diff --git a/libcpu/arm/cortex-a/cp15.h b/libcpu/arm/cortex-a/cp15.h
index 687199e573..6896fd9f71 100644
--- a/libcpu/arm/cortex-a/cp15.h
+++ b/libcpu/arm/cortex-a/cp15.h
@@ -10,6 +10,11 @@
 #ifndef __CP15_H__
 #define __CP15_H__
 
+#define __get_cp(cp, op1, Rt, CRn, CRm, op2) __asm__ volatile("MRC p" # cp ", " # op1 ", %0, c" # CRn ", c" # CRm ", " # op2 : "=r" (Rt) : : "memory" )
+#define __set_cp(cp, op1, Rt, CRn, CRm, op2) __asm__ volatile("MCR p" # cp ", " # op1 ", %0, c" # CRn ", c" # CRm ", " # op2 : : "r" (Rt) : "memory" )
+#define __get_cp64(cp, op1, Rt, CRm) __asm__ volatile("MRRC p" # cp ", " # op1 ", %Q0, %R0, c" # CRm  : "=r" (Rt) : : "memory" )
+#define __set_cp64(cp, op1, Rt, CRm) __asm__ volatile("MCRR p" # cp ", " # op1 ", %Q0, %R0, c" # CRm  : : "r" (Rt) : "memory" )
+
 unsigned long rt_cpu_get_smp_id(void);
 
 void rt_cpu_mmu_disable(void);
diff --git a/libcpu/arm/cortex-a/gic.c b/libcpu/arm/cortex-a/gic.c
index faad418b2d..60e924d405 100644
--- a/libcpu/arm/cortex-a/gic.c
+++ b/libcpu/arm/cortex-a/gic.c
@@ -28,29 +28,35 @@ struct arm_gic
 /* 'ARM_GIC_MAX_NR' is the number of cores */
 static struct arm_gic _gic_table[ARM_GIC_MAX_NR];
 
-#define GIC_CPU_CTRL(hw_base)               __REG32((hw_base) + 0x00)
-#define GIC_CPU_PRIMASK(hw_base)            __REG32((hw_base) + 0x04)
-#define GIC_CPU_BINPOINT(hw_base)           __REG32((hw_base) + 0x08)
-#define GIC_CPU_INTACK(hw_base)             __REG32((hw_base) + 0x0c)
-#define GIC_CPU_EOI(hw_base)                __REG32((hw_base) + 0x10)
-#define GIC_CPU_RUNNINGPRI(hw_base)         __REG32((hw_base) + 0x14)
-#define GIC_CPU_HIGHPRI(hw_base)            __REG32((hw_base) + 0x18)
+/** Macro to access the Generic Interrupt Controller Interface (GICC)
+*/
+#define GIC_CPU_CTRL(hw_base)               __REG32((hw_base) + 0x00U)
+#define GIC_CPU_PRIMASK(hw_base)            __REG32((hw_base) + 0x04U)
+#define GIC_CPU_BINPOINT(hw_base)           __REG32((hw_base) + 0x08U)
+#define GIC_CPU_INTACK(hw_base)             __REG32((hw_base) + 0x0cU)
+#define GIC_CPU_EOI(hw_base)                __REG32((hw_base) + 0x10U)
+#define GIC_CPU_RUNNINGPRI(hw_base)         __REG32((hw_base) + 0x14U)
+#define GIC_CPU_HIGHPRI(hw_base)            __REG32((hw_base) + 0x18U)
+#define GIC_CPU_IIDR(hw_base)               __REG32((hw_base) + 0xFCU)
 
-#define GIC_DIST_CTRL(hw_base)              __REG32((hw_base) + 0x000)
-#define GIC_DIST_TYPE(hw_base)              __REG32((hw_base) + 0x004)
-#define GIC_DIST_IGROUP(hw_base, n)         __REG32((hw_base) + 0x080 + ((n)/32) * 4)
-#define GIC_DIST_ENABLE_SET(hw_base, n)     __REG32((hw_base) + 0x100 + ((n)/32) * 4)
-#define GIC_DIST_ENABLE_CLEAR(hw_base, n)   __REG32((hw_base) + 0x180 + ((n)/32) * 4)
-#define GIC_DIST_PENDING_SET(hw_base, n)    __REG32((hw_base) + 0x200 + ((n)/32) * 4)
-#define GIC_DIST_PENDING_CLEAR(hw_base, n)  __REG32((hw_base) + 0x280 + ((n)/32) * 4)
-#define GIC_DIST_ACTIVE_SET(hw_base, n)     __REG32((hw_base) + 0x300 + ((n)/32) * 4)
-#define GIC_DIST_ACTIVE_CLEAR(hw_base, n)   __REG32((hw_base) + 0x380 + ((n)/32) * 4)
-#define GIC_DIST_PRI(hw_base, n)            __REG32((hw_base) + 0x400 +  ((n)/4) * 4)
-#define GIC_DIST_TARGET(hw_base, n)         __REG32((hw_base) + 0x800 +  ((n)/4) * 4)
-#define GIC_DIST_CONFIG(hw_base, n)         __REG32((hw_base) + 0xc00 + ((n)/16) * 4)
-#define GIC_DIST_SOFTINT(hw_base)           __REG32((hw_base) + 0xf00)
-#define GIC_DIST_CPENDSGI(hw_base, n)       __REG32((hw_base) + 0xf10 + ((n)/4) * 4)
-#define GIC_DIST_ICPIDR2(hw_base)           __REG32((hw_base) + 0xfe8)
+/** Macro to access the Generic Interrupt Controller Distributor (GICD)
+*/
+#define GIC_DIST_CTRL(hw_base)              __REG32((hw_base) + 0x000U)
+#define GIC_DIST_TYPE(hw_base)              __REG32((hw_base) + 0x004U)
+#define GIC_DIST_IGROUP(hw_base, n)         __REG32((hw_base) + 0x080U + ((n)/32U) * 4U)
+#define GIC_DIST_ENABLE_SET(hw_base, n)     __REG32((hw_base) + 0x100U + ((n)/32U) * 4U)
+#define GIC_DIST_ENABLE_CLEAR(hw_base, n)   __REG32((hw_base) + 0x180U + ((n)/32U) * 4U)
+#define GIC_DIST_PENDING_SET(hw_base, n)    __REG32((hw_base) + 0x200U + ((n)/32U) * 4U)
+#define GIC_DIST_PENDING_CLEAR(hw_base, n)  __REG32((hw_base) + 0x280U + ((n)/32U) * 4U)
+#define GIC_DIST_ACTIVE_SET(hw_base, n)     __REG32((hw_base) + 0x300U + ((n)/32U) * 4U)
+#define GIC_DIST_ACTIVE_CLEAR(hw_base, n)   __REG32((hw_base) + 0x380U + ((n)/32U) * 4U)
+#define GIC_DIST_PRI(hw_base, n)            __REG32((hw_base) + 0x400U +  ((n)/4U) * 4U)
+#define GIC_DIST_TARGET(hw_base, n)         __REG32((hw_base) + 0x800U +  ((n)/4U) * 4U)
+#define GIC_DIST_CONFIG(hw_base, n)         __REG32((hw_base) + 0xc00U + ((n)/16U) * 4U)
+#define GIC_DIST_SOFTINT(hw_base)           __REG32((hw_base) + 0xf00U)
+#define GIC_DIST_CPENDSGI(hw_base, n)       __REG32((hw_base) + 0xf10U + ((n)/4U) * 4U)
+#define GIC_DIST_SPENDSGI(hw_base, n)       __REG32((hw_base) + 0xf20U + ((n)/4U) * 4U)
+#define GIC_DIST_ICPIDR2(hw_base)           __REG32((hw_base) + 0xfe8U)
 
 static unsigned int _gic_max_irq;
 
@@ -67,12 +73,12 @@ int arm_gic_get_active_irq(rt_uint32_t index)
 
 void arm_gic_ack(rt_uint32_t index, int irq)
 {
-    rt_uint32_t mask = 1 << (irq % 32);
+    rt_uint32_t mask = 1U << (irq % 32U);
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
     GIC_DIST_PENDING_CLEAR(_gic_table[index].dist_hw_base, irq) = mask;
     GIC_CPU_EOI(_gic_table[index].cpu_hw_base) = irq;
@@ -80,36 +86,136 @@ void arm_gic_ack(rt_uint32_t index, int irq)
 
 void arm_gic_mask(rt_uint32_t index, int irq)
 {
-    rt_uint32_t mask = 1 << (irq % 32);
+    rt_uint32_t mask = 1U << (irq % 32U);
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
     GIC_DIST_ENABLE_CLEAR(_gic_table[index].dist_hw_base, irq) = mask;
 }
 
-void arm_gic_clear_pending(rt_uint32_t index, int irq)
+void arm_gic_umask(rt_uint32_t index, int irq)
 {
-    rt_uint32_t mask = 1 << (irq % 32);
+    rt_uint32_t mask = 1U << (irq % 32U);
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
-    GIC_DIST_PENDING_CLEAR(_gic_table[index].dist_hw_base, irq) = mask;
+    GIC_DIST_ENABLE_SET(_gic_table[index].dist_hw_base, irq) = mask;
+}
+
+rt_uint32_t arm_gic_get_pending_irq(rt_uint32_t index, int irq)
+{
+    rt_uint32_t pend;
+
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    if (irq >= 16U)
+    {
+        pend = (GIC_DIST_PENDING_SET(_gic_table[index].dist_hw_base, irq) >> (irq % 32U)) & 0x1UL;
+    }
+    else
+    {
+        /* INTID 0-15 Software Generated Interrupt */
+        pend = (GIC_DIST_SPENDSGI(_gic_table[index].dist_hw_base, irq) >> ((irq % 4U) * 8U)) & 0xFFUL;
+        /* No CPU identification offered */
+        if (pend != 0U)
+        {
+            pend = 1U;
+        }
+        else
+        {
+            pend = 0U;
+        }
+    }
+
+    return (pend);
+}
+
+void arm_gic_set_pending_irq(rt_uint32_t index, int irq)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    if (irq >= 16U)
+    {
+        GIC_DIST_PENDING_SET(_gic_table[index].dist_hw_base, irq) = 1U << (irq % 32U);
+    }
+    else
+    {
+        /* INTID 0-15 Software Generated Interrupt */
+        /* Forward the interrupt to the CPU interface that requested it */
+        GIC_DIST_SOFTINT(_gic_table[index].dist_hw_base) = (irq | 0x02000000U);
+    }
+}
+
+void arm_gic_clear_pending_irq(rt_uint32_t index, int irq)
+{
+    rt_uint32_t mask;
+
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    if (irq >= 16U)
+    {
+        mask = 1U << (irq % 32U);
+        GIC_DIST_PENDING_CLEAR(_gic_table[index].dist_hw_base, irq) = mask;
+    }
+    else
+    {
+        mask =  1U << ((irq % 4U) * 8U);
+        GIC_DIST_CPENDSGI(_gic_table[index].dist_hw_base, irq) = mask;
+    }
+}
+
+void arm_gic_set_configuration(rt_uint32_t index, int irq, uint32_t config)
+{
+    rt_uint32_t icfgr;
+    rt_uint32_t shift;
+
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    icfgr = GIC_DIST_CONFIG(_gic_table[index].dist_hw_base, irq);
+    shift = (irq % 16U) << 1U;
+
+    icfgr &= (~(3U << shift));
+    icfgr |= (config << shift);
+
+    GIC_DIST_CONFIG(_gic_table[index].dist_hw_base, irq) = icfgr;
+}
+
+rt_uint32_t arm_gic_get_configuration(rt_uint32_t index, int irq)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    return (GIC_DIST_CONFIG(_gic_table[index].dist_hw_base, irq) >> ((irq % 16U) >> 1U));
 }
 
 void arm_gic_clear_active(rt_uint32_t index, int irq)
 {
-    rt_uint32_t mask = 1 << (irq % 32);
+    rt_uint32_t mask = 1U << (irq % 32U);
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
     GIC_DIST_ACTIVE_CLEAR(_gic_table[index].dist_hw_base, irq) = mask;
 }
@@ -122,79 +228,149 @@ void arm_gic_set_cpu(rt_uint32_t index, int irq, unsigned int cpumask)
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
     old_tgt = GIC_DIST_TARGET(_gic_table[index].dist_hw_base, irq);
 
-    old_tgt &= ~(0x0FFUL << ((irq % 4)*8));
-    old_tgt |=   cpumask << ((irq % 4)*8);
+    old_tgt &= ~(0x0FFUL << ((irq % 4U)*8U));
+    old_tgt |= cpumask << ((irq % 4U)*8U);
 
     GIC_DIST_TARGET(_gic_table[index].dist_hw_base, irq) = old_tgt;
 }
 
-void arm_gic_umask(rt_uint32_t index, int irq)
+rt_uint32_t arm_gic_get_target_cpu(rt_uint32_t index, int irq)
 {
-    rt_uint32_t mask = 1 << (irq % 32);
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    return (GIC_DIST_TARGET(_gic_table[index].dist_hw_base, irq) >> ((irq % 4U) * 8U)) & 0xFFUL;
+}
+
+void arm_gic_set_priority(rt_uint32_t index, int irq, rt_uint32_t priority)
+{
+    rt_uint32_t mask;
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
     irq = irq - _gic_table[index].offset;
-    RT_ASSERT(irq >= 0);
+    RT_ASSERT(irq >= 0U);
 
-    GIC_DIST_ENABLE_SET(_gic_table[index].dist_hw_base, irq) = mask;
+    mask = GIC_DIST_PRI(_gic_table[index].dist_hw_base, irq);
+    mask &= ~(0xFFUL << ((irq % 4U) * 8U));
+    mask |= ((priority & 0xFFUL) << ((irq % 4U) * 8U));
+    GIC_DIST_PRI(_gic_table[index].dist_hw_base, irq) = mask;
 }
 
-void arm_gic_dump_type(rt_uint32_t index)
+rt_uint32_t arm_gic_get_priority(rt_uint32_t index, int irq)
 {
-    unsigned int gic_type;
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
 
-    gic_type = GIC_DIST_TYPE(_gic_table[index].dist_hw_base);
-    rt_kprintf("GICv%d on %p, max IRQs: %d, %s security extension(%08x)\n",
-               (GIC_DIST_ICPIDR2(_gic_table[index].dist_hw_base) >> 4) & 0xf,
-               _gic_table[index].dist_hw_base,
-               _gic_max_irq,
-               gic_type & (1 << 10) ? "has" : "no",
-               gic_type);
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    return (GIC_DIST_PRI(_gic_table[index].dist_hw_base, irq) >> ((irq % 4U) * 8U)) & 0xFFUL;
 }
 
-void arm_gic_dump(rt_uint32_t index)
+void arm_gic_set_interface_prior_mask(rt_uint32_t index, rt_uint32_t priority)
 {
-    unsigned int i, k;
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
 
-    k = GIC_CPU_HIGHPRI(_gic_table[index].cpu_hw_base);
-    rt_kprintf("--- high pending priority: %d(%08x)\n", k, k);
-    rt_kprintf("--- hw mask ---\n");
-    for (i = 0; i < _gic_max_irq / 32; i++)
-    {
-        rt_kprintf("0x%08x, ",
-                   GIC_DIST_ENABLE_SET(_gic_table[index].dist_hw_base,
-                                       i * 32));
-    }
-    rt_kprintf("\n--- hw pending ---\n");
-    for (i = 0; i < _gic_max_irq / 32; i++)
-    {
-        rt_kprintf("0x%08x, ",
-                   GIC_DIST_PENDING_SET(_gic_table[index].dist_hw_base,
-                                        i * 32));
-    }
-    rt_kprintf("\n--- hw active ---\n");
-    for (i = 0; i < _gic_max_irq / 32; i++)
-    {
-        rt_kprintf("0x%08x, ",
-                   GIC_DIST_ACTIVE_SET(_gic_table[index].dist_hw_base,
-                                       i * 32));
-    }
-    rt_kprintf("\n");
+    /* set priority mask */
+    GIC_CPU_PRIMASK(_gic_table[index].cpu_hw_base) = priority & 0xFFUL;
+}
+
+rt_uint32_t arm_gic_get_interface_prior_mask(rt_uint32_t index)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    return GIC_CPU_PRIMASK(_gic_table[index].cpu_hw_base);
+}
+
+void arm_gic_set_binary_point(rt_uint32_t index, rt_uint32_t binary_point)
+{
+    GIC_CPU_BINPOINT(_gic_table[index].cpu_hw_base) = binary_point & 0x7U;
+}
+
+rt_uint32_t arm_gic_get_binary_point(rt_uint32_t index)
+{
+    return GIC_CPU_BINPOINT(_gic_table[index].cpu_hw_base);
+}
+
+rt_uint32_t arm_gic_get_irq_status(rt_uint32_t index, int irq)
+{
+    rt_uint32_t pending;
+    rt_uint32_t active;
+
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    active = (GIC_DIST_ACTIVE_SET(_gic_table[index].dist_hw_base, irq) >> (irq % 32U)) & 0x1UL;
+    pending = (GIC_DIST_PENDING_SET(_gic_table[index].dist_hw_base, irq) >> (irq % 32U)) & 0x1UL;
+
+    return ((active << 1U) | pending);
+}
+
+void arm_gic_send_sgi(rt_uint32_t index, int irq, rt_uint32_t target_list, rt_uint32_t filter_list)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    GIC_DIST_SOFTINT(_gic_table[index].dist_hw_base) = ((filter_list & 0x3U) << 24U) | ((target_list & 0xFFUL) << 16U) | (irq & 0x0FUL);
+}
+
+rt_uint32_t arm_gic_get_high_pending_irq(rt_uint32_t index)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    return GIC_CPU_HIGHPRI(_gic_table[index].cpu_hw_base);
+}
+
+rt_uint32_t arm_gic_get_interface_id(rt_uint32_t index)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    return GIC_CPU_IIDR(_gic_table[index].cpu_hw_base);
+}
+
+void arm_gic_set_group(rt_uint32_t index, int irq, rt_uint32_t group)
+{
+    uint32_t igroupr;
+    uint32_t shift;
+
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+    RT_ASSERT(group <= 1U);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    igroupr = GIC_DIST_IGROUP(_gic_table[index].dist_hw_base, irq);
+    shift = (irq % 32U);
+    igroupr &= (~(1U << shift));
+    igroupr |= ( (group & 0x1U) << shift);
+
+    GIC_DIST_IGROUP(_gic_table[index].dist_hw_base, irq) = igroupr;
+}
+
+rt_uint32_t arm_gic_get_group(rt_uint32_t index, int irq)
+{
+    RT_ASSERT(index < ARM_GIC_MAX_NR);
+
+    irq = irq - _gic_table[index].offset;
+    RT_ASSERT(irq >= 0U);
+
+    return (GIC_DIST_IGROUP(_gic_table[index].dist_hw_base, irq) >> (irq % 32U)) & 0x1UL;
 }
-#ifdef RT_USING_FINSH
-#include <finsh.h>
-FINSH_FUNCTION_EXPORT_ALIAS(arm_gic_dump, gic, show gic status);
-#endif
 
 int arm_gic_dist_init(rt_uint32_t index, rt_uint32_t dist_base, int irq_start)
 {
     unsigned int gic_type, i;
-    rt_uint32_t cpumask = 1 << 0;
+    rt_uint32_t cpumask = 1U << 0U;
 
     RT_ASSERT(index < ARM_GIC_MAX_NR);
 
@@ -203,50 +379,50 @@ int arm_gic_dist_init(rt_uint32_t index, rt_uint32_t dist_base, int irq_start)
 
     /* Find out how many interrupts are supported. */
     gic_type = GIC_DIST_TYPE(dist_base);
-    _gic_max_irq = ((gic_type & 0x1f) + 1) * 32;
+    _gic_max_irq = ((gic_type & 0x1fU) + 1U) * 32U;
 
     /*
      * The GIC only supports up to 1020 interrupt sources.
      * Limit this to either the architected maximum, or the
      * platform maximum.
      */
-    if (_gic_max_irq > 1020)
-        _gic_max_irq = 1020;
+    if (_gic_max_irq > 1020U)
+        _gic_max_irq = 1020U;
     if (_gic_max_irq > ARM_GIC_NR_IRQS) /* the platform maximum interrupts */
         _gic_max_irq = ARM_GIC_NR_IRQS;
 
-    cpumask |= cpumask << 8;
-    cpumask |= cpumask << 16;
-    cpumask |= cpumask << 24;
+    cpumask |= cpumask << 8U;
+    cpumask |= cpumask << 16U;
+    cpumask |= cpumask << 24U;
 
-    GIC_DIST_CTRL(dist_base) = 0x0;
+    GIC_DIST_CTRL(dist_base) = 0x0U;
 
     /* Set all global interrupts to be level triggered, active low. */
-    for (i = 32; i < _gic_max_irq; i += 16)
-        GIC_DIST_CONFIG(dist_base, i) = 0x0;
+    for (i = 32U; i < _gic_max_irq; i += 16U)
+        GIC_DIST_CONFIG(dist_base, i) = 0x0U;
 
     /* Set all global interrupts to this CPU only. */
-    for (i = 32; i < _gic_max_irq; i += 4)
+    for (i = 32U; i < _gic_max_irq; i += 4U)
         GIC_DIST_TARGET(dist_base, i) = cpumask;
 
     /* Set priority on all interrupts. */
-    for (i = 0; i < _gic_max_irq; i += 4)
-        GIC_DIST_PRI(dist_base, i) = 0xa0a0a0a0;
+    for (i = 0U; i < _gic_max_irq; i += 4U)
+        GIC_DIST_PRI(dist_base, i) = 0xa0a0a0a0U;
 
     /* Disable all interrupts. */
-    for (i = 0; i < _gic_max_irq; i += 32)
-        GIC_DIST_ENABLE_CLEAR(dist_base, i) = 0xffffffff;
+    for (i = 0U; i < _gic_max_irq; i += 32U)
+        GIC_DIST_ENABLE_CLEAR(dist_base, i) = 0xffffffffU;
 
 #if 0
     /* All interrupts defaults to IGROUP1(IRQ). */
     for (i = 0; i < _gic_max_irq; i += 32)
         GIC_DIST_IGROUP(dist_base, i) = 0xffffffff;
 #endif
-    for (i = 0; i < _gic_max_irq; i += 32)
-        GIC_DIST_IGROUP(dist_base, i) = 0;
+    for (i = 0U; i < _gic_max_irq; i += 32U)
+        GIC_DIST_IGROUP(dist_base, i) = 0U;
 
     /* Enable group0 and group1 interrupt forwarding. */
-    GIC_DIST_CTRL(dist_base) = 0x01;
+    GIC_DIST_CTRL(dist_base) = 0x01U;
 
     return 0;
 }
@@ -257,44 +433,63 @@ int arm_gic_cpu_init(rt_uint32_t index, rt_uint32_t cpu_base)
 
     _gic_table[index].cpu_hw_base = cpu_base;
 
-    GIC_CPU_PRIMASK(cpu_base) = 0xf0;
-    GIC_CPU_BINPOINT(cpu_base) = 0x7;
+    GIC_CPU_PRIMASK(cpu_base) = 0xf0U;
+    GIC_CPU_BINPOINT(cpu_base) = 0x7U;
     /* Enable CPU interrupt */
-    GIC_CPU_CTRL(cpu_base) = 0x01;
+    GIC_CPU_CTRL(cpu_base) = 0x01U;
 
     return 0;
 }
 
-void arm_gic_set_group(rt_uint32_t index, int vector, int group)
+void arm_gic_dump_type(rt_uint32_t index)
 {
-    /* As for GICv2, there are only group0 and group1. */
-    RT_ASSERT(group <= 1);
-    RT_ASSERT(vector < _gic_max_irq);
+    unsigned int gic_type;
 
-    if (group == 0)
-    {
-        GIC_DIST_IGROUP(_gic_table[index].dist_hw_base,
-                        vector) &= ~(1 << (vector % 32));
-    }
-    else if (group == 1)
-    {
-        GIC_DIST_IGROUP(_gic_table[index].dist_hw_base,
-                        vector) |=  (1 << (vector % 32));
-    }
+    gic_type = GIC_DIST_TYPE(_gic_table[index].dist_hw_base);
+    rt_kprintf("GICv%d on %p, max IRQs: %d, %s security extension(%08x)\n",
+               (GIC_DIST_ICPIDR2(_gic_table[index].dist_hw_base) >> 4U) & 0xfUL,
+               _gic_table[index].dist_hw_base,
+               _gic_max_irq,
+               gic_type & (1U << 10U) ? "has" : "no",
+               gic_type);
 }
 
-#ifdef RT_USING_SMP
-void rt_hw_ipi_send(int ipi_vector, unsigned int cpu_mask)
- {
-     /* note: ipi_vector maybe different with irq_vector */
-     GIC_DIST_SOFTINT(_gic_table[0].dist_hw_base) = (cpu_mask << 16) | ipi_vector;
-}
-#endif
-
-#ifdef RT_USING_SMP
-void rt_hw_ipi_handler_install(int ipi_vector, rt_isr_handler_t ipi_isr_handler)
+void arm_gic_dump(rt_uint32_t index)
 {
-    /* note: ipi_vector maybe different with irq_vector */
-    rt_hw_interrupt_install(ipi_vector, ipi_isr_handler, 0, "IPI_HANDLER");
+    unsigned int i, k;
+
+    k = GIC_CPU_HIGHPRI(_gic_table[index].cpu_hw_base);
+    rt_kprintf("--- high pending priority: %d(%08x)\n", k, k);
+    rt_kprintf("--- hw mask ---\n");
+    for (i = 0U; i < _gic_max_irq / 32U; i++)
+    {
+        rt_kprintf("0x%08x, ",
+                   GIC_DIST_ENABLE_SET(_gic_table[index].dist_hw_base,
+                                       i * 32U));
+    }
+    rt_kprintf("\n--- hw pending ---\n");
+    for (i = 0U; i < _gic_max_irq / 32U; i++)
+    {
+        rt_kprintf("0x%08x, ",
+                   GIC_DIST_PENDING_SET(_gic_table[index].dist_hw_base,
+                                        i * 32U));
+    }
+    rt_kprintf("\n--- hw active ---\n");
+    for (i = 0U; i < _gic_max_irq / 32U; i++)
+    {
+        rt_kprintf("0x%08x, ",
+                   GIC_DIST_ACTIVE_SET(_gic_table[index].dist_hw_base,
+                                       i * 32U));
+    }
+    rt_kprintf("\n");
 }
-#endif
+
+long gic_dump(void)
+{
+    arm_gic_dump_type(0);
+    arm_gic_dump(0);
+
+    return 0;
+}
+MSH_CMD_EXPORT(gic_dump, show gic status);
+
diff --git a/libcpu/arm/cortex-a/gic.h b/libcpu/arm/cortex-a/gic.h
index e48f9f854d..e74b760f45 100644
--- a/libcpu/arm/cortex-a/gic.h
+++ b/libcpu/arm/cortex-a/gic.h
@@ -14,19 +14,46 @@
 #include <rthw.h>
 #include <board.h>
 
-int arm_gic_dist_init(rt_uint32_t index, rt_uint32_t dist_base, int irq_start);
-int arm_gic_cpu_init(rt_uint32_t index, rt_uint32_t cpu_base);
-
-void arm_gic_mask(rt_uint32_t index, int irq);
-void arm_gic_umask(rt_uint32_t index, int irq);
-void arm_gic_set_cpu(rt_uint32_t index, int irq, unsigned int cpumask);
-void arm_gic_set_group(rt_uint32_t index, int vector, int group);
-
 int arm_gic_get_active_irq(rt_uint32_t index);
 void arm_gic_ack(rt_uint32_t index, int irq);
 
+void arm_gic_mask(rt_uint32_t index, int irq);
+void arm_gic_umask(rt_uint32_t index, int irq);
+
+rt_uint32_t arm_gic_get_pending_irq(rt_uint32_t index, int irq);
+void arm_gic_set_pending_irq(rt_uint32_t index, int irq);
+void arm_gic_clear_pending_irq(rt_uint32_t index, int irq);
+
+void arm_gic_set_configuration(rt_uint32_t index, int irq, uint32_t config);
+rt_uint32_t arm_gic_get_configuration(rt_uint32_t index, int irq);
+
 void arm_gic_clear_active(rt_uint32_t index, int irq);
-void arm_gic_clear_pending(rt_uint32_t index, int irq);
+
+void arm_gic_set_cpu(rt_uint32_t index, int irq, unsigned int cpumask);
+rt_uint32_t arm_gic_get_target_cpu(rt_uint32_t index, int irq);
+
+void arm_gic_set_priority(rt_uint32_t index, int irq, rt_uint32_t priority);
+rt_uint32_t arm_gic_get_priority(rt_uint32_t index, int irq);
+
+void arm_gic_set_interface_prior_mask(rt_uint32_t index, rt_uint32_t priority);
+rt_uint32_t arm_gic_get_interface_prior_mask(rt_uint32_t index);
+
+void arm_gic_set_binary_point(rt_uint32_t index, rt_uint32_t binary_point);
+rt_uint32_t arm_gic_get_binary_point(rt_uint32_t index);
+
+rt_uint32_t arm_gic_get_irq_status(rt_uint32_t index, int irq);
+
+void arm_gic_send_sgi(rt_uint32_t index, int irq, rt_uint32_t target_list, rt_uint32_t filter_list);
+
+rt_uint32_t arm_gic_get_high_pending_irq(rt_uint32_t index);
+
+rt_uint32_t arm_gic_get_interface_id(rt_uint32_t index);
+
+void arm_gic_set_group(rt_uint32_t index, int irq, rt_uint32_t group);
+rt_uint32_t arm_gic_get_group(rt_uint32_t index, int irq);
+
+int arm_gic_dist_init(rt_uint32_t index, rt_uint32_t dist_base, int irq_start);
+int arm_gic_cpu_init(rt_uint32_t index, rt_uint32_t cpu_base);
 
 void arm_gic_dump_type(rt_uint32_t index);
 void arm_gic_dump(rt_uint32_t index);
diff --git a/libcpu/arm/cortex-a/gtimer.c b/libcpu/arm/cortex-a/gtimer.c
new file mode 100644
index 0000000000..b193a1593d
--- /dev/null
+++ b/libcpu/arm/cortex-a/gtimer.c
@@ -0,0 +1,171 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-03-30     huijie.feng  first version
+ */
+
+#include "cp15.h"
+#include <rtdef.h>
+
+/** Set CNTFRQ
+ *  This function assigns the given value to PL1 Physical Timer Counter Frequency Register (CNTFRQ).
+ *  @param value: CNTFRQ Register value to set
+ */
+static inline void __set_cntfrq(rt_uint32_t value)
+{
+    __set_cp(15, 0, value, 14, 0, 0);
+}
+
+/** Get CNTFRQ
+ *  This function returns the value of the PL1 Physical Timer Counter Frequency Register (CNTFRQ).
+ *  return CNTFRQ Register value
+ */
+static inline rt_uint32_t __get_cntfrq(void)
+{
+    rt_uint32_t result;
+    __get_cp(15, 0, result, 14, 0 , 0);
+    return result;
+}
+
+/** Set CNTP_TVAL
+ *  This function assigns the given value to PL1 Physical Timer Value Register (CNTP_TVAL).
+ *  param value: CNTP_TVAL Register value to set
+ */
+static inline void __set_cntp_tval(rt_uint32_t value)
+{
+    __set_cp(15, 0, value, 14, 2, 0);
+}
+
+/** Get CNTP_TVAL
+ *  This function returns the value of the PL1 Physical Timer Value Register (CNTP_TVAL).
+ *  return CNTP_TVAL Register value
+ */
+static inline rt_uint32_t __get_cntp_tval(void)
+{
+    rt_uint32_t result;
+    __get_cp(15, 0, result, 14, 2, 0);
+    return result;
+}
+
+/** Get CNTPCT
+ *  This function returns the value of the 64 bits PL1 Physical Count Register (CNTPCT).
+ *  return CNTPCT Register value
+ */
+static inline rt_uint64_t __get_cntpct(void)
+{
+    rt_uint64_t result;
+    __get_cp64(15, 0, result, 14);
+    return result;
+}
+
+/** Set CNTP_CVAL
+ *  This function assigns the given value to 64bits PL1 Physical Timer CompareValue Register (CNTP_CVAL).
+ *  param value: CNTP_CVAL Register value to set
+*/
+static inline void __set_cntp_cval(rt_uint64_t value)
+{
+    __set_cp64(15, 2, value, 14);
+}
+
+/** Get CNTP_CVAL
+ *  This function returns the value of the 64 bits PL1 Physical Timer CompareValue Register (CNTP_CVAL).
+ *  return CNTP_CVAL Register value
+ */
+static inline rt_uint64_t __get_cntp_cval(void)
+{
+    rt_uint64_t result;
+    __get_cp64(15, 2, result, 14);
+    return result;
+}
+
+/** Set CNTP_CTL
+ *  This function assigns the given value to PL1 Physical Timer Control Register (CNTP_CTL).
+ *  param value: CNTP_CTL Register value to set
+ */
+static inline void __set_cntp_ctl(uint32_t value)
+{
+    __set_cp(15, 0, value, 14, 2, 1);
+}
+
+/** Get CNTP_CTL register
+ *  return CNTP_CTL Register value
+ */
+static inline rt_uint32_t __get_cntp_ctl(void)
+{
+    rt_uint32_t result;
+    __get_cp(15, 0, result, 14, 2, 1);
+    return result;
+}
+
+/** Configures the frequency the timer shall run at.
+ *  param value The timer frequency in Hz.
+ */
+void gtimer_set_counter_frequency(rt_uint32_t value)
+{
+    __set_cntfrq(value);
+    __asm__ volatile ("isb 0xF":::"memory");
+}
+
+/** Sets the reset value of the timer.
+ *  param value: The value the timer is loaded with.
+ */
+void gtimer_set_load_value(rt_uint32_t value)
+{
+    __set_cntp_tval(value);
+    __asm__ volatile ("isb 0xF":::"memory");
+}
+
+/** Get the current counter value.
+ *  return Current counter value.
+ */
+rt_uint32_t gtimer_get_current_value(void)
+{
+    return(__get_cntp_tval());
+}
+
+/** Get the current physical counter value.
+ *  return Current physical counter value.
+ */
+rt_uint64_t gtimer_get_current_physical_value(void)
+{
+    return(__get_cntpct());
+}
+
+/** Set the physical compare value.
+ *  param value: New physical timer compare value.
+ */
+void gtimer_set_physical_compare_value(rt_uint64_t value)
+{
+    __set_cntp_cval(value);
+    __asm__ volatile ("isb 0xF":::"memory");
+}
+
+/** Get the physical compare value.
+ *  return Physical compare value.
+ */
+rt_uint64_t gtimer_get_physical_compare_value(void)
+{
+    return(__get_cntp_cval());
+}
+
+/** Configure the timer by setting the control value.
+ *  param value: New timer control value.
+ */
+void gtimer_set_control(rt_uint32_t value)
+{
+    __set_cntp_ctl(value);
+    __asm__ volatile ("isb 0xF":::"memory");
+}
+
+/** Get the control value.
+ *  return Control value.
+ */
+rt_uint32_t gtimer_get_control(void)
+{
+    return(__get_cntp_ctl());
+}
+
diff --git a/libcpu/arm/cortex-a/gtimer.h b/libcpu/arm/cortex-a/gtimer.h
new file mode 100644
index 0000000000..160d0cefde
--- /dev/null
+++ b/libcpu/arm/cortex-a/gtimer.h
@@ -0,0 +1,26 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-03-30     huijie.feng  first version
+ */
+
+#ifndef __GTIMER_H__
+#define __GTIMER_H__
+
+#include <rtdef.h>
+
+void gtimer_set_counter_frequency(rt_uint32_t value);
+void gtimer_set_load_value(rt_uint32_t value);
+rt_uint32_t gtimer_get_current_value(void);
+rt_uint64_t gtimer_get_current_physical_value(void);
+void gtimer_set_physical_compare_value(rt_uint64_t value);
+rt_uint64_t gtimer_get_physical_compare_value(void);
+void gtimer_set_control(rt_uint32_t value);
+rt_uint32_t gtimer_get_control(void);
+
+#endif
+
diff --git a/libcpu/arm/cortex-a/interrupt.c b/libcpu/arm/cortex-a/interrupt.c
index 0d440d1d0e..65e51fa3ae 100644
--- a/libcpu/arm/cortex-a/interrupt.c
+++ b/libcpu/arm/cortex-a/interrupt.c
@@ -83,7 +83,7 @@ void rt_hw_interrupt_umask(int vector)
  */
 int rt_hw_interrupt_get_irq(void)
 {
-    return arm_gic_get_active_irq(0) & GIC_ACK_INTID_MASK;
+    return arm_gic_get_active_irq(0);
 }
 
 /**
@@ -94,6 +94,150 @@ void rt_hw_interrupt_ack(int vector)
 {
     arm_gic_ack(0, vector);
 }
+
+/**
+ * This function set interrupt CPU targets.
+ * @param vector:   the interrupt number
+ *        cpu_mask: target cpus mask, one bit for one core
+ */
+void rt_hw_interrupt_set_target_cpus(int vector, unsigned int cpu_mask)
+{
+    arm_gic_set_cpu(0, vector, cpu_mask);
+}
+
+/**
+ * This function get interrupt CPU targets.
+ * @param vector: the interrupt number
+ * @return target cpus mask, one bit for one core
+ */
+unsigned int rt_hw_interrupt_get_target_cpus(int vector)
+{
+    return arm_gic_get_target_cpu(0, vector);
+}
+
+/**
+ * This function set interrupt triger mode.
+ * @param vector: the interrupt number
+ *        mode:   interrupt triger mode; 0: level triger, 1: edge triger
+ */
+void rt_hw_interrupt_set_triger_mode(int vector, unsigned int mode)
+{
+    arm_gic_set_configuration(0, vector, mode);
+}
+
+/**
+ * This function get interrupt triger mode.
+ * @param vector: the interrupt number
+ * @return interrupt triger mode; 0: level triger, 1: edge triger
+ */
+unsigned int rt_hw_interrupt_get_triger_mode(int vector)
+{
+    return arm_gic_get_configuration(0, vector);
+}
+
+/**
+ * This function set interrupt pending flag.
+ * @param vector: the interrupt number
+ */
+void rt_hw_interrupt_set_pending(int vector)
+{
+    arm_gic_set_pending_irq(0, vector);
+}
+
+/**
+ * This function get interrupt pending flag.
+ * @param vector: the interrupt number
+ * @return interrupt pending flag, 0: not pending; 1: pending
+ */
+unsigned int rt_hw_interrupt_get_pending(int vector)
+{
+    return arm_gic_get_pending_irq(0, vector);
+}
+
+/**
+ * This function clear interrupt pending flag.
+ * @param vector: the interrupt number
+ */
+void rt_hw_interrupt_clear_pending(int vector)
+{
+    arm_gic_clear_pending_irq(0, vector);
+}
+
+/**
+ * This function set interrupt priority value.
+ * @param vector: the interrupt number
+ *        priority: the priority of interrupt to set
+ */
+void rt_hw_interrupt_set_priority(int vector, unsigned int priority)
+{
+    arm_gic_set_priority(0, vector, priority);
+}
+
+/**
+ * This function get interrupt priority.
+ * @param vector: the interrupt number
+ * @return interrupt priority value
+ */
+unsigned int rt_hw_interrupt_get_priority(int vector)
+{
+    return arm_gic_get_priority(0, vector);
+}
+
+/**
+ * This function set priority masking threshold.
+ * @param priority: priority masking threshold
+ */
+void rt_hw_interrupt_set_priority_mask(unsigned int priority)
+{
+    arm_gic_set_interface_prior_mask(0, priority);
+}
+
+/**
+ * This function get priority masking threshold.
+ * @param none
+ * @return priority masking threshold
+ */
+unsigned int rt_hw_interrupt_get_priority_mask(void)
+{
+    return arm_gic_get_interface_prior_mask(0);
+}
+
+/**
+ * This function set priority grouping field split point.
+ * @param bits: priority grouping field split point
+ * @return 0: success; -1: failed
+ */
+int rt_hw_interrupt_set_prior_group_bits(unsigned int bits)
+{
+    int status;
+
+    if (bits < 8)
+    {
+        arm_gic_set_binary_point(0, (7 - bits));
+        status = 0;
+    }
+    else
+    {
+        status = -1;
+    }
+
+    return (status);
+}
+
+/**
+ * This function get priority grouping field split point.
+ * @param none
+ * @return priority grouping field split point
+ */
+unsigned int rt_hw_interrupt_get_prior_group_bits(void)
+{
+    unsigned int bp;
+
+    bp = arm_gic_get_binary_point(0) & 0x07;
+
+    return (7 - bp);
+}
+
 /**
  * This function will install a interrupt service routine to a interrupt.
  * @param vector the interrupt number
@@ -121,3 +265,17 @@ rt_isr_handler_t rt_hw_interrupt_install(int vector, rt_isr_handler_t handler,
 
     return old_handler;
 }
+
+#ifdef RT_USING_SMP
+void rt_hw_ipi_send(int ipi_vector, unsigned int cpu_mask)
+{
+    arm_gic_send_sgi(0, ipi_vector, cpu_mask, 0);
+}
+
+void rt_hw_ipi_handler_install(int ipi_vector, rt_isr_handler_t ipi_isr_handler)
+{
+    /* note: ipi_vector maybe different with irq_vector */
+    rt_hw_interrupt_install(ipi_vector, ipi_isr_handler, 0, "IPI_HANDLER");
+}
+#endif
+
diff --git a/libcpu/arm/cortex-a/interrupt.h b/libcpu/arm/cortex-a/interrupt.h
index a8a9afb95c..dda6c5534a 100644
--- a/libcpu/arm/cortex-a/interrupt.h
+++ b/libcpu/arm/cortex-a/interrupt.h
@@ -17,9 +17,10 @@
 #define INT_IRQ     0x00
 #define INT_FIQ     0x01
 
-void rt_hw_vector_init(void);
+#define IRQ_MODE_TRIG_LEVEL         (0x00) /* Trigger: level triggered interrupt */
+#define IRQ_MODE_TRIG_EDGE          (0x01) /* Trigger: edge triggered interrupt */
 
-void rt_hw_interrupt_control(int vector, int priority, int route);
+void rt_hw_vector_init(void);
 
 void rt_hw_interrupt_init(void);
 void rt_hw_interrupt_mask(int vector);
@@ -28,7 +29,32 @@ void rt_hw_interrupt_umask(int vector);
 int rt_hw_interrupt_get_irq(void);
 void rt_hw_interrupt_ack(int vector);
 
+void rt_hw_interrupt_set_target_cpus(int vector, unsigned int cpu_mask);
+unsigned int rt_hw_interrupt_get_target_cpus(int vector);
+
+void rt_hw_interrupt_set_triger_mode(int vector, unsigned int mode);
+unsigned int rt_hw_interrupt_get_triger_mode(int vector);
+
+void rt_hw_interrupt_set_pending(int vector);
+unsigned int rt_hw_interrupt_get_pending(int vector);
+void rt_hw_interrupt_clear_pending(int vector);
+
+void rt_hw_interrupt_set_priority(int vector, unsigned int priority);
+unsigned int rt_hw_interrupt_get_priority(int vector);
+
+void rt_hw_interrupt_set_priority_mask(unsigned int priority);
+unsigned int rt_hw_interrupt_get_priority_mask(void);
+
+int rt_hw_interrupt_set_prior_group_bits(unsigned int bits);
+unsigned int rt_hw_interrupt_get_prior_group_bits(void);
+
 rt_isr_handler_t rt_hw_interrupt_install(int vector, rt_isr_handler_t handler,
         void *param, const char *name);
 
+#ifdef RT_USING_SMP
+void rt_hw_ipi_send(int ipi_vector, unsigned int cpu_mask);
+void rt_hw_ipi_handler_install(int ipi_vector, rt_isr_handler_t ipi_isr_handler);
 #endif
+
+#endif
+