diff --git a/arch/arm/include/asm/hardware/cache-l2x0.h b/arch/arm/include/asm/hardware/cache-l2x0.h
index cc42d5fdee17..5aeec1e1735c 100644
--- a/arch/arm/include/asm/hardware/cache-l2x0.h
+++ b/arch/arm/include/asm/hardware/cache-l2x0.h
@@ -59,7 +59,17 @@
 #define L2X0_CACHE_ID_PART_MASK		(0xf << 6)
 #define L2X0_CACHE_ID_PART_L210		(1 << 6)
 #define L2X0_CACHE_ID_PART_L310		(3 << 6)
-#define L2X0_AUX_CTRL_WAY_SIZE_MASK	(0x3 << 17)
+
+#define L2X0_AUX_CTRL_MASK			0xc0000fff
+#define L2X0_AUX_CTRL_ASSOCIATIVITY_SHIFT	16
+#define L2X0_AUX_CTRL_WAY_SIZE_SHIFT		17
+#define L2X0_AUX_CTRL_WAY_SIZE_MASK		(0x3 << 17)
+#define L2X0_AUX_CTRL_SHARE_OVERRIDE_SHIFT	22
+#define L2X0_AUX_CTRL_NS_LOCKDOWN_SHIFT		26
+#define L2X0_AUX_CTRL_NS_INT_CTRL_SHIFT		27
+#define L2X0_AUX_CTRL_DATA_PREFETCH_SHIFT	28
+#define L2X0_AUX_CTRL_INSTR_PREFETCH_SHIFT	29
+#define L2X0_AUX_CTRL_EARLY_BRESP_SHIFT		30
 
 #ifndef __ASSEMBLY__
 extern void __init l2x0_init(void __iomem *base, __u32 aux_val, __u32 aux_mask);
diff --git a/arch/arm/include/asm/io.h b/arch/arm/include/asm/io.h
index 815efa2d4e07..20e0f7c9e03e 100644
--- a/arch/arm/include/asm/io.h
+++ b/arch/arm/include/asm/io.h
@@ -241,18 +241,15 @@ extern void _memset_io(volatile void __iomem *, int, size_t);
  *
  */
 #ifndef __arch_ioremap
-#define ioremap(cookie,size)		__arm_ioremap(cookie, size, MT_DEVICE)
-#define ioremap_nocache(cookie,size)	__arm_ioremap(cookie, size, MT_DEVICE)
-#define ioremap_cached(cookie,size)	__arm_ioremap(cookie, size, MT_DEVICE_CACHED)
-#define ioremap_wc(cookie,size)		__arm_ioremap(cookie, size, MT_DEVICE_WC)
-#define iounmap(cookie)			__iounmap(cookie)
-#else
+#define __arch_ioremap			__arm_ioremap
+#define __arch_iounmap			__iounmap
+#endif
+
 #define ioremap(cookie,size)		__arch_ioremap((cookie), (size), MT_DEVICE)
 #define ioremap_nocache(cookie,size)	__arch_ioremap((cookie), (size), MT_DEVICE)
 #define ioremap_cached(cookie,size)	__arch_ioremap((cookie), (size), MT_DEVICE_CACHED)
 #define ioremap_wc(cookie,size)		__arch_ioremap((cookie), (size), MT_DEVICE_WC)
-#define iounmap(cookie)			__arch_iounmap(cookie)
-#endif
+#define iounmap				__arch_iounmap
 
 /*
  * io{read,write}{8,16,32} macros
diff --git a/arch/arm/mach-davinci/include/mach/io.h b/arch/arm/mach-davinci/include/mach/io.h
index 62b0a90309ad..d1b954955c12 100644
--- a/arch/arm/mach-davinci/include/mach/io.h
+++ b/arch/arm/mach-davinci/include/mach/io.h
@@ -22,8 +22,8 @@
 #define __mem_isa(a)		(a)
 
 #ifndef __ASSEMBLER__
-#define __arch_ioremap(p, s, t)	davinci_ioremap(p, s, t)
-#define __arch_iounmap(v)	davinci_iounmap(v)
+#define __arch_ioremap		davinci_ioremap
+#define __arch_iounmap		davinci_iounmap
 
 void __iomem *davinci_ioremap(unsigned long phys, size_t size,
 			      unsigned int type);
diff --git a/arch/arm/mach-iop13xx/include/mach/io.h b/arch/arm/mach-iop13xx/include/mach/io.h
index a6e0f9e6ddcf..dffb234bb967 100644
--- a/arch/arm/mach-iop13xx/include/mach/io.h
+++ b/arch/arm/mach-iop13xx/include/mach/io.h
@@ -35,7 +35,7 @@ extern u32 iop13xx_atux_mem_base;
 extern size_t iop13xx_atue_mem_size;
 extern size_t iop13xx_atux_mem_size;
 
-#define __arch_ioremap(a, s, f) __iop13xx_ioremap(a, s, f)
-#define __arch_iounmap(a)	 __iop13xx_iounmap(a)
+#define __arch_ioremap	__iop13xx_ioremap
+#define __arch_iounmap	__iop13xx_iounmap
 
 #endif
diff --git a/arch/arm/mach-iop32x/include/mach/io.h b/arch/arm/mach-iop32x/include/mach/io.h
index 339e5854728b..059c783ce0b2 100644
--- a/arch/arm/mach-iop32x/include/mach/io.h
+++ b/arch/arm/mach-iop32x/include/mach/io.h
@@ -21,7 +21,7 @@ extern void __iop3xx_iounmap(void __iomem *addr);
 #define __io(p)		((void __iomem *)IOP3XX_PCI_IO_PHYS_TO_VIRT(p))
 #define __mem_pci(a)		(a)
 
-#define __arch_ioremap(a, s, f) __iop3xx_ioremap(a, s, f)
-#define __arch_iounmap(a)	 __iop3xx_iounmap(a)
+#define __arch_ioremap	__iop3xx_ioremap
+#define __arch_iounmap	__iop3xx_iounmap
 
 #endif
diff --git a/arch/arm/mach-iop33x/include/mach/io.h b/arch/arm/mach-iop33x/include/mach/io.h
index e99a7ed6d050..39e893e97c21 100644
--- a/arch/arm/mach-iop33x/include/mach/io.h
+++ b/arch/arm/mach-iop33x/include/mach/io.h
@@ -21,7 +21,7 @@ extern void __iop3xx_iounmap(void __iomem *addr);
 #define __io(p)		((void __iomem *)IOP3XX_PCI_IO_PHYS_TO_VIRT(p))
 #define __mem_pci(a)		(a)
 
-#define __arch_ioremap(a, s, f) __iop3xx_ioremap(a, s, f)
-#define __arch_iounmap(a)	 __iop3xx_iounmap(a)
+#define __arch_ioremap	__iop3xx_ioremap
+#define __arch_iounmap	__iop3xx_iounmap
 
 #endif
diff --git a/arch/arm/mach-ixp23xx/include/mach/io.h b/arch/arm/mach-ixp23xx/include/mach/io.h
index fd9ef8e519f7..a1749d0fd896 100644
--- a/arch/arm/mach-ixp23xx/include/mach/io.h
+++ b/arch/arm/mach-ixp23xx/include/mach/io.h
@@ -45,8 +45,8 @@ ixp23xx_iounmap(void __iomem *addr)
 	__iounmap(addr);
 }
 
-#define __arch_ioremap(a,s,f)	ixp23xx_ioremap(a,s,f)
-#define __arch_iounmap(a)	ixp23xx_iounmap(a)
+#define __arch_ioremap	ixp23xx_ioremap
+#define __arch_iounmap	ixp23xx_iounmap
 
 
 #endif
diff --git a/arch/arm/mach-ixp4xx/include/mach/io.h b/arch/arm/mach-ixp4xx/include/mach/io.h
index de274a1f19d7..57b5410c31f4 100644
--- a/arch/arm/mach-ixp4xx/include/mach/io.h
+++ b/arch/arm/mach-ixp4xx/include/mach/io.h
@@ -74,8 +74,8 @@ static inline void __indirect_iounmap(void __iomem *addr)
 		__iounmap(addr);
 }
 
-#define __arch_ioremap(a, s, f)		__indirect_ioremap(a, s, f)
-#define __arch_iounmap(a)		__indirect_iounmap(a)
+#define __arch_ioremap			__indirect_ioremap
+#define __arch_iounmap			__indirect_iounmap
 
 #define writeb(v, p)			__indirect_writeb(v, p)
 #define writew(v, p)			__indirect_writew(v, p)
diff --git a/arch/arm/mach-kirkwood/include/mach/io.h b/arch/arm/mach-kirkwood/include/mach/io.h
index 44e8be04f259..1aaddc364f2e 100644
--- a/arch/arm/mach-kirkwood/include/mach/io.h
+++ b/arch/arm/mach-kirkwood/include/mach/io.h
@@ -42,8 +42,8 @@ __arch_iounmap(void __iomem *addr)
 		__iounmap(addr);
 }
 
-#define __arch_ioremap(p, s, m)	__arch_ioremap(p, s, m)
-#define __arch_iounmap(a)	__arch_iounmap(a)
+#define __arch_ioremap		__arch_ioremap
+#define __arch_iounmap		__arch_iounmap
 #define __io(a)			__io(a)
 #define __mem_pci(a)		(a)
 
diff --git a/arch/arm/mach-omap2/mailbox.c b/arch/arm/mach-omap2/mailbox.c
index 40ddecab93a9..394413dc7deb 100644
--- a/arch/arm/mach-omap2/mailbox.c
+++ b/arch/arm/mach-omap2/mailbox.c
@@ -281,7 +281,7 @@ static struct omap_mbox_ops omap2_mbox_ops = {
 
 /* FIXME: the following structs should be filled automatically by the user id */
 
-#if defined(CONFIG_ARCH_OMAP3430) || defined(CONFIG_ARCH_OMAP2420)
+#if defined(CONFIG_ARCH_OMAP3) || defined(CONFIG_ARCH_OMAP2)
 /* DSP */
 static struct omap_mbox2_priv omap2_mbox_dsp_priv = {
 	.tx_fifo = {
@@ -306,7 +306,7 @@ struct omap_mbox mbox_dsp_info = {
 };
 #endif
 
-#if defined(CONFIG_ARCH_OMAP3430)
+#if defined(CONFIG_ARCH_OMAP3)
 struct omap_mbox *omap3_mboxes[] = { &mbox_dsp_info, NULL };
 #endif
 
@@ -394,15 +394,19 @@ static int __devinit omap2_mbox_probe(struct platform_device *pdev)
 
 	if (false)
 		;
-#if defined(CONFIG_ARCH_OMAP3430)
-	else if (cpu_is_omap3430()) {
+#if defined(CONFIG_ARCH_OMAP3)
+	else if (cpu_is_omap34xx()) {
 		list = omap3_mboxes;
 
 		list[0]->irq = platform_get_irq_byname(pdev, "dsp");
 	}
 #endif
-#if defined(CONFIG_ARCH_OMAP2420)
-	else if (cpu_is_omap2420()) {
+#if defined(CONFIG_ARCH_OMAP2)
+	else if (cpu_is_omap2430()) {
+		list = omap2_mboxes;
+
+		list[0]->irq = platform_get_irq_byname(pdev, "dsp");
+	} else if (cpu_is_omap2420()) {
 		list = omap2_mboxes;
 
 		list[0]->irq = platform_get_irq_byname(pdev, "dsp");
@@ -432,9 +436,8 @@ static int __devinit omap2_mbox_probe(struct platform_device *pdev)
 		iounmap(mbox_base);
 		return ret;
 	}
-	return 0;
 
-	return ret;
+	return 0;
 }
 
 static int __devexit omap2_mbox_remove(struct platform_device *pdev)
diff --git a/arch/arm/mach-omap2/omap-iommu.c b/arch/arm/mach-omap2/omap-iommu.c
index f5a1aad1a5c0..3fc5dc7233da 100644
--- a/arch/arm/mach-omap2/omap-iommu.c
+++ b/arch/arm/mach-omap2/omap-iommu.c
@@ -33,9 +33,11 @@ static struct iommu_device omap3_devices[] = {
 			.name = "isp",
 			.nr_tlb_entries = 8,
 			.clk_name = "cam_ick",
+			.da_start = 0x0,
+			.da_end = 0xFFFFF000,
 		},
 	},
-#if defined(CONFIG_MPU_BRIDGE_IOMMU)
+#if defined(CONFIG_OMAP_IOMMU_IVA2)
 	{
 		.base = 0x5d000000,
 		.irq = 28,
@@ -43,6 +45,8 @@ static struct iommu_device omap3_devices[] = {
 			.name = "iva2",
 			.nr_tlb_entries = 32,
 			.clk_name = "iva2_ck",
+			.da_start = 0x11000000,
+			.da_end = 0xFFFFF000,
 		},
 	},
 #endif
@@ -64,6 +68,8 @@ static struct iommu_device omap4_devices[] = {
 			.name = "ducati",
 			.nr_tlb_entries = 32,
 			.clk_name = "ducati_ick",
+			.da_start = 0x0,
+			.da_end = 0xFFFFF000,
 		},
 	},
 #if defined(CONFIG_MPU_TESLA_IOMMU)
@@ -74,6 +80,8 @@ static struct iommu_device omap4_devices[] = {
 			.name = "tesla",
 			.nr_tlb_entries = 32,
 			.clk_name = "tesla_ick",
+			.da_start = 0x0,
+			.da_end = 0xFFFFF000,
 		},
 	},
 #endif
diff --git a/arch/arm/mach-omap2/omap4-common.c b/arch/arm/mach-omap2/omap4-common.c
index 2f895553e6a8..e7a9b7f13b52 100644
--- a/arch/arm/mach-omap2/omap4-common.c
+++ b/arch/arm/mach-omap2/omap4-common.c
@@ -53,6 +53,8 @@ static void omap4_l2x0_disable(void)
 
 static int __init omap_l2_cache_init(void)
 {
+	u32 aux_ctrl = 0;
+
 	/*
 	 * To avoid code running on other OMAPs in
 	 * multi-omap builds
@@ -64,18 +66,32 @@ static int __init omap_l2_cache_init(void)
 	l2cache_base = ioremap(OMAP44XX_L2CACHE_BASE, SZ_4K);
 	BUG_ON(!l2cache_base);
 
-	/* Enable PL310 L2 Cache controller */
-	omap_smc1(0x102, 0x1);
-
 	/*
 	 * 16-way associativity, parity disabled
 	 * Way size - 32KB (es1.0)
 	 * Way size - 64KB (es2.0 +)
 	 */
-	if (omap_rev() == OMAP4430_REV_ES1_0)
-		l2x0_init(l2cache_base, 0x0e050000, 0xc0000fff);
-	else
-		l2x0_init(l2cache_base, 0x0e070000, 0xc0000fff);
+	aux_ctrl = ((1 << L2X0_AUX_CTRL_ASSOCIATIVITY_SHIFT) |
+			(0x1 << 25) |
+			(0x1 << L2X0_AUX_CTRL_NS_LOCKDOWN_SHIFT) |
+			(0x1 << L2X0_AUX_CTRL_NS_INT_CTRL_SHIFT));
+
+	if (omap_rev() == OMAP4430_REV_ES1_0) {
+		aux_ctrl |= 0x2 << L2X0_AUX_CTRL_WAY_SIZE_SHIFT;
+	} else {
+		aux_ctrl |= ((0x3 << L2X0_AUX_CTRL_WAY_SIZE_SHIFT) |
+			(1 << L2X0_AUX_CTRL_SHARE_OVERRIDE_SHIFT) |
+			(1 << L2X0_AUX_CTRL_DATA_PREFETCH_SHIFT) |
+			(1 << L2X0_AUX_CTRL_INSTR_PREFETCH_SHIFT) |
+			(1 << L2X0_AUX_CTRL_EARLY_BRESP_SHIFT));
+	}
+	if (omap_rev() != OMAP4430_REV_ES1_0)
+		omap_smc1(0x109, aux_ctrl);
+
+	/* Enable PL310 L2 Cache controller */
+	omap_smc1(0x102, 0x1);
+
+	l2x0_init(l2cache_base, aux_ctrl, L2X0_AUX_CTRL_MASK);
 
 	/*
 	 * Override default outer_cache.disable with a OMAP4
diff --git a/arch/arm/mach-orion5x/include/mach/io.h b/arch/arm/mach-orion5x/include/mach/io.h
index c47b033bd999..c5196101a237 100644
--- a/arch/arm/mach-orion5x/include/mach/io.h
+++ b/arch/arm/mach-orion5x/include/mach/io.h
@@ -38,8 +38,8 @@ __arch_iounmap(void __iomem *addr)
 		__iounmap(addr);
 }
 
-#define __arch_ioremap(p, s, m)	__arch_ioremap(p, s, m)
-#define __arch_iounmap(a)	__arch_iounmap(a)
+#define __arch_ioremap		__arch_ioremap
+#define __arch_iounmap		__arch_iounmap
 #define __io(a)			__typesafe_io(a)
 #define __mem_pci(a)		(a)
 
diff --git a/arch/arm/mach-tegra/include/mach/io.h b/arch/arm/mach-tegra/include/mach/io.h
index f0981b1ac59e..4cea2230c8dc 100644
--- a/arch/arm/mach-tegra/include/mach/io.h
+++ b/arch/arm/mach-tegra/include/mach/io.h
@@ -65,8 +65,8 @@
 
 #ifndef __ASSEMBLER__
 
-#define __arch_ioremap(p, s, t)	tegra_ioremap(p, s, t)
-#define __arch_iounmap(v)	tegra_iounmap(v)
+#define __arch_ioremap		tegra_ioremap
+#define __arch_iounmap		tegra_iounmap
 
 void __iomem *tegra_ioremap(unsigned long phys, size_t size, unsigned int type);
 void tegra_iounmap(volatile void __iomem *addr);
diff --git a/arch/arm/plat-omap/Kconfig b/arch/arm/plat-omap/Kconfig
index 92c5bb7909f5..5e63e5069e0d 100644
--- a/arch/arm/plat-omap/Kconfig
+++ b/arch/arm/plat-omap/Kconfig
@@ -109,6 +109,9 @@ config OMAP_IOMMU_DEBUG
 
          Say N unless you know you need this.
 
+config OMAP_IOMMU_IVA2
+	bool
+
 choice
 	prompt "System timer"
 	default OMAP_32K_TIMER if !ARCH_OMAP15XX
diff --git a/arch/arm/plat-omap/include/plat/io.h b/arch/arm/plat-omap/include/plat/io.h
index 128b549c2796..204865f91d93 100644
--- a/arch/arm/plat-omap/include/plat/io.h
+++ b/arch/arm/plat-omap/include/plat/io.h
@@ -294,8 +294,8 @@ static inline void omap44xx_map_common_io(void)
 extern void omap2_init_common_hw(struct omap_sdrc_params *sdrc_cs0,
 				 struct omap_sdrc_params *sdrc_cs1);
 
-#define __arch_ioremap(p,s,t)	omap_ioremap(p,s,t)
-#define __arch_iounmap(v)	omap_iounmap(v)
+#define __arch_ioremap	omap_ioremap
+#define __arch_iounmap	omap_iounmap
 
 void __iomem *omap_ioremap(unsigned long phys, size_t size, unsigned int type);
 void omap_iounmap(volatile void __iomem *addr);
diff --git a/arch/arm/plat-omap/include/plat/iommu.h b/arch/arm/plat-omap/include/plat/iommu.h
index 33c7d41cb6a5..69230d685538 100644
--- a/arch/arm/plat-omap/include/plat/iommu.h
+++ b/arch/arm/plat-omap/include/plat/iommu.h
@@ -50,6 +50,8 @@ struct iommu {
 	int (*isr)(struct iommu *obj);
 
 	void *ctx; /* iommu context: registres saved area */
+	u32 da_start;
+	u32 da_end;
 };
 
 struct cr_regs {
@@ -103,6 +105,8 @@ struct iommu_platform_data {
 	const char *name;
 	const char *clk_name;
 	const int nr_tlb_entries;
+	u32 da_start;
+	u32 da_end;
 };
 
 #if defined(CONFIG_ARCH_OMAP1)
@@ -152,6 +156,7 @@ extern void flush_iotlb_all(struct iommu *obj);
 extern int iopgtable_store_entry(struct iommu *obj, struct iotlb_entry *e);
 extern size_t iopgtable_clear_entry(struct iommu *obj, u32 iova);
 
+extern int iommu_set_da_range(struct iommu *obj, u32 start, u32 end);
 extern struct iommu *iommu_get(const char *name);
 extern void iommu_put(struct iommu *obj);
 
diff --git a/arch/arm/plat-omap/include/plat/mailbox.h b/arch/arm/plat-omap/include/plat/mailbox.h
index 997656552109..cc3921e9059c 100644
--- a/arch/arm/plat-omap/include/plat/mailbox.h
+++ b/arch/arm/plat-omap/include/plat/mailbox.h
@@ -46,8 +46,8 @@ struct omap_mbox_queue {
 	struct kfifo		fifo;
 	struct work_struct	work;
 	struct tasklet_struct	tasklet;
-	int	(*callback)(void *);
 	struct omap_mbox	*mbox;
+	bool full;
 };
 
 struct omap_mbox {
@@ -57,13 +57,15 @@ struct omap_mbox {
 	struct omap_mbox_ops	*ops;
 	struct device		*dev;
 	void			*priv;
+	int			use_count;
+	struct blocking_notifier_head   notifier;
 };
 
 int omap_mbox_msg_send(struct omap_mbox *, mbox_msg_t msg);
 void omap_mbox_init_seq(struct omap_mbox *);
 
-struct omap_mbox *omap_mbox_get(const char *);
-void omap_mbox_put(struct omap_mbox *);
+struct omap_mbox *omap_mbox_get(const char *, struct notifier_block *nb);
+void omap_mbox_put(struct omap_mbox *mbox, struct notifier_block *nb);
 
 int omap_mbox_register(struct device *parent, struct omap_mbox **);
 int omap_mbox_unregister(void);
diff --git a/arch/arm/plat-omap/iommu.c b/arch/arm/plat-omap/iommu.c
index 6cd151b31bc5..b1107c08da56 100644
--- a/arch/arm/plat-omap/iommu.c
+++ b/arch/arm/plat-omap/iommu.c
@@ -829,6 +829,28 @@ static int device_match_by_alias(struct device *dev, void *data)
 	return strcmp(obj->name, name) == 0;
 }
 
+/**
+ * iommu_set_da_range - Set a valid device address range
+ * @obj:		target iommu
+ * @start		Start of valid range
+ * @end			End of valid range
+ **/
+int iommu_set_da_range(struct iommu *obj, u32 start, u32 end)
+{
+
+	if (!obj)
+		return -EFAULT;
+
+	if (end < start || !PAGE_ALIGN(start | end))
+		return -EINVAL;
+
+	obj->da_start = start;
+	obj->da_end = end;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(iommu_set_da_range);
+
 /**
  * iommu_get - Get iommu handler
  * @name:	target iommu name
@@ -922,6 +944,8 @@ static int __devinit omap_iommu_probe(struct platform_device *pdev)
 	obj->name = pdata->name;
 	obj->dev = &pdev->dev;
 	obj->ctx = (void *)obj + sizeof(*obj);
+	obj->da_start = pdata->da_start;
+	obj->da_end = pdata->da_end;
 
 	mutex_init(&obj->iommu_lock);
 	mutex_init(&obj->mmap_lock);
diff --git a/arch/arm/plat-omap/iovmm.c b/arch/arm/plat-omap/iovmm.c
index 8ce0de247c71..6dc1296c8c77 100644
--- a/arch/arm/plat-omap/iovmm.c
+++ b/arch/arm/plat-omap/iovmm.c
@@ -87,35 +87,43 @@ static size_t sgtable_len(const struct sg_table *sgt)
 }
 #define sgtable_ok(x)	(!!sgtable_len(x))
 
+static unsigned max_alignment(u32 addr)
+{
+	int i;
+	unsigned pagesize[] = { SZ_16M, SZ_1M, SZ_64K, SZ_4K, };
+	for (i = 0; i < ARRAY_SIZE(pagesize) && addr & (pagesize[i] - 1); i++)
+		;
+	return (i < ARRAY_SIZE(pagesize)) ? pagesize[i] : 0;
+}
+
 /*
  * calculate the optimal number sg elements from total bytes based on
  * iommu superpages
  */
-static unsigned int sgtable_nents(size_t bytes)
+static unsigned sgtable_nents(size_t bytes, u32 da, u32 pa)
 {
-	int i;
-	unsigned int nr_entries;
-	const unsigned long pagesize[] = { SZ_16M, SZ_1M, SZ_64K, SZ_4K, };
+	unsigned nr_entries = 0, ent_sz;
 
 	if (!IS_ALIGNED(bytes, PAGE_SIZE)) {
 		pr_err("%s: wrong size %08x\n", __func__, bytes);
 		return 0;
 	}
 
-	nr_entries = 0;
-	for (i = 0; i < ARRAY_SIZE(pagesize); i++) {
-		if (bytes >= pagesize[i]) {
-			nr_entries += (bytes / pagesize[i]);
-			bytes %= pagesize[i];
-		}
+	while (bytes) {
+		ent_sz = max_alignment(da | pa);
+		ent_sz = min_t(unsigned, ent_sz, iopgsz_max(bytes));
+		nr_entries++;
+		da += ent_sz;
+		pa += ent_sz;
+		bytes -= ent_sz;
 	}
-	BUG_ON(bytes);
 
 	return nr_entries;
 }
 
 /* allocate and initialize sg_table header(a kind of 'superblock') */
-static struct sg_table *sgtable_alloc(const size_t bytes, u32 flags)
+static struct sg_table *sgtable_alloc(const size_t bytes, u32 flags,
+							u32 da, u32 pa)
 {
 	unsigned int nr_entries;
 	int err;
@@ -127,9 +135,8 @@ static struct sg_table *sgtable_alloc(const size_t bytes, u32 flags)
 	if (!IS_ALIGNED(bytes, PAGE_SIZE))
 		return ERR_PTR(-EINVAL);
 
-	/* FIXME: IOVMF_DA_FIXED should support 'superpages' */
-	if ((flags & IOVMF_LINEAR) && (flags & IOVMF_DA_ANON)) {
-		nr_entries = sgtable_nents(bytes);
+	if (flags & IOVMF_LINEAR) {
+		nr_entries = sgtable_nents(bytes, da, pa);
 		if (!nr_entries)
 			return ERR_PTR(-EINVAL);
 	} else
@@ -273,13 +280,14 @@ static struct iovm_struct *alloc_iovm_area(struct iommu *obj, u32 da,
 	alignement = PAGE_SIZE;
 
 	if (flags & IOVMF_DA_ANON) {
-		/*
-		 * Reserve the first page for NULL
-		 */
-		start = PAGE_SIZE;
+		start = obj->da_start;
+
 		if (flags & IOVMF_LINEAR)
 			alignement = iopgsz_max(bytes);
 		start = roundup(start, alignement);
+	} else if (start < obj->da_start || start > obj->da_end ||
+					obj->da_end - start < bytes) {
+		return ERR_PTR(-EINVAL);
 	}
 
 	tmp = NULL;
@@ -289,19 +297,19 @@ static struct iovm_struct *alloc_iovm_area(struct iommu *obj, u32 da,
 	prev_end = 0;
 	list_for_each_entry(tmp, &obj->mmap, list) {
 
-		if (prev_end >= start)
+		if (prev_end > start)
 			break;
 
-		if (start + bytes < tmp->da_start)
+		if (tmp->da_start > start && (tmp->da_start - start) >= bytes)
 			goto found;
 
-		if (flags & IOVMF_DA_ANON)
+		if (tmp->da_end >= start && flags & IOVMF_DA_ANON)
 			start = roundup(tmp->da_end + 1, alignement);
 
 		prev_end = tmp->da_end;
 	}
 
-	if ((start > prev_end) && (ULONG_MAX - start >= bytes))
+	if ((start >= prev_end) && (obj->da_end - start >= bytes))
 		goto found;
 
 	dev_dbg(obj->dev, "%s: no space to fit %08x(%x) flags: %08x\n",
@@ -409,7 +417,8 @@ static inline void sgtable_drain_vmalloc(struct sg_table *sgt)
 	BUG_ON(!sgt);
 }
 
-static void sgtable_fill_kmalloc(struct sg_table *sgt, u32 pa, size_t len)
+static void sgtable_fill_kmalloc(struct sg_table *sgt, u32 pa, u32 da,
+								size_t len)
 {
 	unsigned int i;
 	struct scatterlist *sg;
@@ -418,9 +427,10 @@ static void sgtable_fill_kmalloc(struct sg_table *sgt, u32 pa, size_t len)
 	va = phys_to_virt(pa);
 
 	for_each_sg(sgt->sgl, sg, sgt->nents, i) {
-		size_t bytes;
+		unsigned bytes;
 
-		bytes = iopgsz_max(len);
+		bytes = max_alignment(da | pa);
+		bytes = min_t(unsigned, bytes, iopgsz_max(len));
 
 		BUG_ON(!iopgsz_ok(bytes));
 
@@ -429,6 +439,7 @@ static void sgtable_fill_kmalloc(struct sg_table *sgt, u32 pa, size_t len)
 		 * 'pa' is cotinuous(linear).
 		 */
 		pa += bytes;
+		da += bytes;
 		len -= bytes;
 	}
 	BUG_ON(len);
@@ -695,18 +706,18 @@ u32 iommu_vmalloc(struct iommu *obj, u32 da, size_t bytes, u32 flags)
 	if (!va)
 		return -ENOMEM;
 
-	sgt = sgtable_alloc(bytes, flags);
+	flags &= IOVMF_HW_MASK;
+	flags |= IOVMF_DISCONT;
+	flags |= IOVMF_ALLOC;
+	flags |= (da ? IOVMF_DA_FIXED : IOVMF_DA_ANON);
+
+	sgt = sgtable_alloc(bytes, flags, da, 0);
 	if (IS_ERR(sgt)) {
 		da = PTR_ERR(sgt);
 		goto err_sgt_alloc;
 	}
 	sgtable_fill_vmalloc(sgt, va);
 
-	flags &= IOVMF_HW_MASK;
-	flags |= IOVMF_DISCONT;
-	flags |= IOVMF_ALLOC;
-	flags |= (da ? IOVMF_DA_FIXED : IOVMF_DA_ANON);
-
 	da = __iommu_vmap(obj, da, sgt, va, bytes, flags);
 	if (IS_ERR_VALUE(da))
 		goto err_iommu_vmap;
@@ -746,11 +757,11 @@ static u32 __iommu_kmap(struct iommu *obj, u32 da, u32 pa, void *va,
 {
 	struct sg_table *sgt;
 
-	sgt = sgtable_alloc(bytes, flags);
+	sgt = sgtable_alloc(bytes, flags, da, pa);
 	if (IS_ERR(sgt))
 		return PTR_ERR(sgt);
 
-	sgtable_fill_kmalloc(sgt, pa, bytes);
+	sgtable_fill_kmalloc(sgt, pa, da, bytes);
 
 	da = map_iommu_region(obj, da, sgt, va, bytes, flags);
 	if (IS_ERR_VALUE(da)) {
@@ -811,7 +822,7 @@ void iommu_kunmap(struct iommu *obj, u32 da)
 	struct sg_table *sgt;
 	typedef void (*func_t)(const void *);
 
-	sgt = unmap_vm_area(obj, da, (func_t)__iounmap,
+	sgt = unmap_vm_area(obj, da, (func_t)iounmap,
 			    IOVMF_LINEAR | IOVMF_MMIO);
 	if (!sgt)
 		dev_dbg(obj->dev, "%s: No sgt\n", __func__);
diff --git a/arch/arm/plat-omap/mailbox.c b/arch/arm/plat-omap/mailbox.c
index d2fafb892f7f..459b319a9fad 100644
--- a/arch/arm/plat-omap/mailbox.c
+++ b/arch/arm/plat-omap/mailbox.c
@@ -28,12 +28,12 @@
 #include <linux/slab.h>
 #include <linux/kfifo.h>
 #include <linux/err.h>
+#include <linux/notifier.h>
 
 #include <plat/mailbox.h>
 
 static struct workqueue_struct *mboxd;
 static struct omap_mbox **mboxes;
-static bool rq_full;
 
 static int mbox_configured;
 static DEFINE_MUTEX(mbox_configured_lock);
@@ -93,20 +93,25 @@ int omap_mbox_msg_send(struct omap_mbox *mbox, mbox_msg_t msg)
 	struct omap_mbox_queue *mq = mbox->txq;
 	int ret = 0, len;
 
-	spin_lock(&mq->lock);
+	spin_lock_bh(&mq->lock);
 
 	if (kfifo_avail(&mq->fifo) < sizeof(msg)) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
+	if (kfifo_is_empty(&mq->fifo) && !__mbox_poll_for_space(mbox)) {
+		mbox_fifo_write(mbox, msg);
+		goto out;
+	}
+
 	len = kfifo_in(&mq->fifo, (unsigned char *)&msg, sizeof(msg));
 	WARN_ON(len != sizeof(msg));
 
 	tasklet_schedule(&mbox->txq->tasklet);
 
 out:
-	spin_unlock(&mq->lock);
+	spin_unlock_bh(&mq->lock);
 	return ret;
 }
 EXPORT_SYMBOL(omap_mbox_msg_send);
@@ -146,8 +151,14 @@ static void mbox_rx_work(struct work_struct *work)
 		len = kfifo_out(&mq->fifo, (unsigned char *)&msg, sizeof(msg));
 		WARN_ON(len != sizeof(msg));
 
-		if (mq->callback)
-			mq->callback((void *)msg);
+		blocking_notifier_call_chain(&mq->mbox->notifier, len,
+								(void *)msg);
+		spin_lock_irq(&mq->lock);
+		if (mq->full) {
+			mq->full = false;
+			omap_mbox_enable_irq(mq->mbox, IRQ_RX);
+		}
+		spin_unlock_irq(&mq->lock);
 	}
 }
 
@@ -170,7 +181,7 @@ static void __mbox_rx_interrupt(struct omap_mbox *mbox)
 	while (!mbox_fifo_empty(mbox)) {
 		if (unlikely(kfifo_avail(&mq->fifo) < sizeof(msg))) {
 			omap_mbox_disable_irq(mbox, IRQ_RX);
-			rq_full = true;
+			mq->full = true;
 			goto nomem;
 		}
 
@@ -239,73 +250,77 @@ static int omap_mbox_startup(struct omap_mbox *mbox)
 	int ret = 0;
 	struct omap_mbox_queue *mq;
 
-	if (mbox->ops->startup) {
-		mutex_lock(&mbox_configured_lock);
-		if (!mbox_configured)
+	mutex_lock(&mbox_configured_lock);
+	if (!mbox_configured++) {
+		if (likely(mbox->ops->startup)) {
 			ret = mbox->ops->startup(mbox);
+			if (unlikely(ret))
+				goto fail_startup;
+		} else
+			goto fail_startup;
+	}
 
-		if (ret) {
-			mutex_unlock(&mbox_configured_lock);
-			return ret;
+	if (!mbox->use_count++) {
+		ret = request_irq(mbox->irq, mbox_interrupt, IRQF_SHARED,
+							mbox->name, mbox);
+		if (unlikely(ret)) {
+			pr_err("failed to register mailbox interrupt:%d\n",
+									ret);
+			goto fail_request_irq;
 		}
-		mbox_configured++;
-		mutex_unlock(&mbox_configured_lock);
-	}
+		mq = mbox_queue_alloc(mbox, NULL, mbox_tx_tasklet);
+		if (!mq) {
+			ret = -ENOMEM;
+			goto fail_alloc_txq;
+		}
+		mbox->txq = mq;
 
-	ret = request_irq(mbox->irq, mbox_interrupt, IRQF_SHARED,
-				mbox->name, mbox);
-	if (ret) {
-		printk(KERN_ERR
-			"failed to register mailbox interrupt:%d\n", ret);
-		goto fail_request_irq;
+		mq = mbox_queue_alloc(mbox, mbox_rx_work, NULL);
+		if (!mq) {
+			ret = -ENOMEM;
+			goto fail_alloc_rxq;
+		}
+		mbox->rxq = mq;
+		mq->mbox = mbox;
 	}
-
-	mq = mbox_queue_alloc(mbox, NULL, mbox_tx_tasklet);
-	if (!mq) {
-		ret = -ENOMEM;
-		goto fail_alloc_txq;
-	}
-	mbox->txq = mq;
-
-	mq = mbox_queue_alloc(mbox, mbox_rx_work, NULL);
-	if (!mq) {
-		ret = -ENOMEM;
-		goto fail_alloc_rxq;
-	}
-	mbox->rxq = mq;
-
+	mutex_unlock(&mbox_configured_lock);
 	return 0;
 
- fail_alloc_rxq:
+fail_alloc_rxq:
 	mbox_queue_free(mbox->txq);
- fail_alloc_txq:
+fail_alloc_txq:
 	free_irq(mbox->irq, mbox);
- fail_request_irq:
+fail_request_irq:
 	if (mbox->ops->shutdown)
 		mbox->ops->shutdown(mbox);
-
+	mbox->use_count--;
+fail_startup:
+	mbox_configured--;
+	mutex_unlock(&mbox_configured_lock);
 	return ret;
 }
 
 static void omap_mbox_fini(struct omap_mbox *mbox)
 {
-	free_irq(mbox->irq, mbox);
-	tasklet_kill(&mbox->txq->tasklet);
-	flush_work(&mbox->rxq->work);
-	mbox_queue_free(mbox->txq);
-	mbox_queue_free(mbox->rxq);
+	mutex_lock(&mbox_configured_lock);
 
-	if (mbox->ops->shutdown) {
-		mutex_lock(&mbox_configured_lock);
-		if (mbox_configured > 0)
-			mbox_configured--;
-		if (!mbox_configured)
-			mbox->ops->shutdown(mbox);
-		mutex_unlock(&mbox_configured_lock);
+	if (!--mbox->use_count) {
+		free_irq(mbox->irq, mbox);
+		tasklet_kill(&mbox->txq->tasklet);
+		flush_work(&mbox->rxq->work);
+		mbox_queue_free(mbox->txq);
+		mbox_queue_free(mbox->rxq);
 	}
+
+	if (likely(mbox->ops->shutdown)) {
+		if (!--mbox_configured)
+			mbox->ops->shutdown(mbox);
+	}
+
+	mutex_unlock(&mbox_configured_lock);
 }
 
-struct omap_mbox *omap_mbox_get(const char *name)
+struct omap_mbox *omap_mbox_get(const char *name, struct notifier_block *nb)
 {
 	struct omap_mbox *mbox;
 	int ret;
@@ -324,12 +339,16 @@ struct omap_mbox *omap_mbox_get(const char *name)
 	if (ret)
 		return ERR_PTR(-ENODEV);
 
+	if (nb)
+		blocking_notifier_chain_register(&mbox->notifier, nb);
+
 	return mbox;
 }
 EXPORT_SYMBOL(omap_mbox_get);
 
-void omap_mbox_put(struct omap_mbox *mbox)
+void omap_mbox_put(struct omap_mbox *mbox, struct notifier_block *nb)
 {
+	blocking_notifier_chain_unregister(&mbox->notifier, nb);
 	omap_mbox_fini(mbox);
 }
 EXPORT_SYMBOL(omap_mbox_put);
@@ -353,6 +372,8 @@ int omap_mbox_register(struct device *parent, struct omap_mbox **list)
 			ret = PTR_ERR(mbox->dev);
 			goto err_out;
 		}
+
+		BLOCKING_INIT_NOTIFIER_HEAD(&mbox->notifier);
 	}
 	return 0;
 
@@ -391,7 +412,8 @@ static int __init omap_mbox_init(void)
 
 	/* kfifo size sanity check: alignment and minimal size */
 	mbox_kfifo_size = ALIGN(mbox_kfifo_size, sizeof(mbox_msg_t));
-	mbox_kfifo_size = max_t(unsigned int, mbox_kfifo_size, sizeof(mbox_msg_t));
+	mbox_kfifo_size = max_t(unsigned int, mbox_kfifo_size,
+							sizeof(mbox_msg_t));
 
 	return 0;
 }