struct dev_pm_ops tegra_smmu_pm_ops = {
-	.suspend	= tegra_smmu_suspend,
-	.resume		= tegra_smmu_resume,
-};
+	value = SMMU_TLB_CONFIG_HIT_UNDER_MISS |
+		SMMU_TLB_CONFIG_ACTIVE_LINES(0x20);
 
-static const struct of_device_id tegra_smmu_of_match[] = {
-	{ .compatible = "nvidia,tegra30-smmu", },
-	{ },
-};
-MODULE_DEVICE_TABLE(of, tegra_smmu_of_match);
-
-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,
-		.of_match_table = tegra_smmu_of_match,
-	},
-};
+	if (soc->supports_round_robin_arbitration)
+		value |= SMMU_TLB_CONFIG_ROUND_ROBIN_ARBITRATION;
 
-static int tegra_smmu_init(void)
-{
-	return platform_driver_register(&tegra_smmu_driver);
-}
+	smmu_writel(smmu, value, SMMU_TLB_CONFIG);
 
-static void __exit tegra_smmu_exit(void)
-{
-	platform_driver_unregister(&tegra_smmu_driver);
-}
+	smmu_flush_ptc(smmu, NULL, 0);
+	smmu_flush_tlb(smmu);
+	smmu_writel(smmu, SMMU_CONFIG_ENABLE, SMMU_CONFIG);
+	smmu_flush(smmu);
+
+	tegra_smmu_ahb_enable();
 
-subsys_initcall(tegra_smmu_init);
-module_exit(tegra_smmu_exit);
+	err = bus_set_iommu(&platform_bus_type, &tegra_smmu_ops);
+	if (err < 0)
+		return ERR_PTR(err);
 
-MODULE_DESCRIPTION("IOMMU API for SMMU in Tegra30");
-MODULE_AUTHOR("Hiroshi DOYU <hdoyu@nvidia.com>");
-MODULE_ALIAS("platform:tegra-smmu");
-MODULE_LICENSE("GPL v2");
+	return smmu;
+}