diff --git a/arch/arc/Kconfig b/arch/arc/Kconfig
index 54b4427d826..5941a87ccfd 100644
--- a/arch/arc/Kconfig
+++ b/arch/arc/Kconfig
@@ -41,6 +41,7 @@ menu "ARCv2 Family Options"
 config	CPU_ARCV2
 	bool
 	select		ARCH_HAS_STACK_PROTECTION
+	select ARCH_HAS_USERSPACE if ARC_CORE_MPU
 	default y
 	help
 	  This option signifies the use of a CPU of the ARCv2 family.
diff --git a/arch/arc/core/CMakeLists.txt b/arch/arc/core/CMakeLists.txt
index 6fd3c7159d8..1427e063218 100644
--- a/arch/arc/core/CMakeLists.txt
+++ b/arch/arc/core/CMakeLists.txt
@@ -22,3 +22,4 @@ zephyr_sources_ifdef(CONFIG_ARC_FIRQ fast_irq.S)
 zephyr_sources_if_kconfig(irq_offload.c)
 zephyr_sources_ifdef(CONFIG_ATOMIC_OPERATIONS_CUSTOM atomic.c)
 add_subdirectory_ifdef(CONFIG_CPU_HAS_MPU mpu)
+zephyr_sources_ifdef(CONFIG_USERSPACE userspace.S)
diff --git a/arch/arc/core/fast_irq.S b/arch/arc/core/fast_irq.S
index 7ce27fe6e02..51f80c9617b 100644
--- a/arch/arc/core/fast_irq.S
+++ b/arch/arc/core/fast_irq.S
@@ -257,21 +257,13 @@ _firq_reschedule:
 	 */
 	_load_callee_saved_regs
 
-#ifdef CONFIG_MPU_STACK_GUARD
+#if defined(CONFIG_MPU_STACK_GUARD) || defined(CONFIG_USERSPACE)
 	push_s r2
 	mov r0, r2
-	bl configure_mpu_stack_guard
+	bl configure_mpu_thread
 	pop_s r2
 #endif
 
-#ifdef CONFIG_USERSPACE
-	push_s r2
-	mov r0, r2
-	bl configure_mpu_mem_domain
-	pop_s r2
-#endif
-
-
 	ld_s r3, [r2, _thread_offset_to_relinquish_cause]
 
 	breq r3, _CAUSE_RIRQ, _firq_return_from_rirq
diff --git a/arch/arc/core/fatal.c b/arch/arc/core/fatal.c
index 9ae2548277b..341616f1f61 100644
--- a/arch/arc/core/fatal.c
+++ b/arch/arc/core/fatal.c
@@ -84,3 +84,10 @@ FUNC_NORETURN void _NanoFatalErrorHandler(unsigned int reason,
 	for (;;)
 		;
 }
+
+
+FUNC_NORETURN void _arch_syscall_oops(void *ssf_ptr)
+{
+	_SysFatalErrorHandler(_NANO_ERR_KERNEL_OOPS, ssf_ptr);
+	CODE_UNREACHABLE;
+}
diff --git a/arch/arc/core/fault_s.S b/arch/arc/core/fault_s.S
index c6af71db13e..f04832fa221 100644
--- a/arch/arc/core/fault_s.S
+++ b/arch/arc/core/fault_s.S
@@ -47,7 +47,7 @@ SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_tlb_miss_d)
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_prot_v)
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_privilege_v)
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_swi)
-#ifndef CONFIG_IRQ_OFFLOAD
+#if !defined(CONFIG_IRQ_OFFLOAD) && !defined(CONFIG_USERSPACE)
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_trap)
 #endif
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_extension)
@@ -109,8 +109,38 @@ exc_nest_handle:
 
 #ifdef CONFIG_IRQ_OFFLOAD
 GTEXT(_irq_do_offload);
+#endif
 
+#if defined(CONFIG_IRQ_OFFLOAD) || defined(CONFIG_USERSPACE)
 SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_trap)
+#ifdef CONFIG_USERSPACE
+	/* get the id of trap_s */
+	lr ilink, [_ARC_V2_ECR]
+	and ilink, ilink, 0x3f
+	cmp ilink, 0x3
+	bne _do_other_trap
+/* do sys_call */
+	mov ilink, _SYSCALL_LIMIT
+	cmp r6, ilink
+	blt valid_syscall_id
+
+	mov r0, r6
+	mov r6, _SYSCALL_BAD
+
+valid_syscall_id:
+	lr ilink, [_ARC_V2_ERET]
+	push ilink
+	lr ilink, [_ARC_V2_ERSTATUS]
+	push ilink
+
+	bclr ilink, ilink, _ARC_V2_STATUS32_U_BIT
+	sr ilink, [_ARC_V2_ERSTATUS]
+
+	mov ilink, _arc_do_syscall
+	sr ilink, [_ARC_V2_ERET]
+
+	rtie
+
 	/*
 	 * Before invoking exception handler, the kernel switches to an exception
 	 * stack to save the faulting thread's registers.
@@ -118,6 +148,8 @@ SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_trap)
 	 * a diagnostic message and halt.
 	 */
 
+_do_other_trap:
+#endif /* CONFIG_USERSPACE */
 #ifdef CONFIG_ARC_STACK_CHECKING
 	push_s r2
 	/* disable stack checking */
@@ -148,7 +180,9 @@ SECTION_SUBSEC_FUNC(TEXT,__fault,__ev_trap)
 trap_nest_handle:
 	push_s r0
 
+#ifdef CONFIG_IRQ_OFFLOAD
 	jl _irq_do_offload
+#endif
 
 	pop sp
 
@@ -200,5 +234,4 @@ _trap_check_for_swap:
 
 	/* Assumption: r2 has current thread */
 	b _rirq_common_interrupt_swap
-
-#endif /* CONFIG_IRQ_OFFLOAD */
+#endif /* CONFIG_IRQ_OFFLOAD || CONFIG_USERSPACE */
diff --git a/arch/arc/core/isr_wrapper.S b/arch/arc/core/isr_wrapper.S
index ad5ef5ac4c4..7899386c7db 100644
--- a/arch/arc/core/isr_wrapper.S
+++ b/arch/arc/core/isr_wrapper.S
@@ -229,6 +229,13 @@ From RIRQ:
  */
 
 SECTION_FUNC(TEXT, _isr_wrapper)
+#if CONFIG_USERSPACE
+/* utilize the fact that Z bit is set if interrupt taken in U mode*/
+	bnz _isr_from_privilege
+/* get the correct stack pointer, don't touch _ARC_V2_USER_SP in the future */
+	aex sp, [_ARC_V2_USER_SP]
+_isr_from_privilege:
+#endif
 #if CONFIG_ARC_FIRQ
 #if CONFIG_RGF_NUM_BANKS == 1
 	st r0,[saved_r0]
diff --git a/arch/arc/core/mpu/arc_core_mpu.c b/arch/arc/core/mpu/arc_core_mpu.c
index 16882a28063..f3b554a7c20 100644
--- a/arch/arc/core/mpu/arc_core_mpu.c
+++ b/arch/arc/core/mpu/arc_core_mpu.c
@@ -11,6 +11,27 @@
 #include <arch/arc/v2/mpu/arc_core_mpu.h>
 #include <logging/sys_log.h>
 
+/*
+ * @brief Configure MPU for the thread
+ *
+ * This function configures per thread memory map reprogramming the MPU.
+ *
+ * @param thread thread info data structure.
+ */
+void configure_mpu_thread(struct k_thread *thread)
+{
+	arc_core_mpu_disable();
+#if defined(CONFIG_MPU_STACK_GUARD)
+	configure_mpu_stack_guard(thread);
+#endif
+
+#if defined(CONFIG_USERSPACE)
+	configure_mpu_user_context(thread);
+	configure_mpu_mem_domain(thread);
+#endif
+	arc_core_mpu_enable();
+}
+
 #if defined(CONFIG_MPU_STACK_GUARD)
 /*
  * @brief Configure MPU stack guard
@@ -22,15 +43,44 @@
  */
 void configure_mpu_stack_guard(struct k_thread *thread)
 {
-	arc_core_mpu_disable();
+#if defined(CONFIG_USERSPACE)
+	if (!thread->arch.priv_stack_start) {
+		/* the areas before and after the user stack of thread is
+		 * kernel only. These area can be used as stack guard.
+		 * -----------------------
+		 * |  kernel only access |
+		 * |---------------------|
+		 * |  user stack         |
+		 * |----------------------
+		 * |  privilege stack    |
+		 * -----------------------
+		 */
+		return;
+	}
+#endif
 	arc_core_mpu_configure(THREAD_STACK_GUARD_REGION,
 			       thread->stack_info.start - STACK_GUARD_SIZE,
 			       STACK_GUARD_SIZE);
-	arc_core_mpu_enable();
+
 }
 #endif
 
 #if defined(CONFIG_USERSPACE)
+/*
+ * @brief Configure MPU user context
+ *
+ * This function configures the thread's user context.
+ * The functionality is meant to be used during context switch.
+ *
+ * @param thread thread info data structure.
+ */
+void configure_mpu_user_context(struct k_thread *thread)
+{
+	SYS_LOG_DBG("configure user thread %p's context", thread);
+	arc_core_mpu_configure_user_context(thread);
+}
+
+
 /*
  * @brief Configure MPU memory domain
  *
@@ -42,9 +92,7 @@ void configure_mpu_stack_guard(struct k_thread *thread)
 void configure_mpu_mem_domain(struct k_thread *thread)
 {
 	SYS_LOG_DBG("configure thread %p's domain", thread);
-	arc_core_mpu_disable();
 	arc_core_mpu_configure_mem_domain(thread->mem_domain_info.mem_domain);
-	arc_core_mpu_enable();
 }
 
 int _arch_mem_domain_max_partitions_get(void)
diff --git a/arch/arc/core/mpu/arc_mpu.c b/arch/arc/core/mpu/arc_mpu.c
index 1f282d91ebc..c64dfd35413 100644
--- a/arch/arc/core/mpu/arc_mpu.c
+++ b/arch/arc/core/mpu/arc_mpu.c
@@ -11,6 +11,7 @@
 #include <arch/arc/v2/aux_regs.h>
 #include <arch/arc/v2/mpu/arc_mpu.h>
 #include <arch/arc/v2/mpu/arc_core_mpu.h>
+#include <linker/linker-defs.h>
 #include <logging/sys_log.h>
 
 
@@ -66,8 +67,12 @@ static inline u8_t _get_num_regions(void)
 static inline u32_t _get_region_attr_by_type(u32_t type)
 {
 	switch (type) {
+	case THREAD_STACK_USER_REGION:
+		return REGION_RAM_ATTR;
 	case THREAD_STACK_REGION:
-		return 0;
+		return  AUX_MPU_RDP_KW | AUX_MPU_RDP_KR;
+	case THREAD_APP_DATA_REGION:
+		return REGION_RAM_ATTR;
 	case THREAD_STACK_GUARD_REGION:
 	/* no Write and Execute to guard region */
 		return AUX_MPU_RDP_UR | AUX_MPU_RDP_KR;
@@ -161,8 +166,11 @@ static inline u32_t _get_region_index_by_type(u32_t type)
 	 */
 	switch (type) {
 #if CONFIG_ARC_MPU_VER  == 2
+	case THREAD_STACK_USER_REGION:
+		return _get_num_regions() - mpu_config.num_regions
+		 - THREAD_STACK_REGION;
 	case THREAD_STACK_REGION:
-		return _get_num_regions() - mpu_config.num_regions - type;
+	case THREAD_APP_DATA_REGION:
 	case THREAD_STACK_GUARD_REGION:
 		return _get_num_regions() - mpu_config.num_regions - type;
 	case THREAD_DOMAIN_PARTITION_REGION:
@@ -176,8 +184,10 @@ static inline u32_t _get_region_index_by_type(u32_t type)
 		return _get_num_regions() - mpu_config.num_regions - type + 1;
 #endif
 #elif CONFIG_ARC_MPU_VER == 3
+	case THREAD_STACK_USER_REGION:
+		return mpu_config.num_regions + THREAD_STACK_REGION - 1;
 	case THREAD_STACK_REGION:
-		return mpu_config.num_regions + type - 1;
+	case THREAD_APP_DATA_REGION:
 	case THREAD_STACK_GUARD_REGION:
 		return mpu_config.num_regions + type - 1;
 	case THREAD_DOMAIN_PARTITION_REGION:
@@ -417,6 +427,37 @@ void arc_core_mpu_region(u32_t index, u32_t base, u32_t size,
 }
 
 #if defined(CONFIG_USERSPACE)
+void arc_core_mpu_configure_user_context(struct k_thread *thread)
+{
+	u32_t base = (u32_t)thread->stack_obj;
+	u32_t size = thread->stack_info.size;
+
+	/* for kernel threads, no need to configure user context */
+	if (!thread->arch.priv_stack_start) {
+		return;
+	}
+
+	arc_core_mpu_configure(THREAD_STACK_USER_REGION, base, size);
+
+	/* configure app data portion */
+#ifdef CONFIG_APPLICATION_MEMORY
+#if CONFIG_ARC_MPU_VER == 2
+	base = (u32_t)&__app_ram_start;
+	size = (u32_t)&__app_ram_end - (u32_t)&__app_ram_start;
+
+	/* set up app data region if exists, otherwise disable */
+	if (size > 0) {
+		arc_core_mpu_configure(THREAD_APP_DATA_REGION, base, size);
+	}
+#elif CONFIG_ARC_MPU_VER == 3
+	/*
+	 * ARC MPV v3 doesn't support MPU region overlap.
+	 * Application memory should be a static memory, defined in mpu_config
+	 */
+#endif
+#endif
+}
+
 /**
  * @brief configure MPU regions for the memory partitions of the memory domain
  *
diff --git a/arch/arc/core/offsets/offsets.c b/arch/arc/core/offsets/offsets.c
index fd2152431db..f9a57920689 100644
--- a/arch/arc/core/offsets/offsets.c
+++ b/arch/arc/core/offsets/offsets.c
@@ -79,6 +79,9 @@ GEN_OFFSET_SYM(_callee_saved_stack_t, r24);
 GEN_OFFSET_SYM(_callee_saved_stack_t, r25);
 GEN_OFFSET_SYM(_callee_saved_stack_t, r26);
 GEN_OFFSET_SYM(_callee_saved_stack_t, fp);
+#ifdef CONFIG_USERSPACE
+GEN_OFFSET_SYM(_callee_saved_stack_t, user_sp);
+#endif
 GEN_OFFSET_SYM(_callee_saved_stack_t, r30);
 #ifdef CONFIG_FP_SHARING
 GEN_OFFSET_SYM(_callee_saved_stack_t, r58);
diff --git a/arch/arc/core/regular_irq.S b/arch/arc/core/regular_irq.S
index 643f34a1839..f8df077c215 100644
--- a/arch/arc/core/regular_irq.S
+++ b/arch/arc/core/regular_irq.S
@@ -161,17 +161,10 @@ _rirq_common_interrupt_swap:
 	 */
 	_load_callee_saved_regs
 
-#ifdef CONFIG_MPU_STACK_GUARD
+#if defined(CONFIG_MPU_STACK_GUARD) || defined(CONFIG_USERSPACE)
 	push_s r2
 	mov r0, r2
-	bl configure_mpu_stack_guard
-	pop_s r2
-#endif
-
-#ifdef CONFIG_USERSPACE
-	push_s r2
-	mov r0, r2
-	bl configure_mpu_mem_domain
+	bl configure_mpu_thread
 	pop_s r2
 #endif
 
diff --git a/arch/arc/core/swap.S b/arch/arc/core/swap.S
index 2f9ac8e8a5b..700bb4783f4 100644
--- a/arch/arc/core/swap.S
+++ b/arch/arc/core/swap.S
@@ -110,17 +110,10 @@ SECTION_FUNC(TEXT, __swap)
 
 	_load_callee_saved_regs
 
-#ifdef CONFIG_MPU_STACK_GUARD
+#if defined(CONFIG_MPU_STACK_GUARD) || defined(CONFIG_USERSPACE)
 	push_s r2
 	mov r0, r2
-	bl configure_mpu_stack_guard
-	pop_s r2
-#endif
-
-#ifdef CONFIG_USERSPACE
-	push_s r2
-	mov r0, r2
-	bl configure_mpu_mem_domain
+	bl configure_mpu_thread
 	pop_s r2
 #endif
 
diff --git a/arch/arc/core/thread.c b/arch/arc/core/thread.c
index 42c763ca670..aacfcb779d6 100644
--- a/arch/arc/core/thread.c
+++ b/arch/arc/core/thread.c
@@ -64,14 +64,41 @@ void _new_thread(struct k_thread *thread, k_thread_stack_t *stack,
 
 	char *stackEnd = pStackMem + stackSize;
 	struct init_stack_frame *pInitCtx;
-
+#if CONFIG_USERSPACE
+	/* for kernel thread, the privilege stack is merged into thread stack */
+	if (!(options & K_USER)) {
+	/* if MPU_STACK_GUARD is enabled, reserve the the stack area
+	 * |---------------------|----------------|
+	 * |                     |(MPU STACK AREA)|
+	 * |			 |----------------|
+	 * |  user stack         | kernel thread  |
+	 * |---------------------| stack          |
+	 * |  privilege stack    |                |
+	 * ----------------------------------------
+	 */
+#ifdef CONFIG_MPU_STACK_GUARD
+		pStackmem += STACK_GUARD_SIZE;
+		stackSize = stackSize + CONFIG_PRIVILEGED_STACK_SIZE
+			 - STACK_GUARD_SIZE;
+#endif
+		stackEnd += CONFIG_PRIVILEGED_STACK_SIZE;
+		stackSize += CONFIG_PRIVILEGED_STACK_SIZE;
+	}
+#endif
 	_new_thread_init(thread, pStackMem, stackSize, priority, options);
 
 	/* carve the thread entry struct from the "base" of the stack */
 	pInitCtx = (struct init_stack_frame *)(STACK_ROUND_DOWN(stackEnd) -
 				       sizeof(struct init_stack_frame));
-
+#if CONFIG_USERSPACE
+	if (options & K_USER) {
+		pInitCtx->pc = ((u32_t)_arch_user_mode_enter);
+	} else {
+		pInitCtx->pc = ((u32_t)_thread_entry_wrapper);
+	}
+#else
 	pInitCtx->pc = ((u32_t)_thread_entry_wrapper);
+#endif
 	pInitCtx->r0 = (u32_t)pEntry;
 	pInitCtx->r1 = (u32_t)parameter1;
 	pInitCtx->r2 = (u32_t)parameter2;
@@ -84,12 +111,24 @@ void _new_thread(struct k_thread *thread, k_thread_stack_t *stack,
 	 * value.
 	 */
 #ifdef CONFIG_ARC_STACK_CHECKING
-	pInitCtx->status32 = _ARC_V2_STATUS32_SC | _ARC_V2_STATUS32_E(_ARC_V2_DEF_IRQ_LEVEL);
+	pInitCtx->status32 = _ARC_V2_STATUS32_SC |
+		 _ARC_V2_STATUS32_E(_ARC_V2_DEF_IRQ_LEVEL);
 	thread->arch.stack_base = (u32_t) stackEnd;
 #else
 	pInitCtx->status32 = _ARC_V2_STATUS32_E(_ARC_V2_DEF_IRQ_LEVEL);
 #endif
 
+#if CONFIG_USERSPACE
+	if (options & K_USER) {
+		thread->arch.priv_stack_start = (u32_t) stackEnd;
+		thread->arch.priv_stack_size =
+			(u32_t)CONFIG_PRIVILEGED_STACK_SIZE;
+	} else {
+		thread->arch.priv_stack_start = 0;
+		thread->arch.priv_stack_size = 0;
+	}
+#endif
+
 #ifdef CONFIG_THREAD_MONITOR
 	/*
 	 * In debug mode thread->entry give direct access to the thread entry
@@ -113,3 +152,17 @@ void _new_thread(struct k_thread *thread, k_thread_stack_t *stack,
 
 	thread_monitor_init(thread);
 }
+
+
+#ifdef CONFIG_USERSPACE
+
+FUNC_NORETURN void _arch_user_mode_enter(k_thread_entry_t user_entry,
+	void *p1, void *p2, void *p3)
+{
+	_arc_userspace_enter(user_entry, p1, p2, p3,
+			     (u32_t)_current->stack_obj,
+			     _current->stack_info.size);
+	CODE_UNREACHABLE;
+}
+
+#endif
diff --git a/arch/arc/core/thread_entry_wrapper.S b/arch/arc/core/thread_entry_wrapper.S
index 08769315b78..17026134392 100644
--- a/arch/arc/core/thread_entry_wrapper.S
+++ b/arch/arc/core/thread_entry_wrapper.S
@@ -15,7 +15,6 @@
 #include <linker/sections.h>
 
 GTEXT(_thread_entry_wrapper)
-GTEXT(_thread_entry)
 
 /*
  * @brief Wrapper for _thread_entry
diff --git a/arch/arc/core/userspace.S b/arch/arc/core/userspace.S
new file mode 100644
index 00000000000..44f51a45b20
--- /dev/null
+++ b/arch/arc/core/userspace.S
@@ -0,0 +1,118 @@
+/*
+ * Copyright (c) 2017 Synopsys.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#include <offsets_short.h>
+#include <toolchain.h>
+#include <linker/sections.h>
+#include <kernel_structs.h>
+#include <arch/cpu.h>
+#include <syscall.h>
+
+GTEXT(_arc_userspace_enter)
+GTEXT(_arc_do_syscall)
+GTEXT(_arch_is_user_context)
+
+
+/**
+ *
+ * User space entry function
+ *
+ * This function is the entry point to user mode from privileged execution.
+ * The conversion is one way, and threads which transition to user mode do
+ * not transition back later, unless they are doing system calls.
+ *
+ */
+SECTION_FUNC(TEXT, _arc_userspace_enter)
+	/*
+	 * In ARCv2, the U bit can only be set through exception return
+	 */
+	/* the end of user stack in r5 */
+	add r5, r4, r5
+	/* start of privilege stack */
+	add r2, r5, CONFIG_PRIVILEGED_STACK_SIZE
+	sub r5, r5, 16 /* skip r0, r1, r2, r3 */
+
+#ifdef CONFIG_INIT_STACKS
+	mov r0, 0xaaaaaaaa
+#else
+	mov r0, 0x0
+#endif
+_clear_user_stack:
+	st.ab r0, [r4, 4]
+	cmp r4, r5
+	jlt _clear_user_stack
+
+	lr r0, [_ARC_V2_STATUS32]
+	bset r0, r0, _ARC_V2_STATUS32_U_BIT
+
+	mov r1, _thread_entry_wrapper
+
+	/* fake exception return */
+	kflag _ARC_V2_STATUS32_AE
+
+	sr r0, [_ARC_V2_ERSTATUS]
+	sr r1, [_ARC_V2_ERET]
+
+	/* when exception returns from kernel to user, sp and _ARC_V2_USER_SP
+	 * will be switched
+	 */
+	sr r5, [_ARC_V2_USER_SP]
+	mov sp, r2
+
+	rtie
+
+/**
+ *
+ * Userspace system call function
+ *
+ * This function is used to do system calls from unprivileged code.  This
+ * function is responsible for the following:
+ * 1) Dispatching the system call
+ * 2) Restoring stack and calling back to the caller of the system call
+ *
+ */
+SECTION_FUNC(TEXT, _arc_do_syscall)
+	/* r0-r5: arg1-arg6, r6 is call id */
+	/* the call id is already checked in trap_s handler */
+	push_s blink
+	push_s r0
+
+	mov r0, _k_syscall_table
+	ld.as blink, [r0, r6]
+
+	pop_s r0
+
+	jl [blink]
+
+
+	pop_s blink
+
+
+	/* through fake exception return, go back to the caller */
+	kflag _ARC_V2_STATUS32_AE
+
+	/* the status and return addesss are saved in trap_s handler */
+	pop r6
+	sr r6, [_ARC_V2_ERSTATUS]
+	pop r6
+	sr r6, [_ARC_V2_ERET]
+	rtie
+
+SECTION_FUNC(TEXT, _arch_is_user_context)
+	lr r0, [_ARC_V2_STATUS32]
+	bbit1 r0, 20, 1f
+	bset r1, r0, 20
+	bclr r1, r1, 31
+	kflag r1
+	lr r1, [_ARC_V2_STATUS32]
+	bbit0 r1, 20, 2f
+	kflag r0
+1:
+	j_s.d [blink]
+	mov r0, 0
+2:
+	j_s.d [blink]
+	mov r0, 1
diff --git a/arch/arc/include/kernel_arch_data.h b/arch/arc/include/kernel_arch_data.h
index 34cf9bb0fc6..de436d4fea6 100644
--- a/arch/arc/include/kernel_arch_data.h
+++ b/arch/arc/include/kernel_arch_data.h
@@ -129,6 +129,10 @@ struct _callee_saved_stack {
 	u32_t r25;
 	u32_t r26;
 	u32_t fp; /* r27 */
+
+#ifdef CONFIG_USERSPACE
+	u32_t user_sp;
+#endif
 	/* r28 is the stack pointer and saved separately */
 	/* r29 is ILINK and does not need to be saved */
 	u32_t r30;
diff --git a/arch/arc/include/kernel_arch_func.h b/arch/arc/include/kernel_arch_func.h
index 9a5808df7e6..760b9326a0f 100644
--- a/arch/arc/include/kernel_arch_func.h
+++ b/arch/arc/include/kernel_arch_func.h
@@ -78,6 +78,9 @@ static inline void _IntLibInit(void)
 	/* nothing needed, here because the kernel requires it */
 }
 
+extern void _arc_userspace_enter(k_thread_entry_t user_entry, void *p1,
+		 void *p2, void *p3, u32_t stack, u32_t size);
+
 #endif /* _ASMLANGUAGE */
 
 #ifdef __cplusplus
diff --git a/arch/arc/include/kernel_arch_thread.h b/arch/arc/include/kernel_arch_thread.h
index 8a4ddca8837..1af17e04e55 100644
--- a/arch/arc/include/kernel_arch_thread.h
+++ b/arch/arc/include/kernel_arch_thread.h
@@ -63,6 +63,11 @@ struct _thread_arch {
 	 */
 	u32_t stack_base;
 #endif
+
+#ifdef CONFIG_USERSPACE
+	u32_t priv_stack_start;
+	u32_t priv_stack_size;
+#endif
 };
 
 typedef struct _thread_arch _thread_arch_t;
diff --git a/arch/arc/include/swap_macros.h b/arch/arc/include/swap_macros.h
index f1da57c07b9..8a9c75e67ac 100644
--- a/arch/arc/include/swap_macros.h
+++ b/arch/arc/include/swap_macros.h
@@ -41,6 +41,11 @@ extern "C" {
 	st r25, [sp, ___callee_saved_stack_t_r25_OFFSET]
 	st r26, [sp, ___callee_saved_stack_t_r26_OFFSET]
 	st fp,  [sp, ___callee_saved_stack_t_fp_OFFSET]
+
+#ifdef CONFIG_USERSPACE
+	lr r13, [_ARC_V2_USER_SP]
+	st r13, [sp, ___callee_saved_stack_t_user_sp_OFFSET]
+#endif
 	st r30, [sp, ___callee_saved_stack_t_r30_OFFSET]
 
 #ifdef CONFIG_FP_SHARING
@@ -93,6 +98,11 @@ extern "C" {
 	sr r13, [_ARC_V2_FPU_DPFP2H]
 #endif
 
+#endif
+
+#ifdef CONFIG_USERSPACE
+	ld_s r13, [sp, ___callee_saved_stack_t_user_sp_OFFSET]
+	sr r13, [_ARC_V2_USER_SP]
 #endif
 
 	ld_s r13, [sp, ___callee_saved_stack_t_r13_OFFSET]
diff --git a/arch/arc/include/v2/irq.h b/arch/arc/include/v2/irq.h
index 18da988cf66..7bdeeedb257 100644
--- a/arch/arc/include/v2/irq.h
+++ b/arch/arc/include/v2/irq.h
@@ -21,6 +21,7 @@ extern "C" {
 
 #define _ARC_V2_AUX_IRQ_CTRL_BLINK (1 << 9)
 #define _ARC_V2_AUX_IRQ_CTRL_LOOP_REGS (1 << 10)
+#define _ARC_V2_AUX_IRQ_CTRL_U (1 << 11)
 #define _ARC_V2_AUX_IRQ_CTRL_LP (1 << 13)
 #define _ARC_V2_AUX_IRQ_CTRL_14_REGS 7
 #define _ARC_V2_AUX_IRQ_CTRL_16_REGS 8
@@ -45,6 +46,9 @@ static ALWAYS_INLINE void _irq_setup(void)
 		_ARC_V2_AUX_IRQ_CTRL_LOOP_REGS | /* save lp_xxx registers */
 #ifdef CONFIG_CODE_DENSITY
 		_ARC_V2_AUX_IRQ_CTRL_LP | /* save code density registers */
+#endif
+#ifdef CONFIG_USERSPACE
+		_ARC_V2_AUX_IRQ_CTRL_U | /* save context into user stack */
 #endif
 		_ARC_V2_AUX_IRQ_CTRL_BLINK     | /* save blink */
 		_ARC_V2_AUX_IRQ_CTRL_14_REGS     /* save r0 -> r13 (caller-saved) */
diff --git a/arch/arc/soc/em7d/CMakeLists.txt b/arch/arc/soc/em7d/CMakeLists.txt
index 1ea2570a48b..0954e6599ff 100644
--- a/arch/arc/soc/em7d/CMakeLists.txt
+++ b/arch/arc/soc/em7d/CMakeLists.txt
@@ -12,5 +12,3 @@ zephyr_sources(
   soc.c
   soc_config.c
   )
-
-zephyr_sources_ifdef(CONFIG_ARC_MPU_ENABLE arc_mpu_regions.c)
diff --git a/arch/arc/soc/em7d/arc_mpu_regions.c b/arch/arc/soc/em7d/arc_mpu_regions.c
deleted file mode 100644
index b8d23a125c6..00000000000
--- a/arch/arc/soc/em7d/arc_mpu_regions.c
+++ /dev/null
@@ -1,42 +0,0 @@
-/*
- * Copyright (c) 2017 Synopsys
- *
- * SPDX-License-Identifier: Apache-2.0
- */
-
-#include <soc.h>
-#include <arch/arc/v2/mpu/arc_mpu.h>
-
-static struct arc_mpu_region mpu_regions[] = {
-#if CONFIG_ICCM_SIZE > 0
-	/* Region ICCM */
-	MPU_REGION_ENTRY("ICCM",
-			 CONFIG_ICCM_BASE_ADDRESS,
-			 CONFIG_ICCM_SIZE * 1024,
-			 REGION_FLASH_ATTR),
-#endif
-#if CONFIG_DCCM_SIZE > 0
-	/* Region DCCM */
-	MPU_REGION_ENTRY("DCCM",
-			 CONFIG_DCCM_BASE_ADDRESS,
-			 CONFIG_DCCM_SIZE * 1024,
-			 REGION_RAM_ATTR),
-#endif
-#if CONFIG_SRAM_SIZE > 0
-	/* Region DDR RAM */
-	MPU_REGION_ENTRY("DDR RAM",
-			CONFIG_SRAM_BASE_ADDRESS,
-			CONFIG_SRAM_SIZE * 1024,
-			REGION_ALL_ATTR),
-#endif
-	/* Region Peripheral */
-	MPU_REGION_ENTRY("PERIPHERAL",
-			 0xF0000000,
-			 64 * 1024,
-			 REGION_IO_ATTR),
-};
-
-struct arc_mpu_config mpu_config = {
-	.num_regions = ARRAY_SIZE(mpu_regions),
-	.mpu_regions = mpu_regions,
-};
diff --git a/boards/arc/em_starterkit/CMakeLists.txt b/boards/arc/em_starterkit/CMakeLists.txt
new file mode 100644
index 00000000000..7a9d4e7049d
--- /dev/null
+++ b/boards/arc/em_starterkit/CMakeLists.txt
@@ -0,0 +1 @@
+zephyr_sources_ifdef(CONFIG_ARC_MPU_ENABLE arc_mpu_regions.c)
\ No newline at end of file
diff --git a/boards/arc/em_starterkit/arc_mpu_regions.c b/boards/arc/em_starterkit/arc_mpu_regions.c
new file mode 100644
index 00000000000..65ff90ada6f
--- /dev/null
+++ b/boards/arc/em_starterkit/arc_mpu_regions.c
@@ -0,0 +1,92 @@
+/*
+ * Copyright (c) 2017 Synopsys
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#include <soc.h>
+#include <arch/arc/v2/mpu/arc_mpu.h>
+#include <linker/linker-defs.h>
+
+#ifdef CONFIG_USERSPACE
+static struct arc_mpu_region mpu_regions[] = {
+#if CONFIG_ARC_MPU_VER == 3 && defined(CONFIG_APPLICATION_MEMORY)
+	/* Region ICCM */
+	MPU_REGION_ENTRY("IMAGE ROM",
+			 _image_rom_start,
+			 _image_rom_end,
+			 REGION_FLASH_ATTR),
+	MPU_REGION_ENTRY("APP MEMORY",
+			 __app_ram_start,
+			 __app_ram_size,
+			 REGION_RAM_ATTR),
+	MPU_REGION_ENTRY("KERNEL MEMORY",
+			 __kernel_ram_start,
+			 __kernel_ram_size,
+			 AUX_MPU_RDP_KW | AUX_MPU_RDP_KR),
+
+#else
+#if CONFIG_ICCM_SIZE > 0
+	/* Region ICCM */
+	MPU_REGION_ENTRY("ICCM",
+			 CONFIG_ICCM_BASE_ADDRESS,
+			 CONFIG_ICCM_SIZE * 1024,
+			 REGION_FLASH_ATTR),
+#endif
+#if CONFIG_DCCM_SIZE > 0
+	/* Region DCCM */
+	MPU_REGION_ENTRY("DCCM",
+			 CONFIG_DCCM_BASE_ADDRESS,
+			 CONFIG_DCCM_SIZE * 1024,
+			 AUX_MPU_RDP_KW | AUX_MPU_RDP_KR),
+#endif
+#if CONFIG_SRAM_SIZE > 0
+	/* Region DDR RAM */
+	MPU_REGION_ENTRY("DDR RAM",
+			CONFIG_SRAM_BASE_ADDRESS,
+			CONFIG_SRAM_SIZE * 1024,
+			AUX_MPU_RDP_KW | AUX_MPU_RDP_KR |
+			AUX_MPU_RDP_KE | AUX_MPU_RDP_UE),
+#endif
+#endif /* ARC_MPU_VER == 3 */
+	/* Region Peripheral */
+	MPU_REGION_ENTRY("PERIPHERAL",
+			 0xF0000000,
+			 64 * 1024,
+			 AUX_MPU_RDP_KW | AUX_MPU_RDP_KR),
+};
+#else
+static struct arc_mpu_region mpu_regions[] = {
+#if CONFIG_ICCM_SIZE > 0
+	/* Region ICCM */
+	MPU_REGION_ENTRY("ICCM",
+			 CONFIG_ICCM_BASE_ADDRESS,
+			 CONFIG_ICCM_SIZE * 1024,
+			 REGION_FLASH_ATTR),
+#endif
+#if CONFIG_DCCM_SIZE > 0
+	/* Region DCCM */
+	MPU_REGION_ENTRY("DCCM",
+			 CONFIG_DCCM_BASE_ADDRESS,
+			 CONFIG_DCCM_SIZE * 1024,
+			 REGION_RAM_ATTR),
+#endif
+#if CONFIG_SRAM_SIZE > 0
+	/* Region DDR RAM */
+	MPU_REGION_ENTRY("DDR RAM",
+			CONFIG_SRAM_BASE_ADDRESS,
+			CONFIG_SRAM_SIZE * 1024,
+			REGION_ALL_ATTR),
+#endif
+	/* Region Peripheral */
+	MPU_REGION_ENTRY("PERIPHERAL",
+			 0xF0000000,
+			 64 * 1024,
+			 REGION_IO_ATTR),
+};
+#endif
+
+struct arc_mpu_config mpu_config = {
+	.num_regions = ARRAY_SIZE(mpu_regions),
+	.mpu_regions = mpu_regions,
+};
diff --git a/include/arch/arc/arch.h b/include/arch/arc/arch.h
index c799aa75c02..f37aeef9b3f 100644
--- a/include/arch/arc/arch.h
+++ b/include/arch/arc/arch.h
@@ -39,7 +39,7 @@ extern "C" {
 #include <arch/arc/v2/addr_types.h>
 #endif
 
-#if defined(CONFIG_MPU_STACK_GUARD)
+#if defined(CONFIG_MPU_STACK_GUARD) || defined(CONFIG_USERSPACE)
 #if defined(CONFIG_ARC_CORE_MPU)
 #if CONFIG_ARC_MPU_VER == 2
 /*
@@ -62,22 +62,93 @@ extern "C" {
 #define STACK_GUARD_SIZE 0
 #endif
 
+#define STACK_SIZE_ALIGN(x)	max(STACK_ALIGN, x)
+
+
+
+/**
+ * @brief Calculate power of two ceiling for a buffer size input
+ *
+ */
+#define POW2_CEIL(x) ((1 << (31 - __builtin_clz(x))) < x ?  \
+		1 << (31 - __builtin_clz(x) + 1) : \
+		1 << (31 - __builtin_clz(x)))
+
+#if defined(CONFIG_USERSPACE)
+
+/*
+ * if user space is enabled, for user thread, no STACK_GUARD area;
+ * for kernel thread, the privilege stack can be used as STACK_GUARD
+ * area, and the privilege stack size must be greater than STACK_GUARD_SIZE
+ */
+#if defined(CONFIG_MPU_STACK_GUARD) && \
+	CONFIG_PRIVILEGED_STACK_SIZE < STACK_GUARD_SIZE
+#error "CONFIG_PRIVILEGED_STACK_SIZE must be larger than STACK_GUARD_SIZE"
+#endif
+
+
+#if CONFIG_ARC_MPU_VER == 2
+
 #define _ARCH_THREAD_STACK_DEFINE(sym, size) \
-	struct _k_thread_stack_element __noinit __aligned(STACK_ALIGN) \
-		sym[size+STACK_GUARD_SIZE]
+	struct _k_thread_stack_element __kernel_noinit \
+		__aligned(POW2_CEIL(STACK_SIZE_ALIGN(size))) \
+		sym[POW2_CEIL(STACK_SIZE_ALIGN(size)) + \
+		CONFIG_PRIVILEGED_STACK_SIZE]
 
 #define _ARCH_THREAD_STACK_ARRAY_DEFINE(sym, nmemb, size) \
-	struct _k_thread_stack_element __noinit __aligned(STACK_ALIGN) \
-		sym[nmemb][size+STACK_GUARD_SIZE]
+	struct _k_thread_stack_element __kernel_noinit \
+		__aligned(POW2_CEIL(STACK_SIZE_ALIGN(size))) \
+		sym[nmemb][POW2_CEIL(STACK_SIZE_ALIGN(size)) + \
+		CONFIG_PRIVILEGED_STACK_SIZE]
+
+#define _ARCH_THREAD_STACK_MEMBER(sym, size) \
+	struct _k_thread_stack_element \
+		__aligned(POW2_CEIL(STACK_SIZE_ALIGN(size))) \
+		sym[POW2_CEIL(size) + CONFIG_PRIVILEGED_STACK_SIZE]
+
+#elif CONFIG_ARC_MPU_VER == 3
+
+#define _ARCH_THREAD_STACK_DEFINE(sym, size) \
+	struct _k_thread_stack_element __kernel_noinit __aligned(STACK_ALIGN) \
+		sym[size + CONFIG_PRIVILEGED_STACK_SIZE]
+
+#define _ARCH_THREAD_STACK_ARRAY_DEFINE(sym, nmemb, size) \
+	struct _k_thread_stack_element __kernel_noinit __aligned(STACK_ALIGN) \
+		sym[nmemb][size + CONFIG_PRIVILEGED_STACK_SIZE]
 
 #define _ARCH_THREAD_STACK_MEMBER(sym, size) \
 	struct _k_thread_stack_element __aligned(STACK_ALIGN) \
-		sym[size+STACK_GUARD_SIZE]
+		sym[size + CONFIG_PRIVILEGED_STACK_SIZE]
+
+#endif /* CONFIG_ARC_MPU_VER */
+
+#define _ARCH_THREAD_STACK_SIZEOF(sym) \
+		(sizeof(sym) - CONFIG_PRIVILEGED_STACK_SIZE)
+
+#define _ARCH_THREAD_STACK_BUFFER(sym) \
+		((char *)(sym))
+
+#else /* CONFIG_USERSPACE */
+
+#define _ARCH_THREAD_STACK_DEFINE(sym, size) \
+	struct _k_thread_stack_element __kernel_noinit __aligned(STACK_ALIGN) \
+		sym[size + STACK_GUARD_SIZE]
+
+#define _ARCH_THREAD_STACK_ARRAY_DEFINE(sym, nmemb, size) \
+	struct _k_thread_stack_element __kernel_noinit __aligned(STACK_ALIGN) \
+		sym[nmemb][size + STACK_GUARD_SIZE]
+
+#define _ARCH_THREAD_STACK_MEMBER(sym, size) \
+	struct _k_thread_stack_element __aligned(STACK_ALIGN) \
+		sym[size + STACK_GUARD_SIZE]
 
 #define _ARCH_THREAD_STACK_SIZEOF(sym) (sizeof(sym) - STACK_GUARD_SIZE)
 
 #define _ARCH_THREAD_STACK_BUFFER(sym) ((char *)(sym + STACK_GUARD_SIZE))
 
+#endif /* CONFIG_USERSPACE */
+
+
 #ifdef CONFIG_USERSPACE
 #ifdef CONFIG_ARC_MPU
 #ifndef _ASMLANGUAGE
@@ -158,46 +229,125 @@ static inline u32_t _arch_syscall_invoke6(u32_t arg1, u32_t arg2, u32_t arg3,
 					  u32_t arg4, u32_t arg5, u32_t arg6,
 					  u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r1 __asm__("r1") = arg2;
+	register u32_t r2 __asm__("r2") = arg3;
+	register u32_t r3 __asm__("r3") = arg4;
+	register u32_t r4 __asm__("r4") = arg5;
+	register u32_t r5 __asm__("r5") = arg6;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r1), "r" (r2), "r" (r3),
+			   "r" (r4), "r" (r5), "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke5(u32_t arg1, u32_t arg2, u32_t arg3,
 					  u32_t arg4, u32_t arg5, u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r1 __asm__("r1") = arg2;
+	register u32_t r2 __asm__("r2") = arg3;
+	register u32_t r3 __asm__("r3") = arg4;
+	register u32_t r4 __asm__("r4") = arg5;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r1), "r" (r2), "r" (r3),
+			   "r" (r4), "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke4(u32_t arg1, u32_t arg2, u32_t arg3,
 					  u32_t arg4, u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r1 __asm__("r1") = arg2;
+	register u32_t r2 __asm__("r2") = arg3;
+	register u32_t r3 __asm__("r3") = arg4;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r1), "r" (r2), "r" (r3),
+			   "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke3(u32_t arg1, u32_t arg2, u32_t arg3,
 					  u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r1 __asm__("r1") = arg2;
+	register u32_t r2 __asm__("r2") = arg3;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r1), "r" (r2), "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke2(u32_t arg1, u32_t arg2, u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r1 __asm__("r1") = arg2;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r1), "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke1(u32_t arg1, u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0") = arg1;
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r6));
+
+	return ret;
 }
 
 static inline u32_t _arch_syscall_invoke0(u32_t call_id)
 {
-	return 0;
+	register u32_t ret __asm__("r0");
+	register u32_t r6 __asm__("r6") = call_id;
+
+	__asm__ volatile(
+			 "trap_s %[trap_s_id]\n"
+			 : "=r"(ret)
+			 : [trap_s_id] "i" (_TRAP_S_CALL_SYSTEM_CALL),
+			   "r" (ret), "r" (r6));
+
+	return ret;
 }
 
-static inline int _arch_is_user_context(void)
-{
-	return 0;
-}
+extern int _arch_is_user_context(void);
+
 #endif /* _ASMLANGUAGE */
 #endif /* CONFIG_USERSPACE */
 #ifdef __cplusplus
diff --git a/include/arch/arc/v2/aux_regs.h b/include/arch/arc/v2/aux_regs.h
index 70f7f85c57a..010d270c299 100644
--- a/include/arch/arc/v2/aux_regs.h
+++ b/include/arch/arc/v2/aux_regs.h
@@ -25,6 +25,7 @@ extern "C" {
 #define _ARC_V2_SEC_STAT 0x09
 #define _ARC_V2_STATUS32 0x00a
 #define _ARC_V2_STATUS32_P0 0x00b
+#define _ARC_V2_USER_SP	0x00d
 #define _ARC_V2_AUX_IRQ_CTRL 0x00e
 #define _ARC_V2_IC_IVIC 0x010
 #define _ARC_V2_IC_CTRL 0x011
@@ -102,7 +103,8 @@ extern "C" {
 #define _ARC_V2_STATUS32_AE_BIT 5
 #define _ARC_V2_STATUS32_AE (1 << _ARC_V2_STATUS32_AE_BIT)
 #define _ARC_V2_STATUS32_DE (1 << 6)
-#define _ARC_V2_STATUS32_U (1 << 7)
+#define _ARC_V2_STATUS32_U_BIT 7
+#define _ARC_V2_STATUS32_U (1 << _ARC_V2_STATUS32_U_BIT)
 #define _ARC_V2_STATUS32_V (1 << 8)
 #define _ARC_V2_STATUS32_C (1 << 9)
 #define _ARC_V2_STATUS32_N (1 << 10)
diff --git a/include/arch/arc/v2/error.h b/include/arch/arc/v2/error.h
index 2dc517c1c8f..a05a0ad93bc 100644
--- a/include/arch/arc/v2/error.h
+++ b/include/arch/arc/v2/error.h
@@ -33,6 +33,11 @@ extern void _SysFatalErrorHandler(unsigned int cause, const NANO_ESF *esf);
 #define _NANO_ERR_KERNEL_OOPS (4)       /* Kernel oops (fatal to thread) */
 #define _NANO_ERR_KERNEL_PANIC (5)	/* Kernel panic (fatal to system) */
 
+
+#define _TRAP_S_SCALL_IRQ_OFFLOAD		1
+#define _TRAP_S_CALL_RUNTIME_EXCEPT		2
+#define _TRAP_S_CALL_SYSTEM_CALL		3
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/include/arch/arc/v2/linker.ld b/include/arch/arc/v2/linker.ld
index 56ab03603e0..769d5063f2e 100644
--- a/include/arch/arc/v2/linker.ld
+++ b/include/arch/arc/v2/linker.ld
@@ -37,6 +37,16 @@
 	#endif
 #endif
 
+#ifdef CONFIG_ARC_MPU_ENABLE
+	#if CONFIG_ARC_MPU_VER == 2
+		#define MPU_ADDR_ALIGN . = ALIGN(2048);
+	#elif CONFIG_ARC_MPU_VER == 3
+		#define MPU_ADDR_ALIGN . = ALIGN(32);
+	#endif
+#else
+	#define MPU_ADDR_ALIGN
+#endif
+
 #if defined(CONFIG_XIP)
 	#define _DATA_IN_ROM __data_rom_start
 #else
@@ -108,7 +118,9 @@ SECTIONS {
 	} GROUP_LINK_IN(ROMABLE_REGION)
 
 	_image_rodata_end = .;
+	MPU_ADDR_ALIGN
 	_image_rom_end = .;
+	_image_rom_size = _image_rom_end - _image_rom_start;
 
 	GROUP_END(ROMABLE_REGION)
 
@@ -117,6 +129,7 @@ SECTIONS {
 #ifdef CONFIG_APPLICATION_MEMORY
 	SECTION_DATA_PROLOGUE(_APP_DATA_SECTION_NAME, (OPTIONAL),)
 	{
+		MPU_ADDR_ALIGN
 		__app_ram_start = .;
 		__app_data_ram_start = .;
 		_image_ram_start = .;
@@ -142,12 +155,19 @@ SECTIONS {
 	{
 		APP_INPUT_SECTION(.noinit)
 		APP_INPUT_SECTION(".noinit.*")
+/*
+ * for MPU v2,the application memory section must be aligned to the size of
+ * section
+ */
+		MPU_ADDR_ALIGN
 	} GROUP_DATA_LINK_IN(RAMABLE_REGION, RAMABLE_REGION)
 
 	__app_ram_end = .;
+	__app_ram_size = __app_ram_end - __app_ram_start;
 #endif /* CONFIG_APPLICATION_MEMORY */
 
 	SECTION_DATA_PROLOGUE(_BSS_SECTION_NAME,(NOLOAD),) {
+		MPU_ADDR_ALIGN
 		/*
 		 * For performance, BSS section is assumed to be 4 byte aligned and
 		 * a multiple of 4 bytes
@@ -204,6 +224,8 @@ SECTIONS {
 
 	__data_ram_end = .;
 
+	MPU_ADDR_ALIGN
+
 	/* Define linker symbols */
 	_image_ram_end = .;
 	_end = .; /* end of image */
diff --git a/include/arch/arc/v2/mpu/arc_core_mpu.h b/include/arch/arc/v2/mpu/arc_core_mpu.h
index 9c67541eb07..89479e2717c 100644
--- a/include/arch/arc/v2/mpu/arc_core_mpu.h
+++ b/include/arch/arc/v2/mpu/arc_core_mpu.h
@@ -28,9 +28,11 @@ extern "C" {
  * be managed inside the MPU driver and not escalated.
  */
 /* Thread Region Intent Type */
+#define THREAD_STACK_USER_REGION 0x0
 #define THREAD_STACK_REGION 0x1
-#define THREAD_STACK_GUARD_REGION 0x2
-#define THREAD_DOMAIN_PARTITION_REGION 0x3
+#define THREAD_APP_DATA_REGION 0x2
+#define THREAD_STACK_GUARD_REGION 0x3
+#define THREAD_DOMAIN_PARTITION_REGION 0x4
 
 #if defined(CONFIG_ARC_CORE_MPU)
 /* ARC Core MPU Driver API */
@@ -95,7 +97,7 @@ void configure_mpu_stack_guard(struct k_thread *thread);
 #endif
 
 #if defined(CONFIG_USERSPACE)
-
+void arc_core_mpu_configure_user_context(struct k_thread *thread);
 void arc_core_mpu_configure_mem_domain(struct k_mem_domain *mem_domain);
 void arc_core_mpu_mem_partition_remove(u32_t part_index);
 void arc_core_mpu_configure_mem_partition(u32_t part_index,
@@ -112,8 +114,11 @@ int arc_core_mpu_buffer_validate(void *addr, size_t size, int write);
  * @param thread thread info data structure.
  */
 void configure_mpu_mem_domain(struct k_thread *thread);
+
+void configure_mpu_user_context(struct k_thread *thread);
 #endif
 
+void configure_mpu_thread(struct k_thread *thread);
 
 #ifdef __cplusplus
 }