crypto: tegra-aes: dual core support

* add bsea engine support for encryption and decryption
* add arbitration semaphore id for bsea

Bug 803932

Original change: http://git-master/r/#change,29672
(cherry picked from commit 0008cdb0f38d0cd0c074671fc067c4321f340b06)

Original-Change-Id: I59fcaab29c47a8b42e7470b30486851cfe90848f
Signed-off-by: Sanjay Singh Rawat <srawat@nvidia.com>
Reviewed-on: http://git-master/r/30190
Tested-by: Varun Wadekar <vwadekar@nvidia.com>
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>

Rebase-Id: R6f1bf287860a24d0a535e49f516581b31092d182

2 files changed, 514 insertions, 331 deletions

diff --git a/drivers/crypto/tegra-aes.c b/drivers/crypto/tegra-aes.c
index 285c8513b075..c691ee5969f7 100644
--- a/drivers/crypto/tegra-aes.c
+++ b/drivers/crypto/tegra-aes.c
@@ -36,6 +36,7 @@
 
 #include <mach/arb_sema.h>
 #include <mach/clk.h>
+#include "../video/tegra/nvmap/nvmap.h"
 
 #include <crypto/scatterwalk.h>
 #include <crypto/aes.h>
@@ -45,15 +46,12 @@
 
 #define FLAGS_MODE_MASK		0x00ff
 #define FLAGS_ENCRYPT		BIT(0)
-#define FLAGS_CBC			BIT(1)
-#define FLAGS_GIV			BIT(2)
-#define FLAGS_RNG			BIT(3)
-#define FLAGS_OFB			BIT(4)
-#define FLAGS_NEW_KEY			BIT(5)
-#define FLAGS_NEW_IV			BIT(6)
-#define FLAGS_INIT			BIT(7)
-#define FLAGS_FAST			BIT(8)
-#define FLAGS_BUSY			9
+#define FLAGS_CBC		BIT(1)
+#define FLAGS_GIV		BIT(2)
+#define FLAGS_RNG		BIT(3)
+#define FLAGS_OFB		BIT(4)
+#define FLAGS_INIT		BIT(5)
+#define FLAGS_BUSY		1
 
 /*
  * Defines AES engine Max process bytes size in one go, which takes 1 msec.
@@ -62,18 +60,18 @@
  * cycles of AVP/BSE is 216K. In this duration, AES can process 216/11 ~= 19KB
  * Based on this AES_HW_DMA_BUFFER_SIZE_BYTES is configured to 16KB.
  */
-#define AES_HW_DMA_BUFFER_SIZE_BYTES 0x4000
+#define AES_HW_DMA_BUFFER_SIZE_BYTES	0x4000
 
 /*
  * The key table length is 64 bytes
  * (This includes first upto 32 bytes key + 16 bytes original initial vector
  * and 16 bytes updated initial vector)
  */
-#define AES_HW_KEY_TABLE_LENGTH_BYTES 64
+#define AES_HW_KEY_TABLE_LENGTH_BYTES	64
 
-#define AES_HW_IV_SIZE 16
-#define AES_HW_KEYSCHEDULE_LEN 256
-#define ARB_SEMA_TIMEOUT 500
+#define AES_HW_IV_SIZE	16
+#define AES_HW_KEYSCHEDULE_LEN	256
+#define ARB_SEMA_TIMEOUT	500
 
 /*
  * The memory being used is divides as follows:
@@ -86,12 +84,12 @@
  */
 #define AES_IVKEY_SIZE (AES_HW_KEY_TABLE_LENGTH_BYTES + AES_HW_KEYSCHEDULE_LEN)
 
-#define DEFAULT_RNG_BLK_SZ 16
+#define DEFAULT_RNG_BLK_SZ	16
 
 /* As of now only 5 commands are USED for AES encryption/Decryption */
-#define AES_HW_MAX_ICQ_LENGTH 5
+#define AES_HW_MAX_ICQ_LENGTH	4
 
-#define ICQBITSHIFT_BLKCNT 0
+#define ICQBITSHIFT_BLKCNT	0
 
 /* memdma_vd command */
 #define MEMDMA_DIR_DTOVRAM	0
@@ -136,120 +134,196 @@ struct tegra_aes_slot {
 	bool available;
 };
 
-static struct tegra_aes_slot ssk = {
-	.slot_num = SSK_SLOT_NUM,
-	.available = true,
-};
-
 struct tegra_aes_reqctx {
 	unsigned long mode;
 };
 
-#define TEGRA_AES_QUEUE_LENGTH 50
+#define TEGRA_AES_QUEUE_LENGTH	50
 
-struct tegra_aes_dev {
-	struct device *dev;
-	phys_addr_t phys_base;
-	void __iomem *io_base;
-	dma_addr_t ivkey_phys_base;
-	void __iomem *ivkey_base;
+struct tegra_aes_engine {
+	struct tegra_aes_dev *dd;
 	struct clk *iclk;
 	struct clk *pclk;
-	struct tegra_aes_ctx *ctx;
-	unsigned long flags;
+	struct ablkcipher_request *req;
+	struct scatterlist *in_sg;
 	struct completion op_complete;
-	u32 *buf_in;
+	struct scatterlist *out_sg;
+	void __iomem *io_base;
+	void __iomem *ivkey_base;
+	unsigned long phys_base;
+	dma_addr_t ivkey_phys_base;
 	dma_addr_t dma_buf_in;
-	u32 *buf_out;
 	dma_addr_t dma_buf_out;
-	u8 *iv;
-	u8 dt[DEFAULT_RNG_BLK_SZ];
-	int ivlen;
-	u64 ctr;
-	int res_id;
-	spinlock_t lock;
-	struct crypto_queue queue;
-	struct tegra_aes_slot *slots;
-	struct ablkcipher_request *req;
 	size_t total;
-	struct scatterlist *in_sg;
 	size_t in_offset;
-	struct scatterlist *out_sg;
 	size_t out_offset;
+	u32 engine_offset;
+	u32 *buf_in;
+	u32 *buf_out;
+	int res_id;
+	int slot_num;
+	int keylen;
+	unsigned long busy;
+	u8 irq;
+	bool new_key;
+	bool use_ssk;
+};
+
+struct tegra_aes_dev {
+	struct device *dev;
+	struct tegra_aes_slot *slots;
+	struct tegra_aes_engine bsev;
+	struct tegra_aes_engine bsea;
+	struct nvmap_client *client;
+	struct nvmap_handle_ref *h_ref;
+	unsigned long bsea_iram_address;
+	void *bsea_iram_base;
+	struct tegra_aes_ctx *ctx;
+	struct crypto_queue queue;
+	spinlock_t lock;
+	u64 ctr;
+	unsigned long flags;
+	u8 dt[DEFAULT_RNG_BLK_SZ];
 };
 
 static struct tegra_aes_dev *aes_dev;
 
 struct tegra_aes_ctx {
 	struct tegra_aes_dev *dd;
-	unsigned long flags;
 	struct tegra_aes_slot *slot;
-	u8 key[AES_MAX_KEY_SIZE];
-	int keylen;
 };
 
-static struct tegra_aes_ctx rng_ctx = {
-	.flags = FLAGS_NEW_KEY,
-	.keylen = AES_KEYSIZE_128,
-};
+static struct tegra_aes_ctx rng_ctx;
 
 /* keep registered devices data here */
-static LIST_HEAD(dev_list);
+static LIST_HEAD(slot_list);
 static DEFINE_SPINLOCK(list_lock);
 static DEFINE_MUTEX(aes_lock);
 
-static void aes_workqueue_handler(struct work_struct *work);
-static DECLARE_WORK(aes_work, aes_workqueue_handler);
-static struct workqueue_struct *aes_wq;
+/* Engine specific work queues */
+static void bsev_workqueue_handler(struct work_struct *work);
+static void bsea_workqueue_handler(struct work_struct *work);
+
+static DECLARE_WORK(bsev_work, bsev_workqueue_handler);
+static DECLARE_WORK(bsea_work, bsea_workqueue_handler);
+
+static struct workqueue_struct *bsev_wq;
+static struct workqueue_struct *bsea_wq;
 
 extern unsigned long long tegra_chip_uid(void);
 
-static inline u32 aes_readl(struct tegra_aes_dev *dd, u32 offset)
+static inline u32 aes_readl(struct tegra_aes_engine *engine, u32 offset)
 {
-	return readl(dd->io_base + offset);
+	return readl(engine->io_base + offset);
 }
 
-static inline void aes_writel(struct tegra_aes_dev *dd, u32 val, u32 offset)
+static inline void aes_writel(struct tegra_aes_engine *engine,
+	u32 val, u32 offset)
 {
-	writel(val, dd->io_base + offset);
+	writel(val, engine->io_base + offset);
 }
 
-static int aes_hw_init(struct tegra_aes_dev *dd)
+static int bsea_alloc_iram(struct tegra_aes_dev *dd)
 {
-	int ret = 0;
+	size_t size, align ;
+	int err;
+
+	dd->h_ref = NULL;
+	/* [key+iv+u-iv=64B] * 8 = 512Bytes */
+	size = align = (AES_HW_KEY_TABLE_LENGTH_BYTES * AES_NR_KEYSLOTS);
+	dd->client = nvmap_create_client(nvmap_dev, "aes_bsea");
+	if (IS_ERR(dd->client)) {
+		dev_err(dd->dev, "nvmap_create_client failed\n");
+		goto out;
+	}
 
-	ret = clk_enable(dd->pclk);
-	if (ret < 0) {
-		dev_err(dd->dev, "%s: pclock enable fail(%d)\n", __func__, ret);
-		return ret;
+	dd->h_ref = nvmap_create_handle(dd->client, size);
+	if (IS_ERR(dd->h_ref)) {
+		dev_err(dd->dev, "nvmap_create_handle failed\n");
+		goto out;
 	}
 
-	ret = clk_enable(dd->iclk);
-	if (ret < 0) {
-		dev_err(dd->dev, "%s: iclock enable fail(%d)\n", __func__, ret);
-		clk_disable(dd->pclk);
-		return ret;
+	/* Allocate memory in the iram */
+	err = nvmap_alloc_handle_id(dd->client, nvmap_ref_to_id(dd->h_ref),
+		NVMAP_HEAP_CARVEOUT_IRAM, align, 0);
+	if (err) {
+		dev_err(dd->dev, "nvmap_alloc_handle_id failed\n");
+		nvmap_free_handle_id(dd->client, nvmap_ref_to_id(dd->h_ref));
+		goto out;
+	}
+	dd->bsea_iram_address = nvmap_handle_address(dd->client,
+				nvmap_ref_to_id(dd->h_ref));
+
+	dd->bsea_iram_base = nvmap_mmap(dd->h_ref);	/* get virtual address */
+	if (!dd->bsea_iram_base) {
+		dev_err(dd->dev, "%s: no mem, BSEA IRAM alloc failure\n",
+		__func__);
+		goto out;
+	}
+	memset(dd->bsea_iram_base, 0, dd->h_ref->handle->size);
+
+	return 0;
+out:
+	if (dd->bsea_iram_base)
+		nvmap_munmap(dd->h_ref, dd->bsea_iram_base);
+	if (dd->client)
+		nvmap_client_put(dd->client);
+	return -ENOMEM;
+}
+
+static void bsea_free_iram(struct tegra_aes_dev *dd)
+{
+	if (dd->bsea_iram_base)
+		nvmap_munmap(dd->h_ref, dd->bsea_iram_base);
+	if (dd->client)
+		nvmap_client_put(dd->client);
+}
+
+static int aes_hw_init(struct tegra_aes_engine *engine)
+{
+	struct tegra_aes_dev *dd = aes_dev;
+	int ret = 0;
+
+	if (engine->pclk) {
+		ret = clk_enable(engine->pclk);
+		if (ret < 0) {
+			dev_err(dd->dev, "%s: pclock enable fail(%d)\n",
+			__func__, ret);
+			return ret;
+		}
+	}
+	if (engine->iclk) {
+		ret = clk_enable(engine->iclk);
+		if (ret < 0) {
+			dev_err(dd->dev, "%s: iclock enable fail(%d)\n",
+			__func__, ret);
+			clk_disable(engine->pclk);
+			return ret;
+		}
 	}
 
 	return ret;
 }
 
-static void aes_hw_deinit(struct tegra_aes_dev *dd)
+static void aes_hw_deinit(struct tegra_aes_engine *engine)
 {
-	clk_disable(dd->iclk);
-	clk_disable(dd->pclk);
+	if (engine->pclk)
+		clk_disable(engine->pclk);
+
+	if (engine->iclk)
+		clk_disable(engine->iclk);
 }
 
-static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
-	int nblocks, int mode, bool upd_iv)
+static int aes_start_crypt(struct tegra_aes_engine *eng, u32 in_addr,
+	u32 out_addr, int nblocks, int mode, bool upd_iv)
 {
 	u32 cmdq[AES_HW_MAX_ICQ_LENGTH];
 	int qlen = 0, i, eng_busy, icq_empty, ret;
 	u32 value;
 
 	/* error, dma xfer complete */
-	aes_writel(dd, 0x33, INT_ENB);
-	enable_irq(INT_VDE_BSE_V);
+	aes_writel(eng, 0x33, INT_ENB);
+	enable_irq(eng->irq);
 
 	cmdq[qlen++] = UCQOPCODE_DMASETUP << ICQBITSHIFT_OPCODE;
 	cmdq[qlen++] = in_addr;
@@ -257,19 +331,17 @@ static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
 		(nblocks-1) << ICQBITSHIFT_BLKCNT;
 	cmdq[qlen++] = UCQOPCODE_DMACOMPLETE << ICQBITSHIFT_OPCODE;
 
-	value = aes_readl(dd, CMDQUE_CONTROL);
+	value = aes_readl(eng, CMDQUE_CONTROL);
 	/* access SDRAM through AHB */
-	value &= ~CMDQ_CTRL_SRC_STM_SEL_FIELD;
-	value &= ~CMDQ_CTRL_DST_STM_SEL_FIELD;
+	value &= (~CMDQ_CTRL_SRC_STM_SEL_FIELD & ~CMDQ_CTRL_DST_STM_SEL_FIELD);
 	value |= (CMDQ_CTRL_SRC_STM_SEL_FIELD | CMDQ_CTRL_DST_STM_SEL_FIELD |
 		CMDQ_CTRL_ICMDQEN_FIELD);
-	aes_writel(dd, value, CMDQUE_CONTROL);
-	dev_dbg(dd->dev, "cmd_q_ctrl=0x%x", value);
+	aes_writel(eng, value, CMDQUE_CONTROL);
 
 	value = 0;
 	if (mode & FLAGS_CBC) {
 		value = ((0x1 << SECURE_INPUT_ALG_SEL_SHIFT) |
-			((dd->ctx->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
+			((eng->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
 			((u32)upd_iv << SECURE_IV_SELECT_SHIFT) |
 			(((mode & FLAGS_ENCRYPT) ? 2 : 3)
 				<< SECURE_XOR_POS_SHIFT) |
@@ -282,7 +354,7 @@ static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
 			(0 << SECURE_HASH_ENB_SHIFT));
 	} else if (mode & FLAGS_OFB) {
 		value = ((0x1 << SECURE_INPUT_ALG_SEL_SHIFT) |
-			((dd->ctx->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
+			((eng->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
 			((u32)upd_iv << SECURE_IV_SELECT_SHIFT) |
 			((u32)0 << SECURE_IV_SELECT_SHIFT) |
 			(SECURE_XOR_POS_FIELD) |
@@ -293,7 +365,7 @@ static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
 			(0 << SECURE_HASH_ENB_SHIFT));
 	} else if (mode & FLAGS_RNG){
 		value = ((0x1 << SECURE_INPUT_ALG_SEL_SHIFT) |
-			((dd->ctx->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
+			((eng->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
 			((u32)upd_iv << SECURE_IV_SELECT_SHIFT) |
 			(0 << SECURE_XOR_POS_SHIFT) |
 			(0 << SECURE_INPUT_SEL_SHIFT) |
@@ -303,7 +375,7 @@ static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
 			(0 << SECURE_HASH_ENB_SHIFT));
 	} else {
 		value = ((0x1 << SECURE_INPUT_ALG_SEL_SHIFT) |
-			((dd->ctx->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
+			((eng->keylen * 8) << SECURE_INPUT_KEY_LEN_SHIFT) |
 			((u32)upd_iv << SECURE_IV_SELECT_SHIFT) |
 			(0 << SECURE_XOR_POS_SHIFT) |
 			(0 << SECURE_INPUT_SEL_SHIFT) |
@@ -312,31 +384,30 @@ static int aes_start_crypt(struct tegra_aes_dev *dd, u32 in_addr, u32 out_addr,
 			(0 << SECURE_RNG_ENB_SHIFT) |
 				(0 << SECURE_HASH_ENB_SHIFT));
 	}
-	dev_dbg(dd->dev, "secure_in_sel=0x%x", value);
-	aes_writel(dd, value, SECURE_INPUT_SELECT);
+	aes_writel(eng, value, SECURE_INPUT_SELECT);
 
-	aes_writel(dd, out_addr, SECURE_DEST_ADDR);
-	INIT_COMPLETION(dd->op_complete);
+	aes_writel(eng, out_addr, SECURE_DEST_ADDR);
+	INIT_COMPLETION(eng->op_complete);
 
 	for (i = 0; i < qlen - 1; i++) {
 		do {
-			value = aes_readl(dd, INTR_STATUS);
+			value = aes_readl(eng, INTR_STATUS);
 			eng_busy = value & BIT(0);
 			icq_empty = value & BIT(3);
-		} while (eng_busy & (!icq_empty));
-		aes_writel(dd, cmdq[i], ICMDQUE_WR);
+		} while (eng_busy || (!icq_empty));
+		aes_writel(eng, cmdq[i], ICMDQUE_WR);
 	}
 
-	ret = wait_for_completion_timeout(&dd->op_complete, msecs_to_jiffies(150));
+	ret = wait_for_completion_timeout(&eng->op_complete,
+		msecs_to_jiffies(150));
 	if (ret == 0) {
-		dev_err(dd->dev, "timed out (0x%x)\n",
-			aes_readl(dd, INTR_STATUS));
-		disable_irq(INT_VDE_BSE_V);
+		dev_err(aes_dev->dev, "engine%d timed out (0x%x)\n",
+			eng->res_id, aes_readl(eng, INTR_STATUS));
+		disable_irq(eng->irq);
 		return -ETIMEDOUT;
 	}
-
-	disable_irq(INT_VDE_BSE_V);
-	aes_writel(dd, cmdq[qlen - 1], ICMDQUE_WR);
+	disable_irq(eng->irq);
+	aes_writel(eng, cmdq[qlen - 1], ICMDQUE_WR);
 	return 0;
 }
 
@@ -351,174 +422,169 @@ static void aes_release_key_slot(struct tegra_aes_ctx *ctx)
 static struct tegra_aes_slot *aes_find_key_slot(struct tegra_aes_dev *dd)
 {
 	struct tegra_aes_slot *slot = NULL;
-	bool found = false;
+	bool found = 0;
 
 	spin_lock(&list_lock);
-	list_for_each_entry(slot, &dev_list, node) {
+	list_for_each_entry(slot, &slot_list, node) {
 		dev_dbg(dd->dev, "empty:%d, num:%d\n", slot->available,
 			slot->slot_num);
 		if (slot->available) {
 			slot->available = false;
-			found = true;
+			found = 1;
 			break;
 		}
 	}
+
 	spin_unlock(&list_lock);
 	return found ? slot : NULL;
 }
 
-static int aes_set_key(struct tegra_aes_dev *dd)
+static int aes_set_key(struct tegra_aes_engine *eng)
 {
+	struct tegra_aes_dev *dd = aes_dev;
 	u32 value, cmdq[2];
-	struct tegra_aes_ctx *ctx = dd->ctx;
 	int i, eng_busy, icq_empty, dma_busy;
-	bool use_ssk = false;
 
-	if (!ctx) {
+	if (!eng) {
 		dev_err(dd->dev, "%s: context invalid\n", __func__);
 		return -EINVAL;
 	}
 
-	/* use ssk? */
-	if (!dd->ctx->slot) {
-		dev_dbg(dd->dev, "using ssk");
-		dd->ctx->slot = &ssk;
-		use_ssk = true;
-	}
-
 	/* enable key schedule generation in hardware */
-	value = aes_readl(dd, SECURE_CONFIG_EXT);
+	value = aes_readl(eng, SECURE_CONFIG_EXT);
 	value &= ~SECURE_KEY_SCH_DIS_FIELD;
-	aes_writel(dd, value, SECURE_CONFIG_EXT);
+	aes_writel(eng, value, SECURE_CONFIG_EXT);
 
 	/* select the key slot */
-	value = aes_readl(dd, SECURE_CONFIG);
+	value = aes_readl(eng, SECURE_CONFIG);
 	value &= ~SECURE_KEY_INDEX_FIELD;
-	value |= (ctx->slot->slot_num << SECURE_KEY_INDEX_SHIFT);
-	aes_writel(dd, value, SECURE_CONFIG);
+	value |= (eng->slot_num << SECURE_KEY_INDEX_SHIFT);
+	aes_writel(eng, value, SECURE_CONFIG);
 
-	if (use_ssk)
-		goto out;
-
-	/* copy the key table from sdram to vram */
-	cmdq[0] = 0;
-	cmdq[0] = UCQOPCODE_MEMDMAVD << ICQBITSHIFT_OPCODE |
-		(MEMDMA_DIR_DTOVRAM << MEMDMABITSHIFT_DIR) |
-		(AES_HW_KEY_TABLE_LENGTH_BYTES/sizeof(u32))
-			<< MEMDMABITSHIFT_NUM_WORDS;
-	cmdq[1] = (u32)dd->ivkey_phys_base;
+	if (eng->res_id == TEGRA_ARB_BSEV) {
 
-	for (i = 0; i < ARRAY_SIZE(cmdq); i++)
-		aes_writel(dd, cmdq[i], ICMDQUE_WR);
-
-	do {
-		value = aes_readl(dd, INTR_STATUS);
-		eng_busy = value & BIT(0);
-		icq_empty = value & BIT(3);
-		dma_busy = value & BIT(23);
-	} while (eng_busy & (!icq_empty) & dma_busy);
+		if (eng->use_ssk)
+			goto out;
 
-	/* settable command to get key into internal registers */
-	value = 0;
-	value = UCQOPCODE_SETTABLE << ICQBITSHIFT_OPCODE |
-		UCQCMD_CRYPTO_TABLESEL << ICQBITSHIFT_TABLESEL |
-		UCQCMD_VRAM_SEL << ICQBITSHIFT_VRAMSEL |
-		(UCQCMD_KEYTABLESEL | ctx->slot->slot_num)
+		/* copy the key table from sdram to vram */
+		cmdq[0] = 0;
+		cmdq[0] = UCQOPCODE_MEMDMAVD << ICQBITSHIFT_OPCODE |
+				(MEMDMA_DIR_DTOVRAM << MEMDMABITSHIFT_DIR) |
+			(AES_HW_KEY_TABLE_LENGTH_BYTES/sizeof(u32))
+			<< MEMDMABITSHIFT_NUM_WORDS;
+		cmdq[1] = (u32)eng->ivkey_phys_base;
+		for (i = 0; i < ARRAY_SIZE(cmdq); i++)
+			aes_writel(eng, cmdq[i], ICMDQUE_WR);
+		do {
+			value = aes_readl(eng, INTR_STATUS);
+			eng_busy = value & BIT(0);
+			icq_empty = value & BIT(3);
+			dma_busy = value & BIT(23);
+		} while (eng_busy & (!icq_empty) & dma_busy);
+
+		/* settable command to get key into internal registers */
+		value = 0;
+		value = UCQOPCODE_SETTABLE << ICQBITSHIFT_OPCODE |
+			UCQCMD_CRYPTO_TABLESEL << ICQBITSHIFT_TABLESEL |
+			UCQCMD_VRAM_SEL << ICQBITSHIFT_VRAMSEL |
+			(UCQCMD_KEYTABLESEL | eng->slot_num)
 			<< ICQBITSHIFT_KEYTABLEID;
-	aes_writel(dd, value, ICMDQUE_WR);
-	do {
-		value = aes_readl(dd, INTR_STATUS);
-		eng_busy = value & BIT(0);
-		icq_empty = value & BIT(3);
-	} while (eng_busy & (!icq_empty));
+		aes_writel(eng, value, ICMDQUE_WR);
+		do {
+			value = aes_readl(eng, INTR_STATUS);
+			eng_busy = value & BIT(0);
+			icq_empty = value & BIT(3);
+		} while (eng_busy & (!icq_empty));
+	} else {
+		if (eng->use_ssk)
+			goto out;
 
+		/* settable command to get key into internal registers */
+		value = 0;
+		value = UCQOPCODE_SETTABLE << ICQBITSHIFT_OPCODE |
+			UCQCMD_CRYPTO_TABLESEL << ICQBITSHIFT_TABLESEL |
+			(UCQCMD_KEYTABLESEL | eng->slot_num)
+			<< ICQBITSHIFT_KEYTABLEID |
+			dd->bsea_iram_address >> 2;
+			aes_writel(eng, value, ICMDQUE_WR);
+		do {
+			value = aes_readl(eng, INTR_STATUS);
+			eng_busy = value & BIT(0);
+			icq_empty = value & BIT(3);
+		} while (eng_busy & (!icq_empty));
+	}
 out:
 	return 0;
 }
 
-static int tegra_aes_handle_req(struct tegra_aes_dev *dd)
+static int tegra_aes_handle_req(struct tegra_aes_engine *eng)
 {
-	struct crypto_async_request *async_req, *backlog;
+	struct tegra_aes_dev *dd = aes_dev;
 	struct tegra_aes_ctx *ctx;
+	struct crypto_async_request *async_req, *backlog;
 	struct tegra_aes_reqctx *rctx;
 	struct ablkcipher_request *req;
-	unsigned long flags;
+	unsigned long irq_flags;
 	int dma_max = AES_HW_DMA_BUFFER_SIZE_BYTES;
-	int ret = 0, nblocks, total;
-	int count = 0;
+	int nblocks, total, ret = 0, count = 0;
 	dma_addr_t addr_in, addr_out;
 	struct scatterlist *in_sg, *out_sg;
 
-	if (!dd)
-		return -EINVAL;
-
-	spin_lock_irqsave(&dd->lock, flags);
+	spin_lock_irqsave(&dd->lock, irq_flags);
 	backlog = crypto_get_backlog(&dd->queue);
 	async_req = crypto_dequeue_request(&dd->queue);
 	if (!async_req)
-		clear_bit(FLAGS_BUSY, &dd->flags);
-	spin_unlock_irqrestore(&dd->lock, flags);
+		clear_bit(FLAGS_BUSY, &eng->busy);
 
+	spin_unlock_irqrestore(&dd->lock, irq_flags);
 	if (!async_req)
 		return -ENODATA;
-
 	if (backlog)
 		backlog->complete(backlog, -EINPROGRESS);
 
 	req = ablkcipher_request_cast(async_req);
-
-	dev_dbg(dd->dev, "%s: get new req\n", __func__);
-
+	dev_dbg(dd->dev, "%s: get new req (engine #%d)\n", __func__,
+		eng->res_id);
 	if (!req->src || !req->dst)
 		return -EINVAL;
 
 	/* take the hardware semaphore */
-	if (tegra_arb_mutex_lock_timeout(dd->res_id, ARB_SEMA_TIMEOUT) < 0) {
-		dev_err(dd->dev, "aes hardware not available\n");
+	if (tegra_arb_mutex_lock_timeout(eng->res_id, ARB_SEMA_TIMEOUT) < 0) {
+		dev_err(dd->dev, "aes hardware (%d) not available\n",
+		eng->res_id);
 		return -EBUSY;
 	}
-
 	/* assign new request to device */
-	dd->req = req;
-	dd->total = req->nbytes;
-	dd->in_offset = 0;
-	dd->in_sg = req->src;
-	dd->out_offset = 0;
-	dd->out_sg = req->dst;
+	eng->req = req;
+	eng->total = req->nbytes;
+	eng->in_offset = 0;
+	eng->in_sg = req->src;
+	eng->out_offset = 0;
+	eng->out_sg = req->dst;
 
-	in_sg = dd->in_sg;
-	out_sg = dd->out_sg;
+	in_sg = eng->in_sg;
+	out_sg = eng->out_sg;
+	total = eng->total;
 
-	total = dd->total;
 	rctx = ablkcipher_request_ctx(req);
 	ctx = crypto_ablkcipher_ctx(crypto_ablkcipher_reqtfm(req));
 	rctx->mode &= FLAGS_MODE_MASK;
 	dd->flags = (dd->flags & ~FLAGS_MODE_MASK) | rctx->mode;
 
-	dd->iv = (u8 *)req->info;
-	dd->ivlen = AES_BLOCK_SIZE;
-
-	/* assign new context to device */
-	ctx->dd = dd;
-	dd->ctx = ctx;
-
-	if (ctx->flags & FLAGS_NEW_KEY) {
-		/* copy the key */
-		memset(dd->ivkey_base, 0, AES_HW_KEY_TABLE_LENGTH_BYTES);
-		memcpy(dd->ivkey_base, ctx->key, ctx->keylen);
-		aes_set_key(dd);
-		ctx->flags &= ~FLAGS_NEW_KEY;
+	if (eng->new_key) {
+		aes_set_key(eng);
+		eng->new_key = false;
 	}
 
-	if (((dd->flags & FLAGS_CBC) || (dd->flags & FLAGS_OFB)) && dd->iv) {
+	if (((dd->flags & FLAGS_CBC) || (dd->flags & FLAGS_OFB)) && req->info) {
 		/* set iv to the aes hw slot
 		 * Hw generates updated iv only after iv is set in slot.
 		 * So key and iv is passed asynchronously.
 		 */
-		memcpy(dd->buf_in, dd->iv, dd->ivlen);
+		memcpy(eng->buf_in, (u8 *)req->info, AES_BLOCK_SIZE);
 
-		ret = aes_start_crypt(dd, (u32)dd->dma_buf_in,
-		  (u32)dd->dma_buf_out, 1, FLAGS_CBC, false);
+		ret = aes_start_crypt(eng, (u32)eng->dma_buf_in,
+			(u32)eng->dma_buf_out, 1, FLAGS_CBC, false);
 		if (ret < 0) {
 			dev_err(dd->dev, "aes_start_crypt fail(%d)\n", ret);
 			goto out;
@@ -536,29 +602,25 @@ static int tegra_aes_handle_req(struct tegra_aes_dev *dd)
 		ret = dma_map_sg(dd->dev, out_sg, 1, DMA_FROM_DEVICE);
 		if (!ret) {
 			dev_err(dd->dev, "dma_map_sg() error\n");
-			dma_unmap_sg(dd->dev, dd->in_sg,
+			dma_unmap_sg(dd->dev, eng->in_sg,
 				1, DMA_TO_DEVICE);
 			goto out;
 		}
 
 		addr_in = sg_dma_address(in_sg);
 		addr_out = sg_dma_address(out_sg);
-		dd->flags |= FLAGS_FAST;
 		count = min((int)sg_dma_len(in_sg), (int)dma_max);
 		WARN_ON(sg_dma_len(in_sg) != sg_dma_len(out_sg));
 		nblocks = DIV_ROUND_UP(count, AES_BLOCK_SIZE);
-
-		ret = aes_start_crypt(dd, addr_in, addr_out, nblocks,
+		ret = aes_start_crypt(eng, addr_in, addr_out, nblocks,
 			dd->flags, true);
 
 		dma_unmap_sg(dd->dev, out_sg, 1, DMA_FROM_DEVICE);
 		dma_unmap_sg(dd->dev, in_sg, 1, DMA_TO_DEVICE);
-
 		if (ret < 0) {
 			dev_err(dd->dev, "aes_start_crypt fail(%d)\n", ret);
 			goto out;
 		}
-		dd->flags &= ~FLAGS_FAST;
 
 		dev_dbg(dd->dev, "out: copied %d\n", count);
 		total -= count;
@@ -569,12 +631,11 @@ static int tegra_aes_handle_req(struct tegra_aes_dev *dd)
 
 out:
 	/* release the hardware semaphore */
-	tegra_arb_mutex_unlock(dd->res_id);
-
-	dd->total = total;
+	tegra_arb_mutex_unlock(eng->res_id);
+	eng->total = total;
 
-	if (dd->req->base.complete)
-		dd->req->base.complete(&dd->req->base, ret);
+	if (eng->req->base.complete)
+		eng->req->base.complete(&eng->req->base, ret);
 
 	dev_dbg(dd->dev, "%s: exit\n", __func__);
 	return ret;
@@ -610,43 +671,95 @@ static int tegra_aes_setkey(struct crypto_ablkcipher *tfm, const u8 *key,
 				dev_err(dd->dev, "no empty slot\n");
 				return -ENOMEM;
 			}
-
 			ctx->slot = key_slot;
 		}
 
-		memcpy(ctx->key, key, keylen);
-		ctx->keylen = keylen;
+		/* copy the key */
+		memset(dd->bsev.ivkey_base, 0, AES_HW_KEY_TABLE_LENGTH_BYTES);
+		memset(dd->bsea_iram_base, 0, AES_HW_KEY_TABLE_LENGTH_BYTES);
+		memcpy(dd->bsev.ivkey_base, key, keylen);
+		memcpy(dd->bsea_iram_base, key, keylen);
+	} else {
+		dd->bsev.slot_num = SSK_SLOT_NUM;
+		dd->bsea.slot_num = SSK_SLOT_NUM;
+		dd->bsev.use_ssk = true;
+		dd->bsea.use_ssk = true;
+		keylen = AES_KEYSIZE_128;
 	}
 
-	ctx->flags |= FLAGS_NEW_KEY;
+	dd->bsev.keylen = keylen;
+	dd->bsea.keylen = keylen;
+	dd->bsev.new_key = true;
+	dd->bsea.new_key = true;
 	dev_dbg(dd->dev, "done\n");
 	return 0;
 }
 
-static void aes_workqueue_handler(struct work_struct *work)
+static void bsev_workqueue_handler(struct work_struct *work)
 {
 	struct tegra_aes_dev *dd = aes_dev;
 	int ret;
 
-	aes_hw_init(dd);
+	aes_hw_init(&dd->bsev);
 
 	/* empty the crypto queue and then return */
 	do {
-		ret = tegra_aes_handle_req(dd);
+		ret = tegra_aes_handle_req(&dd->bsev);
 	} while (!ret);
 
-	aes_hw_deinit(dd);
+	aes_hw_deinit(&dd->bsev);
 }
 
-static irqreturn_t aes_irq(int irq, void *dev_id)
+static void bsea_workqueue_handler(struct work_struct *work)
+{
+	struct tegra_aes_dev *dd = aes_dev;
+	int ret;
+
+	aes_hw_init(&dd->bsea);
+
+	/* empty the crypto queue and then return */
+	do {
+		ret = tegra_aes_handle_req(&dd->bsea);
+	} while (!ret);
+
+	aes_hw_deinit(&dd->bsea);
+}
+
+static irqreturn_t aes_bsev_irq(int irq, void *dev_id)
+{
+	struct tegra_aes_dev *dd = (struct tegra_aes_dev *)dev_id;
+	u32 value = aes_readl(&dd->bsev, INTR_STATUS);
+
+	if (!(value & ENGINE_BUSY_FIELD))
+		complete(&dd->bsev.op_complete);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t aes_bsea_irq(int irq, void *dev_id)
 {
 	struct tegra_aes_dev *dd = (struct tegra_aes_dev *)dev_id;
-	u32 value = aes_readl(dd, INTR_STATUS);
+	u32 value = aes_readl(&dd->bsea, INTR_STATUS);
+	u32 intr_err_mask = (value & (BIT(19) | BIT(20)));
+	u32 cmdq_ctrl = 0;
 
 	dev_dbg(dd->dev, "irq_stat: 0x%x", value);
-	if (!((value & ENGINE_BUSY_FIELD) & !(value & ICQ_EMPTY_FIELD)))
-		complete(&dd->op_complete);
+	if (intr_err_mask) {
+		dev_err(dd->dev, "BSEA Error Intrrupt: INTR_STATUS=0x%x\n",
+		value);
+		if (value & BIT(2)) {
+			cmdq_ctrl = aes_readl(&dd->bsea, CMDQUE_CONTROL);
+			cmdq_ctrl |= BIT(2);
+			aes_writel(&dd->bsea, cmdq_ctrl, CMDQUE_CONTROL);
+			aes_writel(&dd->bsea, intr_err_mask, INTR_STATUS);
+		} else
+			aes_writel(&dd->bsea, intr_err_mask, INTR_STATUS);
+		goto done;
+	}
+	if (!(value & ENGINE_BUSY_FIELD))
+		complete(&dd->bsea.op_complete);
 
+done:
 	return IRQ_HANDLED;
 }
 
@@ -656,7 +769,8 @@ static int tegra_aes_crypt(struct ablkcipher_request *req, unsigned long mode)
 	struct tegra_aes_dev *dd = aes_dev;
 	unsigned long flags;
 	int err = 0;
-	int busy;
+	int bsev_busy;
+	int bsea_busy;
 
 	dev_dbg(dd->dev, "nbytes: %d, enc: %d, cbc: %d\n", req->nbytes,
 		!!(mode & FLAGS_ENCRYPT),
@@ -666,11 +780,14 @@ static int tegra_aes_crypt(struct ablkcipher_request *req, unsigned long mode)
 
 	spin_lock_irqsave(&dd->lock, flags);
 	err = ablkcipher_enqueue_request(&dd->queue, req);
-	busy = test_and_set_bit(FLAGS_BUSY, &dd->flags);
+	bsev_busy = test_and_set_bit(FLAGS_BUSY, &dd->bsev.busy);
+	bsea_busy = test_and_set_bit(FLAGS_BUSY, &dd->bsea.busy);
 	spin_unlock_irqrestore(&dd->lock, flags);
 
-	if (!busy)
-		queue_work(aes_wq, &aes_work);
+	if (!bsev_busy)
+		queue_work(bsev_wq, &bsev_work);
+	if (!bsea_busy)
+		queue_work(bsea_wq, &bsea_work);
 
 	return err;
 }
@@ -707,7 +824,7 @@ static int tegra_aes_get_random(struct crypto_rng *tfm, u8 *rdata,
 	unsigned int dlen)
 {
 	struct tegra_aes_dev *dd = aes_dev;
-	struct tegra_aes_ctx *ctx = &rng_ctx;
+	struct tegra_aes_engine *eng = &dd->bsev;
 	int ret, i;
 	u8 *dest = rdata, *dt = dd->dt;
 
@@ -715,34 +832,31 @@ static int tegra_aes_get_random(struct crypto_rng *tfm, u8 *rdata,
 	mutex_lock(&aes_lock);
 
 	/* take the hardware semaphore */
-	if (tegra_arb_mutex_lock_timeout(dd->res_id, ARB_SEMA_TIMEOUT) < 0) {
-		dev_err(dd->dev, "aes hardware not available\n");
+	if (tegra_arb_mutex_lock_timeout(eng->res_id, ARB_SEMA_TIMEOUT) < 0) {
+		dev_err(dd->dev, "aes hardware (%d) not available\n",
+		eng->res_id);
 		mutex_unlock(&aes_lock);
 		return -EBUSY;
 	}
 
-	ret = aes_hw_init(dd);
+	ret = aes_hw_init(eng);
 	if (ret < 0) {
 		dev_err(dd->dev, "%s: hw init fail(%d)\n", __func__, ret);
 		dlen = ret;
 		goto fail;
 	}
 
-	ctx->dd = dd;
-	dd->ctx = ctx;
-	dd->flags = FLAGS_ENCRYPT | FLAGS_RNG;
-
-	memset(dd->buf_in, 0, AES_BLOCK_SIZE);
-	memcpy(dd->buf_in, dt, DEFAULT_RNG_BLK_SZ);
+	memset(eng->buf_in, 0, AES_BLOCK_SIZE);
+	memcpy(eng->buf_in, dt, DEFAULT_RNG_BLK_SZ);
 
-	ret = aes_start_crypt(dd, (u32)dd->dma_buf_in,
-		(u32)dd->dma_buf_out, 1, dd->flags, true);
+	ret = aes_start_crypt(eng, (u32)eng->dma_buf_in, (u32)eng->dma_buf_out,
+		1, FLAGS_ENCRYPT | FLAGS_RNG, true);
 	if (ret < 0) {
 		dev_err(dd->dev, "aes_start_crypt fail(%d)\n", ret);
 		dlen = ret;
 		goto out;
 	}
-	memcpy(dest, dd->buf_out, dlen);
+	memcpy(dest, eng->buf_out, dlen);
 
 	/* update the DT */
 	for (i = DEFAULT_RNG_BLK_SZ - 1; i >= 0; i--) {
@@ -752,11 +866,11 @@ static int tegra_aes_get_random(struct crypto_rng *tfm, u8 *rdata,
 	}
 
 out:
-	aes_hw_deinit(dd);
+	aes_hw_deinit(eng);
 
 fail:
 	/* release the hardware semaphore */
-	tegra_arb_mutex_unlock(dd->res_id);
+	tegra_arb_mutex_unlock(eng->res_id);
 	mutex_unlock(&aes_lock);
 	dev_dbg(dd->dev, "%s: done\n", __func__);
 	return dlen;
@@ -767,23 +881,25 @@ static int tegra_aes_rng_reset(struct crypto_rng *tfm, u8 *seed,
 {
 	struct tegra_aes_dev *dd = aes_dev;
 	struct tegra_aes_ctx *ctx = &rng_ctx;
+	struct tegra_aes_engine *eng = &dd->bsev;
 	struct tegra_aes_slot *key_slot;
 	struct timespec ts;
 	int ret = 0;
 	u64 nsec, tmp[2];
 	u8 *dt;
 
-	if (!ctx || !dd) {
-		dev_err(dd->dev, "ctx=0x%x, dd=0x%x\n",
-			(unsigned int)ctx, (unsigned int)dd);
+	if (!eng || !dd) {
+		dev_err(dd->dev, "eng=0x%x, dd=0x%x\n",
+			(unsigned int)eng, (unsigned int)dd);
 		return -EINVAL;
 	}
 
 	if (slen < (DEFAULT_RNG_BLK_SZ + AES_KEYSIZE_128)) {
-		dev_err(dd->dev, "seed size invalid");
 		return -ENOMEM;
 	}
 
+	dd->flags = FLAGS_ENCRYPT | FLAGS_RNG;
+
 	/* take mutex to access the aes hw */
 	mutex_lock(&aes_lock);
 
@@ -797,49 +913,41 @@ static int tegra_aes_rng_reset(struct crypto_rng *tfm, u8 *seed,
 		ctx->slot = key_slot;
 	}
 
-	ctx->dd = dd;
-	dd->ctx = ctx;
-	dd->ctr = 0;
-
-	ctx->keylen = AES_KEYSIZE_128;
-	ctx->flags |= FLAGS_NEW_KEY;
-
 	/* copy the key to the key slot */
-	memset(dd->ivkey_base, 0, AES_HW_KEY_TABLE_LENGTH_BYTES);
-	memcpy(dd->ivkey_base, seed + DEFAULT_RNG_BLK_SZ, AES_KEYSIZE_128);
-
-	dd->iv = seed;
-	dd->ivlen = slen;
-
-	dd->flags = FLAGS_ENCRYPT | FLAGS_RNG;
+	memset(eng->ivkey_base, 0, AES_HW_KEY_TABLE_LENGTH_BYTES);
+	memcpy(eng->ivkey_base, seed + DEFAULT_RNG_BLK_SZ, AES_KEYSIZE_128);
 
 	/* take the hardware semaphore */
-	if (tegra_arb_mutex_lock_timeout(dd->res_id, ARB_SEMA_TIMEOUT) < 0) {
-		dev_err(dd->dev, "aes hardware not available\n");
+	if (tegra_arb_mutex_lock_timeout(eng->res_id, ARB_SEMA_TIMEOUT) < 0) {
+		dev_err(dd->dev, "aes hardware (%d) not available\n",
+		eng->res_id);
 		mutex_unlock(&aes_lock);
 		return -EBUSY;
 	}
 
-	ret = aes_hw_init(dd);
+	ret = aes_hw_init(eng);
 	if (ret < 0) {
 		dev_err(dd->dev, "%s: hw init fail(%d)\n", __func__, ret);
 		goto fail;
 	}
 
-	aes_set_key(dd);
+	dd->ctx = ctx;
+	eng->slot_num = ctx->slot->slot_num;
+	eng->keylen = AES_KEYSIZE_128;
+	aes_set_key(eng);
 
 	/* set seed to the aes hw slot */
-	memset(dd->buf_in, 0, AES_BLOCK_SIZE);
-	memcpy(dd->buf_in, dd->iv, DEFAULT_RNG_BLK_SZ);
-	ret = aes_start_crypt(dd, (u32)dd->dma_buf_in,
-	  (u32)dd->dma_buf_out, 1, FLAGS_CBC, false);
+	memset(eng->buf_in, 0, AES_BLOCK_SIZE);
+	memcpy(eng->buf_in, seed, DEFAULT_RNG_BLK_SZ);
+	ret = aes_start_crypt(eng, (u32)eng->dma_buf_in,
+	  (u32)eng->dma_buf_out, 1, FLAGS_CBC, false);
 	if (ret < 0) {
 		dev_err(dd->dev, "aes_start_crypt fail(%d)\n", ret);
 		goto out;
 	}
 
-	if (dd->ivlen >= (2 * DEFAULT_RNG_BLK_SZ + AES_KEYSIZE_128)) {
-		dt = dd->iv + DEFAULT_RNG_BLK_SZ + AES_KEYSIZE_128;
+	if (slen >= (2 * DEFAULT_RNG_BLK_SZ + AES_KEYSIZE_128)) {
+		dt = seed + DEFAULT_RNG_BLK_SZ + AES_KEYSIZE_128;
 	} else {
 		getnstimeofday(&ts);
 		nsec = timespec_to_ns(&ts);
@@ -853,11 +961,11 @@ static int tegra_aes_rng_reset(struct crypto_rng *tfm, u8 *seed,
 	memcpy(dd->dt, dt, DEFAULT_RNG_BLK_SZ);
 
 out:
-	aes_hw_deinit(dd);
+	aes_hw_deinit(eng);
 
 fail:
 	/* release the hardware semaphore */
-	tegra_arb_mutex_unlock(dd->res_id);
+	tegra_arb_mutex_unlock(eng->res_id);
 	mutex_unlock(&aes_lock);
 
 	dev_dbg(dd->dev, "%s: done\n", __func__);
@@ -867,7 +975,6 @@ fail:
 static int tegra_aes_cra_init(struct crypto_tfm *tfm)
 {
 	tfm->crt_ablkcipher.reqsize = sizeof(struct tegra_aes_reqctx);
-
 	return 0;
 }
 
@@ -961,7 +1068,7 @@ static int tegra_aes_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct tegra_aes_dev *dd;
-	struct resource *res;
+	struct resource *res[2];
 	int err = -ENOMEM, i = 0, j;
 
 	if (aes_dev)
@@ -986,42 +1093,57 @@ static int tegra_aes_probe(struct platform_device *pdev)
 	crypto_init_queue(&dd->queue, TEGRA_AES_QUEUE_LENGTH);
 
 	/* Get the module base address */
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res) {
+	res[0] = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	res[1] = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (!res[0] || !res[1]) {
 		dev_err(dev, "invalid resource type: base\n");
 		err = -ENODEV;
 		goto out;
 	}
-	dd->phys_base = res->start;
+	dd->bsev.phys_base = res[0]->start;
+	dd->bsev.io_base = ioremap(dd->bsev.phys_base, resource_size(res[0]));
+	dd->bsea.phys_base = res[1]->start;
+	dd->bsea.io_base = ioremap(dd->bsea.phys_base, resource_size(res[1]));
 
-	dd->io_base = ioremap(dd->phys_base, resource_size(res));
-	if (!dd->io_base) {
+	if (!dd->bsev.io_base || !dd->bsea.io_base) {
 		dev_err(dev, "can't ioremap phys_base\n");
 		err = -ENOMEM;
 		goto out;
 	}
 
-	dd->res_id = TEGRA_ARB_AES;
+	err = bsea_alloc_iram(dd);
+	if (err < 0) {
+		dev_err(dev, "Failed to allocate IRAM for BSEA\n");
+		goto out;
+	}
+
+	dd->bsev.res_id = TEGRA_ARB_BSEV;
+	dd->bsea.res_id = TEGRA_ARB_BSEA;
 
-	/* Initialise the master bsev clock */
-	dd->pclk = clk_get(dev, "bsev");
-	if (!dd->pclk) {
+	dd->bsev.pclk = clk_get(dev, "bsev");
+	if (!dd->bsev.pclk) {
 		dev_err(dev, "pclock intialization failed.\n");
 		err = -ENODEV;
 		goto out;
 	}
 
-	/* Initialize the vde clock */
-	dd->iclk = clk_get(dev, "vde");
-	if (!dd->iclk) {
+	dd->bsev.iclk = clk_get(dev, "vde");
+	if (!dd->bsev.iclk) {
 		dev_err(dev, "iclock intialization failed.\n");
 		err = -ENODEV;
 		goto out;
 	}
 
-	err = clk_set_rate(dd->iclk, ULONG_MAX);
+	dd->bsea.pclk = clk_get(dev, "bsea");
+	if (!dd->bsea.pclk) {
+		dev_err(dev, "pclock intialization failed.\n");
+		err = -ENODEV;
+		goto out;
+	}
+
+	err = clk_set_rate(dd->bsev.iclk, ULONG_MAX);
 	if (err) {
-		dev_err(dd->dev, "iclk set_rate fail(%d)\n", err);
+		dev_err(dd->dev, "bsev iclk set_rate fail(%d)\n", err);
 		goto out;
 	}
 
@@ -1030,45 +1152,64 @@ static int tegra_aes_probe(struct platform_device *pdev)
 	 * - hardware key table
 	 * - key schedule
 	 */
-	dd->ivkey_base = dma_alloc_coherent(dev, SZ_512, &dd->ivkey_phys_base,
-		GFP_KERNEL);
-	if (!dd->ivkey_base) {
-		dev_err(dev, "can not allocate iv/key buffer\n");
+	dd->bsev.ivkey_base = dma_alloc_coherent(dev, SZ_512,
+		&dd->bsev.ivkey_phys_base, GFP_KERNEL);
+	/* Allocated IRAM for BSEA */
+	dd->bsea.ivkey_base = NULL;
+	if (!dd->bsev.ivkey_base) {
+		dev_err(dev, "can not allocate iv/key buffer for BSEV\n");
 		err = -ENOMEM;
 		goto out;
 	}
 
-	dd->buf_in = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-		&dd->dma_buf_in, GFP_KERNEL);
-	if (!dd->buf_in) {
+	dd->bsev.buf_in = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+		&dd->bsev.dma_buf_in, GFP_KERNEL);
+	dd->bsea.buf_in = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+		&dd->bsea.dma_buf_in, GFP_KERNEL);
+	if (!dd->bsev.buf_in || !dd->bsea.buf_in) {
 		dev_err(dev, "can not allocate dma-in buffer\n");
 		err = -ENOMEM;
 		goto out;
 	}
 
-	dd->buf_out = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-		&dd->dma_buf_out, GFP_KERNEL);
-	if (!dd->buf_out) {
+	dd->bsev.buf_out = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+		&dd->bsev.dma_buf_out, GFP_KERNEL);
+	dd->bsea.buf_out = dma_alloc_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+		&dd->bsea.dma_buf_out, GFP_KERNEL);
+	if (!dd->bsev.buf_out || !dd->bsea.buf_out) {
 		dev_err(dev, "can not allocate dma-out buffer\n");
 		err = -ENOMEM;
 		goto out;
 	}
 
-	init_completion(&dd->op_complete);
-	aes_wq = alloc_workqueue("aes_wq", WQ_HIGHPRI, 16);
-	if (!aes_wq) {
+	init_completion(&dd->bsev.op_complete);
+	init_completion(&dd->bsea.op_complete);
+
+	bsev_wq = alloc_workqueue("bsev_wq", WQ_HIGHPRI, 16);
+	bsea_wq = alloc_workqueue("bsea_wq", WQ_HIGHPRI, 16);
+	if (!bsev_wq || !bsea_wq) {
 		dev_err(dev, "alloc_workqueue failed\n");
 		goto out;
 	}
 
 	/* get the irq */
-	err = request_irq(INT_VDE_BSE_V, aes_irq, IRQF_TRIGGER_HIGH,
+	dd->bsev.irq = INT_VDE_BSE_V;
+	err = request_irq(dd->bsev.irq, aes_bsev_irq, IRQF_TRIGGER_HIGH,
 		"tegra-aes", dd);
 	if (err) {
-		dev_err(dev, "request_irq failed\n");
+		dev_err(dev, "request_irq failed fir BSEV Engine\n");
 		goto out;
 	}
-	disable_irq(INT_VDE_BSE_V);
+	disable_irq(dd->bsev.irq);
+
+	dd->bsea.irq = INT_VDE_BSE_A;
+	err = request_irq(dd->bsea.irq, aes_bsea_irq, IRQF_TRIGGER_HIGH,
+		"tegra-aes", dd);
+	if (err) {
+		dev_err(dev, "request_irq failed for BSEA Engine\n");
+		goto out;
+	}
+	disable_irq(dd->bsea.irq);
 
 	spin_lock_init(&list_lock);
 	spin_lock(&list_lock);
@@ -1078,11 +1219,12 @@ static int tegra_aes_probe(struct platform_device *pdev)
 		dd->slots[i].available = true;
 		dd->slots[i].slot_num = i;
 		INIT_LIST_HEAD(&dd->slots[i].node);
-		list_add_tail(&dd->slots[i].node, &dev_list);
+		list_add_tail(&dd->slots[i].node, &slot_list);
 	}
 	spin_unlock(&list_lock);
 
 	aes_dev = dd;
+
 	for (i = 0; i < ARRAY_SIZE(algs); i++) {
 		INIT_LIST_HEAD(&algs[i].cra_list);
 		err = crypto_register_alg(&algs[i]);
@@ -1096,31 +1238,61 @@ static int tegra_aes_probe(struct platform_device *pdev)
 out:
 	for (j = 0; j < i; j++)
 		crypto_unregister_alg(&algs[j]);
-	if (dd->ivkey_base)
-		dma_free_coherent(dev, SZ_512, dd->ivkey_base,
-			dd->ivkey_phys_base);
-	if (dd->buf_in)
+
+	bsea_free_iram(dd);
+	if (dd->bsev.ivkey_base) {
+		dma_free_coherent(dev, SZ_512, dd->bsev.ivkey_base,
+			dd->bsev.ivkey_phys_base);
+	}
+
+	if (dd->bsev.buf_in && dd->bsea.buf_in) {
+		dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+			dd->bsev.buf_in, dd->bsev.dma_buf_in);
+		dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
+			dd->bsea.buf_in, dd->bsea.dma_buf_in);
+	}
+
+	if (dd->bsev.buf_out && dd->bsea.buf_out) {
 		dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-			dd->buf_in, dd->dma_buf_in);
-	if (dd->buf_out)
+			dd->bsev.buf_out, dd->bsev.dma_buf_out);
 		dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-			dd->buf_out, dd->dma_buf_out);
-	if (dd->io_base)
-		iounmap(dd->io_base);
-	if (dd->iclk)
-		clk_put(dd->iclk);
-	if (dd->pclk)
-		clk_put(dd->pclk);
-	if (aes_wq)
-		destroy_workqueue(aes_wq);
-	free_irq(INT_VDE_BSE_V, dd);
+			dd->bsea.buf_out, dd->bsea.dma_buf_out);
+	}
+
+	if (dd->bsev.io_base && dd->bsea.io_base) {
+		iounmap(dd->bsev.io_base);
+		iounmap(dd->bsea.io_base);
+	}
+
+	if (dd->bsev.pclk)
+		clk_put(dd->bsev.pclk);
+
+	if (dd->bsev.iclk)
+		clk_put(dd->bsev.iclk);
+
+	if (dd->bsea.pclk)
+		clk_put(dd->bsea.pclk);
+
+	if (bsev_wq)
+		destroy_workqueue(bsev_wq);
+
+	if (bsea_wq)
+		destroy_workqueue(bsea_wq);
+
+	if (dd->bsev.irq)
+		free_irq(dd->bsev.irq, dd);
+
+	if (dd->bsea.irq)
+		free_irq(dd->bsea.irq, dd);
+
 	spin_lock(&list_lock);
-	list_del(&dev_list);
+	list_del(&slot_list);
 	spin_unlock(&list_lock);
 
 	kfree(dd->slots);
 	kfree(dd);
 	aes_dev = NULL;
+
 	dev_err(dev, "%s: initialization failed.\n", __func__);
 	return err;
 }
@@ -1134,25 +1306,36 @@ static int __devexit tegra_aes_remove(struct platform_device *pdev)
 	if (!dd)
 		return -ENODEV;
 
-	cancel_work_sync(&aes_work);
-	destroy_workqueue(aes_wq);
-	free_irq(INT_VDE_BSE_V, dd);
+	cancel_work_sync(&bsev_work);
+	cancel_work_sync(&bsea_work);
+	destroy_workqueue(bsev_wq);
+	destroy_workqueue(bsea_wq);
+	free_irq(dd->bsev.irq, dd);
+	free_irq(dd->bsea.irq, dd);
 	spin_lock(&list_lock);
-	list_del(&dev_list);
+	list_del(&slot_list);
 	spin_unlock(&list_lock);
 
 	for (i = 0; i < ARRAY_SIZE(algs); i++)
 		crypto_unregister_alg(&algs[i]);
 
-	dma_free_coherent(dev, SZ_512, dd->ivkey_base,
-		dd->ivkey_phys_base);
-	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-		dd->buf_in, dd->dma_buf_in);
-	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES,
-		dd->buf_out, dd->dma_buf_out);
-	iounmap(dd->io_base);
-	clk_put(dd->iclk);
-	clk_put(dd->pclk);
+	bsea_free_iram(dd);
+	dma_free_coherent(dev, SZ_512, dd->bsev.ivkey_base,
+		dd->bsev.ivkey_phys_base);
+	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES, dd->bsev.buf_in,
+		dd->bsev.dma_buf_in);
+	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES, dd->bsea.buf_in,
+		dd->bsea.dma_buf_in);
+	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES, dd->bsev.buf_out,
+		dd->bsev.dma_buf_out);
+	dma_free_coherent(dev, AES_HW_DMA_BUFFER_SIZE_BYTES, dd->bsea.buf_out,
+		dd->bsea.dma_buf_out);
+
+	iounmap(dd->bsev.io_base);
+	iounmap(dd->bsea.io_base);
+	clk_put(dd->bsev.iclk);
+	clk_put(dd->bsev.pclk);
+	clk_put(dd->bsea.pclk);
 	kfree(dd->slots);
 	kfree(dd);
 	aes_dev = NULL;
@@ -1172,7 +1355,7 @@ static struct platform_driver tegra_aes_driver = {
 static int __init tegra_aes_mod_init(void)
 {
 	mutex_init(&aes_lock);
-	INIT_LIST_HEAD(&dev_list);
+	INIT_LIST_HEAD(&slot_list);
 	return  platform_driver_register(&tegra_aes_driver);
 }
 
diff --git a/drivers/crypto/tegra-aes.h b/drivers/crypto/tegra-aes.h
index bb311bb58c48..9c53fe4165cc 100644
--- a/drivers/crypto/tegra-aes.h
+++ b/drivers/crypto/tegra-aes.h
@@ -22,7 +22,7 @@
 #define ICMDQUE_WR		0x1000
 #define CMDQUE_CONTROL		0x1008
 #define INTR_STATUS		0x1018
-#define INT_ENB		0x1040
+#define INT_ENB			0x1040
 #define CONFIG			0x1044
 #define IRAM_ACCESS_CFG		0x10A0
 #define SECURE_DEST_ADDR	0x1100