diff --git a/.github/workflows/compilation_on_android_ubuntu.yml b/.github/workflows/compilation_on_android_ubuntu.yml
index e0c9326b..c1c31f5f 100644
--- a/.github/workflows/compilation_on_android_ubuntu.yml
+++ b/.github/workflows/compilation_on_android_ubuntu.yml
@@ -566,9 +566,7 @@ jobs:
             test_option: $GC_TEST_OPTIONS
           - running_mode: "multi-tier-jit"
             test_option: $GC_TEST_OPTIONS
-          # aot, fast-interp, fast-jit, llvm-jit, multi-tier-jit don't support Memory64
-          - running_mode: "aot"
-            test_option: $MEMORY64_TEST_OPTIONS
+          # fast-interp, fast-jit, llvm-jit, multi-tier-jit don't support Memory64
           - running_mode: "fast-interp"
             test_option: $MEMORY64_TEST_OPTIONS
           - running_mode: "fast-jit"
@@ -616,6 +614,7 @@ jobs:
         if: >
           ((matrix.test_option == '$DEFAULT_TEST_OPTIONS' || matrix.test_option == '$THREADS_TEST_OPTIONS'
             || matrix.test_option == '$WASI_TEST_OPTIONS' || matrix.test_option == '$GC_TEST_OPTIONS')
+           && matrix.test_option != '$MEMORY64_TEST_OPTIONS'
            && matrix.running_mode != 'fast-jit' && matrix.running_mode != 'jit' && matrix.running_mode != 'multi-tier-jit')
         run: echo "TEST_ON_X86_32=true" >> $GITHUB_ENV
 
diff --git a/core/iwasm/aot/aot_loader.c b/core/iwasm/aot/aot_loader.c
index b67f9c68..48ed4e84 100644
--- a/core/iwasm/aot/aot_loader.c
+++ b/core/iwasm/aot/aot_loader.c
@@ -9,6 +9,7 @@
 #include "aot_reloc.h"
 #include "../common/wasm_runtime_common.h"
 #include "../common/wasm_native.h"
+#include "../common/wasm_loader_common.h"
 #include "../compilation/aot.h"
 
 #if WASM_ENABLE_DEBUG_AOT != 0
@@ -1043,6 +1044,12 @@ load_memory_info(const uint8 **p_buf, const uint8 *buf_end, AOTModule *module,
 
     for (i = 0; i < module->memory_count; i++) {
         read_uint32(buf, buf_end, module->memories[i].memory_flags);
+
+        if (!wasm_memory_check_flags(module->memories[i].memory_flags,
+                                     error_buf, error_buf_size, true)) {
+            return false;
+        }
+
         read_uint32(buf, buf_end, module->memories[i].num_bytes_per_page);
         read_uint32(buf, buf_end, module->memories[i].mem_init_page_count);
         read_uint32(buf, buf_end, module->memories[i].mem_max_page_count);
@@ -3634,6 +3641,21 @@ fail:
     return ret;
 }
 
+#if WASM_ENABLE_MEMORY64 != 0
+static bool
+has_module_memory64(AOTModule *module)
+{
+    /* TODO: multi-memories for now assuming the memory idx type is consistent
+     * across multi-memories */
+    if (module->import_memory_count > 0)
+        return !!(module->import_memories[0].memory_flags & MEMORY64_FLAG);
+    else if (module->memory_count > 0)
+        return !!(module->memories[0].memory_flags & MEMORY64_FLAG);
+
+    return false;
+}
+#endif
+
 static bool
 load_from_sections(AOTModule *module, AOTSection *sections,
                    bool is_load_from_file_buf, char *error_buf,
@@ -3645,6 +3667,7 @@ load_from_sections(AOTModule *module, AOTSection *sections,
     uint32 i, func_index, func_type_index;
     AOTFuncType *func_type;
     AOTExport *exports;
+    uint8 malloc_free_io_type = VALUE_TYPE_I32;
 
     while (section) {
         buf = section->section_body;
@@ -3719,7 +3742,10 @@ load_from_sections(AOTModule *module, AOTSection *sections,
     module->malloc_func_index = (uint32)-1;
     module->free_func_index = (uint32)-1;
     module->retain_func_index = (uint32)-1;
-
+#if WASM_ENABLE_MEMORY64 != 0
+    if (has_module_memory64(module))
+        malloc_free_io_type = VALUE_TYPE_I64;
+#endif
     exports = module->exports;
     for (i = 0; i < module->export_count; i++) {
         if (exports[i].kind == EXPORT_KIND_FUNC
@@ -3729,8 +3755,8 @@ load_from_sections(AOTModule *module, AOTSection *sections,
                 func_type_index = module->func_type_indexes[func_index];
                 func_type = (AOTFuncType *)module->types[func_type_index];
                 if (func_type->param_count == 1 && func_type->result_count == 1
-                    && func_type->types[0] == VALUE_TYPE_I32
-                    && func_type->types[1] == VALUE_TYPE_I32) {
+                    && func_type->types[0] == malloc_free_io_type
+                    && func_type->types[1] == malloc_free_io_type) {
                     bh_assert(module->malloc_func_index == (uint32)-1);
                     module->malloc_func_index = func_index;
                     LOG_VERBOSE("Found malloc function, name: %s, index: %u",
@@ -3742,9 +3768,9 @@ load_from_sections(AOTModule *module, AOTSection *sections,
                 func_type_index = module->func_type_indexes[func_index];
                 func_type = (AOTFuncType *)module->types[func_type_index];
                 if (func_type->param_count == 2 && func_type->result_count == 1
-                    && func_type->types[0] == VALUE_TYPE_I32
+                    && func_type->types[0] == malloc_free_io_type
                     && func_type->types[1] == VALUE_TYPE_I32
-                    && func_type->types[2] == VALUE_TYPE_I32) {
+                    && func_type->types[2] == malloc_free_io_type) {
                     uint32 j;
                     WASMExport *export_tmp;
 
@@ -3768,8 +3794,8 @@ load_from_sections(AOTModule *module, AOTSection *sections,
                                 (AOTFuncType *)module->types[func_type_index];
                             if (func_type->param_count == 1
                                 && func_type->result_count == 1
-                                && func_type->types[0] == VALUE_TYPE_I32
-                                && func_type->types[1] == VALUE_TYPE_I32) {
+                                && func_type->types[0] == malloc_free_io_type
+                                && func_type->types[1] == malloc_free_io_type) {
                                 bh_assert(module->retain_func_index
                                           == (uint32)-1);
                                 module->retain_func_index = export_tmp->index;
@@ -3795,7 +3821,7 @@ load_from_sections(AOTModule *module, AOTSection *sections,
                 func_type_index = module->func_type_indexes[func_index];
                 func_type = (AOTFuncType *)module->types[func_type_index];
                 if (func_type->param_count == 1 && func_type->result_count == 0
-                    && func_type->types[0] == VALUE_TYPE_I32) {
+                    && func_type->types[0] == malloc_free_io_type) {
                     bh_assert(module->free_func_index == (uint32)-1);
                     module->free_func_index = func_index;
                     LOG_VERBOSE("Found free function, name: %s, index: %u",
diff --git a/core/iwasm/aot/aot_runtime.c b/core/iwasm/aot/aot_runtime.c
index 1668ed01..610213cb 100644
--- a/core/iwasm/aot/aot_runtime.c
+++ b/core/iwasm/aot/aot_runtime.c
@@ -792,16 +792,18 @@ memory_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
     uint32 max_page_count =
         wasm_runtime_get_max_mem(max_memory_pages, memory->mem_init_page_count,
                                  memory->mem_max_page_count);
+    uint32 default_max_pages;
     uint32 inc_page_count, global_idx;
     uint32 bytes_of_last_page, bytes_to_page_end;
     uint64 aux_heap_base,
         heap_offset = (uint64)num_bytes_per_page * init_page_count;
     uint64 memory_data_size, max_memory_data_size;
     uint8 *p = NULL, *global_addr;
+    bool is_memory64 = memory->memory_flags & MEMORY64_FLAG;
 
     bool is_shared_memory = false;
 #if WASM_ENABLE_SHARED_MEMORY != 0
-    is_shared_memory = memory->memory_flags & 0x02 ? true : false;
+    is_shared_memory = memory->memory_flags & SHARED_MEMORY_FLAG ? true : false;
     /* Shared memory */
     if (is_shared_memory && parent != NULL) {
         AOTMemoryInstance *shared_memory_instance;
@@ -813,6 +815,16 @@ memory_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
     }
 #endif
 
+#if WASM_ENABLE_MEMORY64 != 0
+    if (is_memory64) {
+        default_max_pages = DEFAULT_MEM64_MAX_PAGES;
+    }
+    else
+#endif
+    {
+        default_max_pages = DEFAULT_MAX_PAGES;
+    }
+
     if (heap_size > 0 && module->malloc_func_index != (uint32)-1
         && module->free_func_index != (uint32)-1) {
         /* Disable app heap, use malloc/free function exported
@@ -893,14 +905,14 @@ memory_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
         }
         init_page_count += inc_page_count;
         max_page_count += inc_page_count;
-        if (init_page_count > DEFAULT_MAX_PAGES) {
+        if (init_page_count > default_max_pages) {
             set_error_buf(error_buf, error_buf_size,
                           "failed to insert app heap into linear memory, "
                           "try using `--heap-size=0` option");
             return NULL;
         }
-        if (max_page_count > DEFAULT_MAX_PAGES)
-            max_page_count = DEFAULT_MAX_PAGES;
+        if (max_page_count > default_max_pages)
+            max_page_count = default_max_pages;
     }
 
     LOG_VERBOSE("Memory instantiate:");
@@ -912,11 +924,11 @@ memory_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
                 heap_size);
 
     max_memory_data_size = (uint64)num_bytes_per_page * max_page_count;
-    bh_assert(max_memory_data_size <= MAX_LINEAR_MEMORY_SIZE);
+    bh_assert(max_memory_data_size <= GET_MAX_LINEAR_MEMORY_SIZE(is_memory64));
     (void)max_memory_data_size;
 
     /* TODO: memory64 uses is_memory64 flag */
-    if (wasm_allocate_linear_memory(&p, is_shared_memory, false,
+    if (wasm_allocate_linear_memory(&p, is_shared_memory, is_memory64,
                                     num_bytes_per_page, init_page_count,
                                     max_page_count, &memory_data_size)
         != BHT_OK) {
@@ -930,6 +942,11 @@ memory_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
     memory_inst->cur_page_count = init_page_count;
     memory_inst->max_page_count = max_page_count;
     memory_inst->memory_data_size = memory_data_size;
+#if WASM_ENABLE_MEMORY64 != 0
+    if (is_memory64) {
+        memory_inst->is_memory64 = 1;
+    }
+#endif
 
     /* Init memory info */
     memory_inst->memory_data = p;
@@ -993,11 +1010,12 @@ memories_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
                      uint32 max_memory_pages, char *error_buf,
                      uint32 error_buf_size)
 {
-    uint32 global_index, global_data_offset, base_offset, length;
+    uint32 global_index, global_data_offset, length;
     uint32 i, memory_count = module->memory_count;
     AOTMemoryInstance *memories, *memory_inst;
     AOTMemInitData *data_seg;
     uint64 total_size;
+    mem_offset_t base_offset;
 
     module_inst->memory_count = memory_count;
     total_size = sizeof(AOTMemoryInstance *) * (uint64)memory_count;
@@ -1036,7 +1054,9 @@ memories_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
                initialized */
             continue;
 
-        bh_assert(data_seg->offset.init_expr_type == INIT_EXPR_TYPE_I32_CONST
+        bh_assert(data_seg->offset.init_expr_type
+                      == (memory_inst->is_memory64 ? INIT_EXPR_TYPE_I64_CONST
+                                                   : INIT_EXPR_TYPE_I32_CONST)
                   || data_seg->offset.init_expr_type
                          == INIT_EXPR_TYPE_GET_GLOBAL);
 
@@ -1057,11 +1077,28 @@ memories_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
                     module->globals[global_index - module->import_global_count]
                         .data_offset;
 
-            base_offset =
-                *(uint32 *)(module_inst->global_data + global_data_offset);
+#if WASM_ENABLE_MEMORY64 != 0
+            if (memory_inst->is_memory64) {
+                base_offset =
+                    *(uint64 *)(module_inst->global_data + global_data_offset);
+            }
+            else
+#endif
+            {
+                base_offset =
+                    *(uint32 *)(module_inst->global_data + global_data_offset);
+            }
         }
         else {
-            base_offset = (uint32)data_seg->offset.u.i32;
+#if WASM_ENABLE_MEMORY64 != 0
+            if (memory_inst->is_memory64) {
+                base_offset = data_seg->offset.u.i64;
+            }
+            else
+#endif
+            {
+                base_offset = data_seg->offset.u.u32;
+            }
         }
 
         /* Copy memory data */
@@ -1071,7 +1108,8 @@ memories_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
         /* Check memory data */
         /* check offset since length might negative */
         if (base_offset > memory_inst->memory_data_size) {
-            LOG_DEBUG("base_offset(%d) > memory_data_size(%" PRIu64 ")",
+            LOG_DEBUG("base_offset(%" PR_MEM_OFFSET
+                      ") > memory_data_size(%" PRIu64 ")",
                       base_offset, memory_inst->memory_data_size);
 #if WASM_ENABLE_REF_TYPES != 0
             set_error_buf(error_buf, error_buf_size,
@@ -1086,8 +1124,8 @@ memories_instantiate(AOTModuleInstance *module_inst, AOTModuleInstance *parent,
         /* check offset + length(could be zero) */
         length = data_seg->byte_count;
         if (base_offset + length > memory_inst->memory_data_size) {
-            LOG_DEBUG("base_offset(%d) + length(%d) > memory_data_size(%" PRIu64
-                      ")",
+            LOG_DEBUG("base_offset(%" PR_MEM_OFFSET
+                      ") + length(%d) > memory_data_size(%" PRIu64 ")",
                       base_offset, length, memory_inst->memory_data_size);
 #if WASM_ENABLE_REF_TYPES != 0
             set_error_buf(error_buf, error_buf_size,
@@ -2334,22 +2372,44 @@ aot_copy_exception(AOTModuleInstance *module_inst, char *exception_buf)
 static bool
 execute_malloc_function(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
                         AOTFunctionInstance *malloc_func,
-                        AOTFunctionInstance *retain_func, uint32 size,
-                        uint32 *p_result)
+                        AOTFunctionInstance *retain_func, uint64 size,
+                        uint64 *p_result)
 {
 #ifdef OS_ENABLE_HW_BOUND_CHECK
     WASMExecEnv *exec_env_tls = wasm_runtime_get_exec_env_tls();
 #endif
     WASMExecEnv *exec_env_created = NULL;
     WASMModuleInstanceCommon *module_inst_old = NULL;
-    uint32 argv[2], argc;
+    union {
+        uint32 u32[3];
+        uint64 u64;
+    } argv;
+    uint32 argc;
     bool ret;
 
-    argv[0] = size;
-    argc = 1;
-    if (retain_func) {
-        argv[1] = 0;
+#if WASM_ENABLE_MEMORY64 != 0
+    bool is_memory64 = module_inst->memories[0]->is_memory64;
+    if (is_memory64) {
         argc = 2;
+        PUT_I64_TO_ADDR(&argv.u64, size);
+    }
+    else
+#endif
+    {
+        argc = 1;
+        argv.u32[0] = (uint32)size;
+    }
+
+    /* if __retain is exported, then this module is compiled by
+        assemblyscript, the memory should be managed by as's runtime,
+        in this case we need to call the retain function after malloc
+        the memory */
+    if (retain_func) {
+        /* the malloc function from assemblyscript is:
+            function __new(size: usize, id: u32)
+            id = 0 means this is an ArrayBuffer object */
+        argv.u32[argc] = 0;
+        argc++;
     }
 
     if (exec_env) {
@@ -2389,10 +2449,10 @@ execute_malloc_function(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
         }
     }
 
-    ret = aot_call_function(exec_env, malloc_func, argc, argv);
+    ret = aot_call_function(exec_env, malloc_func, argc, argv.u32);
 
     if (retain_func && ret)
-        ret = aot_call_function(exec_env, retain_func, 1, argv);
+        ret = aot_call_function(exec_env, retain_func, 1, argv.u32);
 
     if (module_inst_old)
         /* Restore the existing exec_env's module inst */
@@ -2401,24 +2461,46 @@ execute_malloc_function(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
     if (exec_env_created)
         wasm_exec_env_destroy(exec_env_created);
 
-    if (ret)
-        *p_result = argv[0];
+    if (ret) {
+#if WASM_ENABLE_MEMORY64 != 0
+        if (is_memory64)
+            *p_result = GET_I64_FROM_ADDR(&argv.u64);
+        else
+#endif
+        {
+            *p_result = argv.u32[0];
+        }
+    }
     return ret;
 }
 
 static bool
 execute_free_function(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
-                      AOTFunctionInstance *free_func, uint32 offset)
+                      AOTFunctionInstance *free_func, uint64 offset)
 {
 #ifdef OS_ENABLE_HW_BOUND_CHECK
     WASMExecEnv *exec_env_tls = wasm_runtime_get_exec_env_tls();
 #endif
     WASMExecEnv *exec_env_created = NULL;
     WASMModuleInstanceCommon *module_inst_old = NULL;
-    uint32 argv[2];
+    union {
+        uint32 u32[2];
+        uint64 u64;
+    } argv;
+    uint32 argc;
     bool ret;
 
-    argv[0] = offset;
+#if WASM_ENABLE_MEMORY64 != 0
+    if (module_inst->memories[0]->is_memory64) {
+        PUT_I64_TO_ADDR(&argv.u64, offset);
+        argc = 2;
+    }
+    else
+#endif
+    {
+        argv.u32[0] = (uint32)offset;
+        argc = 1;
+    }
 
     if (exec_env) {
 #ifdef OS_ENABLE_HW_BOUND_CHECK
@@ -2457,7 +2539,7 @@ execute_free_function(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
         }
     }
 
-    ret = aot_call_function(exec_env, free_func, 1, argv);
+    ret = aot_call_function(exec_env, free_func, argc, argv.u32);
 
     if (module_inst_old)
         /* Restore the existing exec_env's module inst */
@@ -2477,7 +2559,7 @@ aot_module_malloc_internal(AOTModuleInstance *module_inst,
     AOTMemoryInstance *memory_inst = aot_get_default_memory(module_inst);
     AOTModule *module = (AOTModule *)module_inst->module;
     uint8 *addr = NULL;
-    uint32 offset = 0;
+    uint64 offset = 0;
 
     /* TODO: Memory64 size check based on memory idx type */
     bh_assert(size <= UINT32_MAX);
@@ -2509,7 +2591,7 @@ aot_module_malloc_internal(AOTModuleInstance *module_inst,
 
         if (!malloc_func
             || !execute_malloc_function(module_inst, exec_env, malloc_func,
-                                        retain_func, (uint32)size, &offset)) {
+                                        retain_func, size, &offset)) {
             return 0;
         }
         addr = offset ? (uint8 *)memory_inst->memory_data + offset : NULL;
@@ -2620,8 +2702,7 @@ aot_module_free_internal(AOTModuleInstance *module_inst, WASMExecEnv *exec_env,
                 free_func = aot_lookup_function(module_inst, "__unpin");
 
             if (free_func)
-                execute_free_function(module_inst, exec_env, free_func,
-                                      (uint32)ptr);
+                execute_free_function(module_inst, exec_env, free_func, ptr);
         }
     }
 }
@@ -2983,7 +3064,7 @@ aot_sqrtf(float x)
 #if WASM_ENABLE_BULK_MEMORY != 0
 bool
 aot_memory_init(AOTModuleInstance *module_inst, uint32 seg_index, uint32 offset,
-                uint32 len, uint32 dst)
+                uint32 len, size_t dst)
 {
     AOTMemoryInstance *memory_inst = aot_get_default_memory(module_inst);
     AOTModule *aot_module;
@@ -3016,7 +3097,7 @@ aot_memory_init(AOTModuleInstance *module_inst, uint32 seg_index, uint32 offset,
         (WASMModuleInstanceCommon *)module_inst, (uint64)dst);
 
     SHARED_MEMORY_LOCK(memory_inst);
-    bh_memcpy_s(maddr, (uint32)(memory_inst->memory_data_size - dst),
+    bh_memcpy_s(maddr, CLAMP_U64_TO_U32(memory_inst->memory_data_size - dst),
                 data + offset, len);
     SHARED_MEMORY_UNLOCK(memory_inst);
     return true;
diff --git a/core/iwasm/aot/aot_runtime.h b/core/iwasm/aot/aot_runtime.h
index 0dbf9a5e..79bdb1df 100644
--- a/core/iwasm/aot/aot_runtime.h
+++ b/core/iwasm/aot/aot_runtime.h
@@ -627,7 +627,7 @@ aot_sqrtf(float x);
 #if WASM_ENABLE_BULK_MEMORY != 0
 bool
 aot_memory_init(AOTModuleInstance *module_inst, uint32 seg_index, uint32 offset,
-                uint32 len, uint32 dst);
+                uint32 len, size_t dst);
 
 bool
 aot_data_drop(AOTModuleInstance *module_inst, uint32 seg_index);
diff --git a/core/iwasm/common/wasm_loader_common.c b/core/iwasm/common/wasm_loader_common.c
new file mode 100644
index 00000000..ceb1cf67
--- /dev/null
+++ b/core/iwasm/common/wasm_loader_common.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright (C) 2024 Amazon Inc.  All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ */
+#include "wasm_loader_common.h"
+#include "bh_log.h"
+#include "../interpreter/wasm.h"
+
+static void
+set_error_buf(char *error_buf, uint32 error_buf_size, const char *string,
+              bool is_aot)
+{
+    if (error_buf != NULL) {
+        snprintf(error_buf, error_buf_size, "%s module load failed: %s",
+                 is_aot ? "AOT" : "WASM", string);
+    }
+}
+
+bool
+wasm_memory_check_flags(const uint8 mem_flag, char *error_buf,
+                        uint32 error_buf_size, bool is_aot)
+{
+    /* Check whether certain features indicated by mem_flag are enabled in
+     * runtime */
+    if (mem_flag > MAX_PAGE_COUNT_FLAG) {
+#if WASM_ENABLE_SHARED_MEMORY == 0
+        if (mem_flag & SHARED_MEMORY_FLAG) {
+            LOG_VERBOSE("shared memory flag was found, please enable shared "
+                        "memory, lib-pthread or lib-wasi-threads");
+            set_error_buf(error_buf, error_buf_size, "invalid limits flags",
+                          is_aot);
+            return false;
+        }
+#endif
+#if WASM_ENABLE_MEMORY64 == 0
+        if (mem_flag & MEMORY64_FLAG) {
+            LOG_VERBOSE("memory64 flag was found, please enable memory64");
+            set_error_buf(error_buf, error_buf_size, "invalid limits flags",
+                          is_aot);
+            return false;
+        }
+#endif
+    }
+
+    if (mem_flag > MAX_PAGE_COUNT_FLAG + SHARED_MEMORY_FLAG + MEMORY64_FLAG) {
+        set_error_buf(error_buf, error_buf_size, "invalid limits flags",
+                      is_aot);
+        return false;
+    }
+    else if ((mem_flag & SHARED_MEMORY_FLAG)
+             && !(mem_flag & MAX_PAGE_COUNT_FLAG)) {
+        set_error_buf(error_buf, error_buf_size,
+                      "shared memory must have maximum", is_aot);
+        return false;
+    }
+
+    return true;
+}
diff --git a/core/iwasm/common/wasm_loader_common.h b/core/iwasm/common/wasm_loader_common.h
new file mode 100644
index 00000000..05808176
--- /dev/null
+++ b/core/iwasm/common/wasm_loader_common.h
@@ -0,0 +1,23 @@
+/*
+ * Copyright (C) 2024 Amazon Inc.  All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ */
+
+#ifndef _WASM_LOADER_COMMON_H
+#define _WASM_LOADER_COMMON_H
+
+#include "platform_common.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+bool
+wasm_memory_check_flags(const uint8 mem_flag, char *error_buf,
+                        uint32 error_buf_size, bool is_aot);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* end of _WASM_LOADER_COMMON_H */
\ No newline at end of file
diff --git a/core/iwasm/common/wasm_memory.c b/core/iwasm/common/wasm_memory.c
index c49c1825..f47215f2 100644
--- a/core/iwasm/common/wasm_memory.c
+++ b/core/iwasm/common/wasm_memory.c
@@ -1005,15 +1005,7 @@ wasm_allocate_linear_memory(uint8 **data, bool is_shared_memory,
     page_size = os_getpagesize();
     *memory_data_size = init_page_count * num_bytes_per_page;
 
-#if WASM_ENABLE_MEMORY64 != 0
-    if (is_memory64) {
-        bh_assert(*memory_data_size <= MAX_LINEAR_MEM64_MEMORY_SIZE);
-    }
-    else
-#endif
-    {
-        bh_assert(*memory_data_size <= MAX_LINEAR_MEMORY_SIZE);
-    }
+    bh_assert(*memory_data_size <= GET_MAX_LINEAR_MEMORY_SIZE(is_memory64));
     *memory_data_size = align_as_and_cast(*memory_data_size, page_size);
 
     if (map_size > 0) {
diff --git a/core/iwasm/common/wasm_runtime_common.h b/core/iwasm/common/wasm_runtime_common.h
index dc41f72d..f38eb717 100644
--- a/core/iwasm/common/wasm_runtime_common.h
+++ b/core/iwasm/common/wasm_runtime_common.h
@@ -362,6 +362,9 @@ LOAD_I16(void *addr)
 #define SHARED_MEMORY_UNLOCK(memory) (void)0
 #endif
 
+#define CLAMP_U64_TO_U32(value) \
+    ((value) > UINT32_MAX ? UINT32_MAX : (uint32)(value))
+
 typedef struct WASMModuleCommon {
     /* Module type, for module loaded from WASM bytecode binary,
        this field is Wasm_Module_Bytecode, and this structure should
diff --git a/core/iwasm/compilation/aot_compiler.c b/core/iwasm/compilation/aot_compiler.c
index 694307ae..5c66e0fe 100644
--- a/core/iwasm/compilation/aot_compiler.c
+++ b/core/iwasm/compilation/aot_compiler.c
@@ -84,37 +84,46 @@ read_leb(const uint8 *buf, const uint8 *buf_end, uint32 *p_offset,
 }
 
 /* NOLINTNEXTLINE */
-#define read_leb_uint32(p, p_end, res)                    \
-    do {                                                  \
-        uint32 off = 0;                                   \
-        uint64 res64;                                     \
-        if (!read_leb(p, p_end, &off, 32, false, &res64)) \
-            return false;                                 \
-        p += off;                                         \
-        res = (uint32)res64;                              \
+#define read_leb_generic(p, p_end, res, res_type, sign)                     \
+    do {                                                                    \
+        uint32 off = 0;                                                     \
+        uint64 res64;                                                       \
+        if (!read_leb(p, p_end, &off, sizeof(res_type) << 3, sign, &res64)) \
+            return false;                                                   \
+        p += off;                                                           \
+        res = (res_type)res64;                                              \
     } while (0)
 
 /* NOLINTNEXTLINE */
-#define read_leb_int32(p, p_end, res)                    \
-    do {                                                 \
-        uint32 off = 0;                                  \
-        uint64 res64;                                    \
-        if (!read_leb(p, p_end, &off, 32, true, &res64)) \
-            return false;                                \
-        p += off;                                        \
-        res = (int32)res64;                              \
-    } while (0)
+#define read_leb_int32(p, p_end, res) \
+    read_leb_generic(p, p_end, res, int32, true)
 
 /* NOLINTNEXTLINE */
-#define read_leb_int64(p, p_end, res)                    \
-    do {                                                 \
-        uint32 off = 0;                                  \
-        uint64 res64;                                    \
-        if (!read_leb(p, p_end, &off, 64, true, &res64)) \
-            return false;                                \
-        p += off;                                        \
-        res = (int64)res64;                              \
+#define read_leb_int64(p, p_end, res) \
+    read_leb_generic(p, p_end, res, int64, true)
+
+/* NOLINTNEXTLINE */
+#define read_leb_uint32(p, p_end, res) \
+    read_leb_generic(p, p_end, res, uint32, false)
+
+/* NOLINTNEXTLINE */
+#define read_leb_uint64(p, p_end, res) \
+    read_leb_generic(p, p_end, res, uint64, false)
+
+/* NOLINTNEXTLINE */
+#if WASM_ENABLE_MEMORY64 != 0
+#define read_leb_mem_offset(p, p_end, res)  \
+    do {                                    \
+        if (IS_MEMORY64) {                  \
+            read_leb_uint64(p, p_end, res); \
+        }                                   \
+        else {                              \
+            read_leb_uint32(p, p_end, res); \
+        }                                   \
     } while (0)
+#else
+#define read_leb_mem_offset read_leb_uint32
+#endif
 
 /**
  * Since wamrc uses a full feature Wasm loader,
@@ -135,6 +144,13 @@ aot_validate_wasm(AOTCompContext *comp_ctx)
         }
     }
 
+#if WASM_ENABLE_MEMORY64 != 0
+    if (comp_ctx->pointer_size < sizeof(uint64) && IS_MEMORY64) {
+        aot_set_last_error("Compiling wasm64 to 32bit platform is not allowed");
+        return false;
+    }
+#endif
+
     return true;
 }
 
@@ -933,7 +949,8 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
     uint16 result_count;
     uint32 br_depth, *br_depths, br_count;
     uint32 func_idx, type_idx, mem_idx, local_idx, global_idx, i;
-    uint32 bytes = 4, align, offset;
+    uint32 bytes = 4, align;
+    mem_offset_t offset;
     uint32 type_index;
     bool sign = true;
     int32 i32_const;
@@ -1892,7 +1909,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                 sign = (opcode == WASM_OP_I32_LOAD16_S) ? true : false;
             op_i32_load:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_i32_load(comp_ctx, func_ctx, align, offset,
                                              bytes, sign, false))
                     return false;
@@ -1918,7 +1935,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                 sign = (opcode == WASM_OP_I64_LOAD32_S) ? true : false;
             op_i64_load:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_i64_load(comp_ctx, func_ctx, align, offset,
                                              bytes, sign, false))
                     return false;
@@ -1926,14 +1943,14 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
 
             case WASM_OP_F32_LOAD:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_f32_load(comp_ctx, func_ctx, align, offset))
                     return false;
                 break;
 
             case WASM_OP_F64_LOAD:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_f64_load(comp_ctx, func_ctx, align, offset))
                     return false;
                 break;
@@ -1948,7 +1965,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                 bytes = 2;
             op_i32_store:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_i32_store(comp_ctx, func_ctx, align, offset,
                                               bytes, false))
                     return false;
@@ -1967,7 +1984,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                 bytes = 4;
             op_i64_store:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_i64_store(comp_ctx, func_ctx, align, offset,
                                               bytes, false))
                     return false;
@@ -1975,7 +1992,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
 
             case WASM_OP_F32_STORE:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_f32_store(comp_ctx, func_ctx, align,
                                               offset))
                     return false;
@@ -1983,7 +2000,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
 
             case WASM_OP_F64_STORE:
                 read_leb_uint32(frame_ip, frame_ip_end, align);
-                read_leb_uint32(frame_ip, frame_ip_end, offset);
+                read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 if (!aot_compile_op_f64_store(comp_ctx, func_ctx, align,
                                               offset))
                     return false;
@@ -2540,7 +2557,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
 
                 if (opcode != WASM_OP_ATOMIC_FENCE) {
                     read_leb_uint32(frame_ip, frame_ip_end, align);
-                    read_leb_uint32(frame_ip, frame_ip_end, offset);
+                    read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                 }
                 switch (opcode) {
                     case WASM_OP_ATOMIC_WAIT32:
@@ -2705,7 +2722,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_load:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_v128_load(comp_ctx, func_ctx,
                                                         align, offset))
                             return false;
@@ -2720,7 +2737,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_load32x2_u:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_load_extend(
                                 comp_ctx, func_ctx, opcode, align, offset))
                             return false;
@@ -2733,7 +2750,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_load64_splat:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_load_splat(comp_ctx, func_ctx,
                                                          opcode, align, offset))
                             return false;
@@ -2743,7 +2760,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_store:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_v128_store(comp_ctx, func_ctx,
                                                          align, offset))
                             return false;
@@ -3006,7 +3023,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_load64_lane:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_load_lane(comp_ctx, func_ctx,
                                                         opcode, align, offset,
                                                         *frame_ip++))
@@ -3020,7 +3037,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_store64_lane:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_store_lane(comp_ctx, func_ctx,
                                                          opcode, align, offset,
                                                          *frame_ip++))
@@ -3032,7 +3049,7 @@ aot_compile_func(AOTCompContext *comp_ctx, uint32 func_index)
                     case SIMD_v128_load64_zero:
                     {
                         read_leb_uint32(frame_ip, frame_ip_end, align);
-                        read_leb_uint32(frame_ip, frame_ip_end, offset);
+                        read_leb_mem_offset(frame_ip, frame_ip_end, offset);
                         if (!aot_compile_simd_load_zero(comp_ctx, func_ctx,
                                                         opcode, align, offset))
                             return false;
diff --git a/core/iwasm/compilation/aot_compiler.h b/core/iwasm/compilation/aot_compiler.h
index 08e9db7b..f9016ac6 100644
--- a/core/iwasm/compilation/aot_compiler.h
+++ b/core/iwasm/compilation/aot_compiler.h
@@ -519,6 +519,15 @@ set_local_gc_ref(AOTCompFrame *frame, int n, LLVMValueRef value, uint8 ref_type)
         wasm_runtime_free(aot_value);                                        \
     } while (0)
 
+#if WASM_ENABLE_MEMORY64 != 0
+#define IS_MEMORY64 \
+    (comp_ctx->comp_data->memories[0].memory_flags & MEMORY64_FLAG)
+#define MEMORY64_COND_VALUE(VAL_IF_ENABLED, VAL_IF_DISABLED) \
+    (IS_MEMORY64 ? VAL_IF_ENABLED : VAL_IF_DISABLED)
+#else
+#define MEMORY64_COND_VALUE(VAL_IF_ENABLED, VAL_IF_DISABLED) (VAL_IF_DISABLED)
+#endif
+
 #define POP_I32(v) POP(v, VALUE_TYPE_I32)
 #define POP_I64(v) POP(v, VALUE_TYPE_I64)
 #define POP_F32(v) POP(v, VALUE_TYPE_F32)
@@ -527,6 +536,10 @@ set_local_gc_ref(AOTCompFrame *frame, int n, LLVMValueRef value, uint8 ref_type)
 #define POP_FUNCREF(v) POP(v, VALUE_TYPE_FUNCREF)
 #define POP_EXTERNREF(v) POP(v, VALUE_TYPE_EXTERNREF)
 #define POP_GC_REF(v) POP(v, VALUE_TYPE_GC_REF)
+#define POP_MEM_OFFSET(v) \
+    POP(v, MEMORY64_COND_VALUE(VALUE_TYPE_I64, VALUE_TYPE_I32))
+#define POP_PAGE_COUNT(v) \
+    POP(v, MEMORY64_COND_VALUE(VALUE_TYPE_I64, VALUE_TYPE_I32))
 
 #define POP_COND(llvm_value)                                                   \
     do {                                                                       \
@@ -590,6 +603,8 @@ set_local_gc_ref(AOTCompFrame *frame, int n, LLVMValueRef value, uint8 ref_type)
 #define PUSH_FUNCREF(v) PUSH(v, VALUE_TYPE_FUNCREF)
 #define PUSH_EXTERNREF(v) PUSH(v, VALUE_TYPE_EXTERNREF)
 #define PUSH_GC_REF(v) PUSH(v, VALUE_TYPE_GC_REF)
+#define PUSH_PAGE_COUNT(v) \
+    PUSH(v, MEMORY64_COND_VALUE(VALUE_TYPE_I64, VALUE_TYPE_I32))
 
 #define TO_LLVM_TYPE(wasm_type) \
     wasm_type_to_llvm_type(comp_ctx, &comp_ctx->basic_types, wasm_type)
diff --git a/core/iwasm/compilation/aot_emit_memory.c b/core/iwasm/compilation/aot_emit_memory.c
index 4582cdd0..a506f2a7 100644
--- a/core/iwasm/compilation/aot_emit_memory.c
+++ b/core/iwasm/compilation/aot_emit_memory.c
@@ -38,6 +38,20 @@
 
 #define SET_BUILD_POS(block) LLVMPositionBuilderAtEnd(comp_ctx->builder, block)
 
+static bool
+zero_extend_u64(AOTCompContext *comp_ctx, LLVMValueRef *value, const char *name)
+{
+    if (comp_ctx->pointer_size == sizeof(uint64)) {
+        /* zero extend to uint64 if the target is 64-bit */
+        *value = LLVMBuildZExt(comp_ctx->builder, *value, I64_TYPE, name);
+        if (!*value) {
+            aot_set_last_error("llvm build zero extend failed.");
+            return false;
+        }
+    }
+    return true;
+}
+
 static LLVMValueRef
 get_memory_check_bound(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
                        uint32 bytes)
@@ -82,9 +96,10 @@ get_memory_curr_page_count(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx);
 
 LLVMValueRef
 aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                          uint32 offset, uint32 bytes, bool enable_segue)
+                          mem_offset_t offset, uint32 bytes, bool enable_segue)
 {
-    LLVMValueRef offset_const = I32_CONST(offset);
+    LLVMValueRef offset_const =
+        MEMORY64_COND_VALUE(I64_CONST(offset), I32_CONST(offset));
     LLVMValueRef addr, maddr, offset1, cmp1, cmp2, cmp;
     LLVMValueRef mem_base_addr, mem_check_bound;
     LLVMBasicBlockRef block_curr = LLVMGetInsertBlock(comp_ctx->builder);
@@ -94,17 +109,27 @@ aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
     bool is_target_64bit, is_local_of_aot_value = false;
 #if WASM_ENABLE_SHARED_MEMORY != 0
     bool is_shared_memory =
-        comp_ctx->comp_data->memories[0].memory_flags & 0x02;
+        comp_ctx->comp_data->memories[0].memory_flags & SHARED_MEMORY_FLAG;
 #endif
 
     is_target_64bit = (comp_ctx->pointer_size == sizeof(uint64)) ? true : false;
 
     if (comp_ctx->is_indirect_mode
-        && aot_intrinsic_check_capability(comp_ctx, "i32.const")) {
+        && aot_intrinsic_check_capability(
+            comp_ctx, MEMORY64_COND_VALUE("i64.const", "i32.const"))) {
         WASMValue wasm_value;
-        wasm_value.i32 = offset;
+#if WASM_ENABLE_MEMORY64 != 0
+        if (IS_MEMORY64) {
+            wasm_value.i64 = offset;
+        }
+        else
+#endif
+        {
+            wasm_value.i32 = (int32)offset;
+        }
         offset_const = aot_load_const_from_table(
-            comp_ctx, func_ctx->native_symbol, &wasm_value, VALUE_TYPE_I32);
+            comp_ctx, func_ctx->native_symbol, &wasm_value,
+            MEMORY64_COND_VALUE(VALUE_TYPE_I64, VALUE_TYPE_I32));
         if (!offset_const) {
             return NULL;
         }
@@ -139,7 +164,7 @@ aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
         local_idx_of_aot_value = aot_value_top->local_idx;
     }
 
-    POP_I32(addr);
+    POP_MEM_OFFSET(addr);
 
     /*
      * Note: not throw the integer-overflow-exception here since it must
@@ -158,7 +183,7 @@ aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
         if (mem_offset + bytes <= mem_data_size) {
             /* inside memory space */
             if (comp_ctx->pointer_size == sizeof(uint64))
-                offset1 = I64_CONST((uint32)mem_offset);
+                offset1 = I64_CONST(mem_offset);
             else
                 offset1 = I32_CONST((uint32)mem_offset);
             CHECK_LLVM_CONST(offset1);
@@ -206,7 +231,8 @@ aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
             if (!(mem_size = get_memory_curr_page_count(comp_ctx, func_ctx))) {
                 goto fail;
             }
-            BUILD_ICMP(LLVMIntEQ, mem_size, I32_ZERO, cmp, "is_zero");
+            BUILD_ICMP(LLVMIntEQ, mem_size,
+                       MEMORY64_COND_VALUE(I64_ZERO, I32_ZERO), cmp, "is_zero");
             ADD_BASIC_BLOCK(check_succ, "check_mem_size_succ");
             LLVMMoveBasicBlockAfter(check_succ, block_curr);
             if (!aot_emit_exception(comp_ctx, func_ctx,
@@ -412,8 +438,8 @@ fail:
 
 bool
 aot_compile_op_i32_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset, uint32 bytes, bool sign,
-                        bool atomic)
+                        uint32 align, mem_offset_t offset, uint32 bytes,
+                        bool sign, bool atomic)
 {
     LLVMValueRef maddr, value = NULL;
     LLVMTypeRef data_type;
@@ -482,8 +508,8 @@ fail:
 
 bool
 aot_compile_op_i64_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset, uint32 bytes, bool sign,
-                        bool atomic)
+                        uint32 align, mem_offset_t offset, uint32 bytes,
+                        bool sign, bool atomic)
 {
     LLVMValueRef maddr, value = NULL;
     LLVMTypeRef data_type;
@@ -560,7 +586,7 @@ fail:
 
 bool
 aot_compile_op_f32_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset)
+                        uint32 align, mem_offset_t offset)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_f32_load;
@@ -583,7 +609,7 @@ fail:
 
 bool
 aot_compile_op_f64_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset)
+                        uint32 align, mem_offset_t offset)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_f64_load;
@@ -606,7 +632,8 @@ fail:
 
 bool
 aot_compile_op_i32_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset, uint32 bytes, bool atomic)
+                         uint32 align, mem_offset_t offset, uint32 bytes,
+                         bool atomic)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_i32_store;
@@ -656,7 +683,8 @@ fail:
 
 bool
 aot_compile_op_i64_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset, uint32 bytes, bool atomic)
+                         uint32 align, mem_offset_t offset, uint32 bytes,
+                         bool atomic)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_i64_store;
@@ -713,7 +741,7 @@ fail:
 
 bool
 aot_compile_op_f32_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset)
+                         uint32 align, mem_offset_t offset)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_f32_store;
@@ -736,7 +764,7 @@ fail:
 
 bool
 aot_compile_op_f64_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset)
+                         uint32 align, mem_offset_t offset)
 {
     LLVMValueRef maddr, value;
     bool enable_segue = comp_ctx->enable_segue_f64_store;
@@ -774,7 +802,8 @@ get_memory_curr_page_count(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
         }
     }
 
-    return mem_size;
+    return LLVMBuildIntCast(comp_ctx->builder, mem_size,
+                            MEMORY64_COND_VALUE(I64_TYPE, I32_TYPE), "");
 fail:
     return NULL;
 }
@@ -785,7 +814,7 @@ aot_compile_op_memory_size(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     LLVMValueRef mem_size = get_memory_curr_page_count(comp_ctx, func_ctx);
 
     if (mem_size)
-        PUSH_I32(mem_size);
+        PUSH_PAGE_COUNT(mem_size);
     return mem_size ? true : false;
 fail:
     return false;
@@ -798,11 +827,14 @@ aot_compile_op_memory_grow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     LLVMValueRef delta, param_values[2], ret_value, func, value;
     LLVMTypeRef param_types[2], ret_type, func_type, func_ptr_type;
     int32 func_index;
+#if WASM_ENABLE_MEMORY64 != 0
+    LLVMValueRef u32_max, u32_cmp_result;
+#endif
 
     if (!mem_size)
         return false;
 
-    POP_I32(delta);
+    POP_PAGE_COUNT(delta);
 
     /* Function type of aot_enlarge_memory() */
     param_types[0] = INT8_PTR_TYPE;
@@ -854,7 +886,7 @@ aot_compile_op_memory_grow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
 
     /* Call function aot_enlarge_memory() */
     param_values[0] = func_ctx->aot_inst;
-    param_values[1] = delta;
+    param_values[1] = LLVMBuildTrunc(comp_ctx->builder, delta, I32_TYPE, "");
     if (!(ret_value = LLVMBuildCall2(comp_ctx->builder, func_type, func,
                                      param_values, 2, "call"))) {
         aot_set_last_error("llvm build call failed.");
@@ -862,15 +894,26 @@ aot_compile_op_memory_grow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     }
 
     BUILD_ICMP(LLVMIntUGT, ret_value, I8_ZERO, ret_value, "mem_grow_ret");
+#if WASM_ENABLE_MEMORY64 != 0
+    if (IS_MEMORY64) {
+        if (!(u32_max = I64_CONST(UINT32_MAX))) {
+            aot_set_last_error("llvm build const failed");
+            return false;
+        }
+        BUILD_ICMP(LLVMIntULE, delta, u32_max, u32_cmp_result, "page_size_cmp");
+        BUILD_OP(And, ret_value, u32_cmp_result, ret_value, "and");
+    }
+#endif
 
-    /* ret_value = ret_value == true ? delta : pre_page_count */
-    if (!(ret_value = LLVMBuildSelect(comp_ctx->builder, ret_value, mem_size,
-                                      I32_NEG_ONE, "mem_grow_ret"))) {
+    /* ret_value = ret_value == true ? pre_page_count : -1 */
+    if (!(ret_value = LLVMBuildSelect(
+              comp_ctx->builder, ret_value, mem_size,
+              MEMORY64_COND_VALUE(I64_NEG_ONE, I32_NEG_ONE), "mem_grow_ret"))) {
         aot_set_last_error("llvm build select failed.");
         return false;
     }
 
-    PUSH_I32(ret_value);
+    PUSH_PAGE_COUNT(ret_value);
     return true;
 fail:
     return false;
@@ -987,13 +1030,17 @@ aot_compile_op_memory_init(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
 
     POP_I32(len);
     POP_I32(offset);
-    POP_I32(dst);
+    POP_MEM_OFFSET(dst);
+
+    if (!zero_extend_u64(comp_ctx, &dst, "dst64")) {
+        return false;
+    }
 
     param_types[0] = INT8_PTR_TYPE;
     param_types[1] = I32_TYPE;
     param_types[2] = I32_TYPE;
     param_types[3] = I32_TYPE;
-    param_types[4] = I32_TYPE;
+    param_types[4] = SIZE_T_TYPE;
     ret_type = INT8_TYPE;
 
     if (comp_ctx->is_jit_mode)
@@ -1080,9 +1127,9 @@ aot_compile_op_memory_copy(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     LLVMValueRef src, dst, src_addr, dst_addr, len, res;
     bool call_aot_memmove = false;
 
-    POP_I32(len);
-    POP_I32(src);
-    POP_I32(dst);
+    POP_MEM_OFFSET(len);
+    POP_MEM_OFFSET(src);
+    POP_MEM_OFFSET(dst);
 
     if (!(src_addr = check_bulk_memory_overflow(comp_ctx, func_ctx, src, len)))
         return false;
@@ -1090,13 +1137,8 @@ aot_compile_op_memory_copy(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     if (!(dst_addr = check_bulk_memory_overflow(comp_ctx, func_ctx, dst, len)))
         return false;
 
-    if (comp_ctx->pointer_size == sizeof(uint64)) {
-        /* zero extend to uint64 if the target is 64-bit */
-        len = LLVMBuildZExt(comp_ctx->builder, len, I64_TYPE, "len64");
-        if (!len) {
-            aot_set_last_error("llvm build zero extend failed.");
-            return false;
-        }
+    if (!zero_extend_u64(comp_ctx, &len, "len64")) {
+        return false;
     }
 
     call_aot_memmove = comp_ctx->is_indirect_mode || comp_ctx->is_jit_mode;
@@ -1174,20 +1216,15 @@ aot_compile_op_memory_fill(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx)
     LLVMTypeRef param_types[3], ret_type, func_type, func_ptr_type;
     LLVMValueRef func, params[3];
 
-    POP_I32(len);
+    POP_MEM_OFFSET(len);
     POP_I32(val);
-    POP_I32(dst);
+    POP_MEM_OFFSET(dst);
 
     if (!(dst_addr = check_bulk_memory_overflow(comp_ctx, func_ctx, dst, len)))
         return false;
 
-    if (comp_ctx->pointer_size == sizeof(uint64)) {
-        /* zero extend to uint64 if the target is 64-bit */
-        len = LLVMBuildZExt(comp_ctx->builder, len, I64_TYPE, "len64");
-        if (!len) {
-            aot_set_last_error("llvm build zero extend failed.");
-            return false;
-        }
+    if (!zero_extend_u64(comp_ctx, &len, "len64")) {
+        return false;
     }
 
     param_types[0] = INT8_PTR_TYPE;
@@ -1251,7 +1288,7 @@ fail:
 bool
 aot_compile_op_atomic_rmw(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
                           uint8 atomic_op, uint8 op_type, uint32 align,
-                          uint32 offset, uint32 bytes)
+                          mem_offset_t offset, uint32 bytes)
 {
     LLVMValueRef maddr, value, result;
     bool enable_segue = (op_type == VALUE_TYPE_I32)
@@ -1337,7 +1374,7 @@ fail:
 bool
 aot_compile_op_atomic_cmpxchg(AOTCompContext *comp_ctx,
                               AOTFuncContext *func_ctx, uint8 op_type,
-                              uint32 align, uint32 offset, uint32 bytes)
+                              uint32 align, mem_offset_t offset, uint32 bytes)
 {
     LLVMValueRef maddr, value, expect, result;
     bool enable_segue = (op_type == VALUE_TYPE_I32)
@@ -1442,7 +1479,7 @@ fail:
 
 bool
 aot_compile_op_atomic_wait(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 op_type, uint32 align, uint32 offset,
+                           uint8 op_type, uint32 align, mem_offset_t offset,
                            uint32 bytes)
 {
     LLVMValueRef maddr, value, timeout, expect, cmp;
@@ -1534,7 +1571,7 @@ fail:
 bool
 aot_compiler_op_atomic_notify(AOTCompContext *comp_ctx,
                               AOTFuncContext *func_ctx, uint32 align,
-                              uint32 offset, uint32 bytes)
+                              mem_offset_t offset, uint32 bytes)
 {
     LLVMValueRef maddr, value, count;
     LLVMValueRef param_values[3], ret_value, func;
diff --git a/core/iwasm/compilation/aot_emit_memory.h b/core/iwasm/compilation/aot_emit_memory.h
index e174aa3d..f1828f17 100644
--- a/core/iwasm/compilation/aot_emit_memory.h
+++ b/core/iwasm/compilation/aot_emit_memory.h
@@ -17,43 +17,43 @@ extern "C" {
 
 bool
 aot_compile_op_i32_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset, uint32 bytes, bool sign,
-                        bool atomic);
+                        uint32 align, mem_offset_t offset, uint32 bytes,
+                        bool sign, bool atomic);
 
 bool
 aot_compile_op_i64_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset, uint32 bytes, bool sign,
-                        bool atomic);
+                        uint32 align, mem_offset_t offset, uint32 bytes,
+                        bool sign, bool atomic);
 
 bool
 aot_compile_op_f32_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset);
+                        uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_op_f64_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                        uint32 align, uint32 offset);
+                        uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_op_i32_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset, uint32 bytes,
+                         uint32 align, mem_offset_t offset, uint32 bytes,
                          bool atomic);
 
 bool
 aot_compile_op_i64_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset, uint32 bytes,
+                         uint32 align, mem_offset_t offset, uint32 bytes,
                          bool atomic);
 
 bool
 aot_compile_op_f32_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset);
+                         uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_op_f64_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                         uint32 align, uint32 offset);
+                         uint32 align, mem_offset_t offset);
 
 LLVMValueRef
 aot_check_memory_overflow(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                          uint32 offset, uint32 bytes, bool enable_segue);
+                          mem_offset_t offset, uint32 bytes, bool enable_segue);
 
 bool
 aot_compile_op_memory_size(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx);
@@ -89,22 +89,22 @@ aot_compile_op_memory_fill(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx);
 bool
 aot_compile_op_atomic_rmw(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
                           uint8 atomic_op, uint8 op_type, uint32 align,
-                          uint32 offset, uint32 bytes);
+                          mem_offset_t offset, uint32 bytes);
 
 bool
 aot_compile_op_atomic_cmpxchg(AOTCompContext *comp_ctx,
                               AOTFuncContext *func_ctx, uint8 op_type,
-                              uint32 align, uint32 offset, uint32 bytes);
+                              uint32 align, mem_offset_t offset, uint32 bytes);
 
 bool
 aot_compile_op_atomic_wait(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 op_type, uint32 align, uint32 offset,
+                           uint8 op_type, uint32 align, mem_offset_t offset,
                            uint32 bytes);
 
 bool
 aot_compiler_op_atomic_notify(AOTCompContext *comp_ctx,
                               AOTFuncContext *func_ctx, uint32 align,
-                              uint32 offset, uint32 bytes);
+                              mem_offset_t offset, uint32 bytes);
 
 bool
 aot_compiler_op_atomic_fence(AOTCompContext *comp_ctx,
diff --git a/core/iwasm/compilation/simd/simd_load_store.c b/core/iwasm/compilation/simd/simd_load_store.c
index 0e869727..45829f08 100644
--- a/core/iwasm/compilation/simd/simd_load_store.c
+++ b/core/iwasm/compilation/simd/simd_load_store.c
@@ -13,7 +13,7 @@
 /* data_length in bytes */
 static LLVMValueRef
 simd_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx, uint32 align,
-          uint32 offset, uint32 data_length, LLVMTypeRef ptr_type,
+          mem_offset_t offset, uint32 data_length, LLVMTypeRef ptr_type,
           LLVMTypeRef data_type, bool enable_segue)
 {
     LLVMValueRef maddr, data;
@@ -42,7 +42,7 @@ simd_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx, uint32 align,
 
 bool
 aot_compile_simd_v128_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint32 align, uint32 offset)
+                           uint32 align, mem_offset_t offset)
 {
     bool enable_segue = comp_ctx->enable_segue_v128_load;
     LLVMTypeRef v128_ptr_type = enable_segue ? V128_PTR_TYPE_GS : V128_PTR_TYPE;
@@ -62,7 +62,7 @@ fail:
 
 bool
 aot_compile_simd_load_extend(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                             uint8 opcode, uint32 align, uint32 offset)
+                             uint8 opcode, uint32 align, mem_offset_t offset)
 {
     LLVMValueRef sub_vector, result;
     uint32 opcode_index = opcode - SIMD_v128_load8x8_s;
@@ -117,7 +117,7 @@ aot_compile_simd_load_extend(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
 
 bool
 aot_compile_simd_load_splat(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint8 opcode, uint32 align, uint32 offset)
+                            uint8 opcode, uint32 align, mem_offset_t offset)
 {
     uint32 opcode_index = opcode - SIMD_v128_load8_splat;
     LLVMValueRef element, result;
@@ -173,7 +173,7 @@ aot_compile_simd_load_splat(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
 
 bool
 aot_compile_simd_load_lane(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 opcode, uint32 align, uint32 offset,
+                           uint8 opcode, uint32 align, mem_offset_t offset,
                            uint8 lane_id)
 {
     LLVMValueRef element, vector;
@@ -218,7 +218,7 @@ aot_compile_simd_load_lane(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
 
 bool
 aot_compile_simd_load_zero(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 opcode, uint32 align, uint32 offset)
+                           uint8 opcode, uint32 align, mem_offset_t offset)
 {
     LLVMValueRef element, result, mask;
     uint32 opcode_index = opcode - SIMD_v128_load32_zero;
@@ -308,7 +308,7 @@ simd_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx, uint32 align,
 
 bool
 aot_compile_simd_v128_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint32 align, uint32 offset)
+                            uint32 align, mem_offset_t offset)
 {
     bool enable_segue = comp_ctx->enable_segue_v128_store;
     LLVMTypeRef v128_ptr_type = enable_segue ? V128_PTR_TYPE_GS : V128_PTR_TYPE;
@@ -324,7 +324,7 @@ fail:
 
 bool
 aot_compile_simd_store_lane(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint8 opcode, uint32 align, uint32 offset,
+                            uint8 opcode, uint32 align, mem_offset_t offset,
                             uint8 lane_id)
 {
     LLVMValueRef element, vector;
diff --git a/core/iwasm/compilation/simd/simd_load_store.h b/core/iwasm/compilation/simd/simd_load_store.h
index fd118ec1..7a98cbb6 100644
--- a/core/iwasm/compilation/simd/simd_load_store.h
+++ b/core/iwasm/compilation/simd/simd_load_store.h
@@ -14,32 +14,32 @@ extern "C" {
 
 bool
 aot_compile_simd_v128_load(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint32 align, uint32 offset);
+                           uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_simd_load_extend(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                             uint8 opcode, uint32 align, uint32 offset);
+                             uint8 opcode, uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_simd_load_splat(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint8 opcode, uint32 align, uint32 offset);
+                            uint8 opcode, uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_simd_load_lane(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 opcode, uint32 align, uint32 offset,
+                           uint8 opcode, uint32 align, mem_offset_t offset,
                            uint8 lane_id);
 
 bool
 aot_compile_simd_load_zero(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                           uint8 opcode, uint32 align, uint32 offset);
+                           uint8 opcode, uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_simd_v128_store(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint32 align, uint32 offset);
+                            uint32 align, mem_offset_t offset);
 
 bool
 aot_compile_simd_store_lane(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
-                            uint8 opcode, uint32 align, uint32 offset,
+                            uint8 opcode, uint32 align, mem_offset_t offset,
                             uint8 lane_id);
 
 #ifdef __cplusplus
diff --git a/core/iwasm/interpreter/wasm.h b/core/iwasm/interpreter/wasm.h
index 5401153a..d56ef3b3 100644
--- a/core/iwasm/interpreter/wasm.h
+++ b/core/iwasm/interpreter/wasm.h
@@ -500,8 +500,10 @@ typedef struct WASMTable {
 
 #if WASM_ENABLE_MEMORY64 != 0
 typedef uint64 mem_offset_t;
+#define PR_MEM_OFFSET PRIu64
 #else
 typedef uint32 mem_offset_t;
+#define PR_MEM_OFFSET PRIu32
 #endif
 
 typedef struct WASMMemory {
diff --git a/core/iwasm/interpreter/wasm_loader.c b/core/iwasm/interpreter/wasm_loader.c
index 8af0fd9f..fd447ec3 100644
--- a/core/iwasm/interpreter/wasm_loader.c
+++ b/core/iwasm/interpreter/wasm_loader.c
@@ -9,6 +9,7 @@
 #include "wasm.h"
 #include "wasm_opcode.h"
 #include "wasm_runtime.h"
+#include "wasm_loader_common.h"
 #include "../common/wasm_native.h"
 #include "../common/wasm_memory.h"
 #if WASM_ENABLE_GC != 0
@@ -2755,43 +2756,6 @@ check_memory_max_size(bool is_memory64, uint32 init_size, uint32 max_size,
     return true;
 }
 
-static bool
-check_memory_flag(const uint8 mem_flag, char *error_buf, uint32 error_buf_size)
-{
-    /* Check whether certain features indicated by mem_flag are enabled in
-     * runtime */
-    if (mem_flag > MAX_PAGE_COUNT_FLAG) {
-#if WASM_ENABLE_SHARED_MEMORY == 0
-        if (mem_flag & SHARED_MEMORY_FLAG) {
-            LOG_VERBOSE("shared memory flag was found, please enable shared "
-                        "memory, lib-pthread or lib-wasi-threads");
-            set_error_buf(error_buf, error_buf_size, "invalid limits flags");
-            return false;
-        }
-#endif
-#if WASM_ENABLE_MEMORY64 == 0
-        if (mem_flag & MEMORY64_FLAG) {
-            LOG_VERBOSE("memory64 flag was found, please enable memory64");
-            set_error_buf(error_buf, error_buf_size, "invalid limits flags");
-            return false;
-        }
-#endif
-    }
-
-    if (mem_flag > MAX_PAGE_COUNT_FLAG + SHARED_MEMORY_FLAG + MEMORY64_FLAG) {
-        set_error_buf(error_buf, error_buf_size, "invalid limits flags");
-        return false;
-    }
-    else if ((mem_flag & SHARED_MEMORY_FLAG)
-             && !(mem_flag & MAX_PAGE_COUNT_FLAG)) {
-        set_error_buf(error_buf, error_buf_size,
-                      "shared memory must have maximum");
-        return false;
-    }
-
-    return true;
-}
-
 static bool
 load_memory_import(const uint8 **p_buf, const uint8 *buf_end,
                    WASMModule *parent_module, const char *sub_module_name,
@@ -2824,7 +2788,7 @@ load_memory_import(const uint8 **p_buf, const uint8 *buf_end,
         return false;
     }
 
-    if (!check_memory_flag(mem_flag, error_buf, error_buf_size)) {
+    if (!wasm_memory_check_flags(mem_flag, error_buf, error_buf_size, false)) {
         return false;
     }
 
@@ -3226,7 +3190,8 @@ load_memory(const uint8 **p_buf, const uint8 *buf_end, WASMMemory *memory,
         return false;
     }
 
-    if (!check_memory_flag(memory->flags, error_buf, error_buf_size)) {
+    if (!wasm_memory_check_flags(memory->flags, error_buf, error_buf_size,
+                                 false)) {
         return false;
     }
 
@@ -14762,9 +14727,9 @@ re_scan:
                             goto fail;
                         }
 
-                        read_leb_uint32(p, p_end, mem_offset); /* offset */
+                        read_leb_mem_offset(p, p_end, mem_offset); /* offset */
 
-                        POP_AND_PUSH(VALUE_TYPE_I32, VALUE_TYPE_V128);
+                        POP_AND_PUSH(mem_offset_type, VALUE_TYPE_V128);
 #if WASM_ENABLE_JIT != 0 || WASM_ENABLE_WAMR_COMPILER != 0
                         func->has_memory_operations = true;
 #endif
@@ -14781,10 +14746,10 @@ re_scan:
                             goto fail;
                         }
 
-                        read_leb_uint32(p, p_end, mem_offset); /* offset */
+                        read_leb_mem_offset(p, p_end, mem_offset); /* offset */
 
                         POP_V128();
-                        POP_I32();
+                        POP_MEM_OFFSET();
 #if WASM_ENABLE_JIT != 0 || WASM_ENABLE_WAMR_COMPILER != 0
                         func->has_memory_operations = true;
 #endif
@@ -14999,7 +14964,7 @@ re_scan:
                             goto fail;
                         }
 
-                        read_leb_uint32(p, p_end, mem_offset); /* offset */
+                        read_leb_mem_offset(p, p_end, mem_offset); /* offset */
 
                         CHECK_BUF(p, p_end, 1);
                         lane = read_uint8(p);
@@ -15009,7 +14974,7 @@ re_scan:
                         }
 
                         POP_V128();
-                        POP_I32();
+                        POP_MEM_OFFSET();
                         if (opcode1 < SIMD_v128_store8_lane) {
                             PUSH_V128();
                         }
@@ -15030,9 +14995,9 @@ re_scan:
                             goto fail;
                         }
 
-                        read_leb_uint32(p, p_end, mem_offset); /* offset */
+                        read_leb_mem_offset(p, p_end, mem_offset); /* offset */
 
-                        POP_AND_PUSH(VALUE_TYPE_I32, VALUE_TYPE_V128);
+                        POP_AND_PUSH(mem_offset_type, VALUE_TYPE_V128);
 #if WASM_ENABLE_JIT != 0 || WASM_ENABLE_WAMR_COMPILER != 0
                         func->has_memory_operations = true;
 #endif
diff --git a/core/iwasm/interpreter/wasm_mini_loader.c b/core/iwasm/interpreter/wasm_mini_loader.c
index ff37d818..efdb06bc 100644
--- a/core/iwasm/interpreter/wasm_mini_loader.c
+++ b/core/iwasm/interpreter/wasm_mini_loader.c
@@ -11,6 +11,7 @@
 #include "wasm_runtime.h"
 #include "../common/wasm_native.h"
 #include "../common/wasm_memory.h"
+#include "wasm_loader_common.h"
 #if WASM_ENABLE_FAST_JIT != 0
 #include "../fast-jit/jit_compiler.h"
 #include "../fast-jit/jit_codecache.h"
@@ -714,38 +715,6 @@ load_table_import(const uint8 **p_buf, const uint8 *buf_end,
     return true;
 }
 
-static bool
-check_memory_flag(const uint8 mem_flag)
-{
-    /* Check whether certain features indicated by mem_flag are enabled in
-     * runtime */
-    if (mem_flag > MAX_PAGE_COUNT_FLAG) {
-#if WASM_ENABLE_SHARED_MEMORY == 0
-        if (mem_flag & SHARED_MEMORY_FLAG) {
-            LOG_VERBOSE("shared memory flag was found, please enable shared "
-                        "memory, lib-pthread or lib-wasi-threads");
-            return false;
-        }
-#endif
-#if WASM_ENABLE_MEMORY64 == 0
-        if (mem_flag & MEMORY64_FLAG) {
-            LOG_VERBOSE("memory64 flag was found, please enable memory64");
-            return false;
-        }
-#endif
-    }
-
-    if (mem_flag > MAX_PAGE_COUNT_FLAG + SHARED_MEMORY_FLAG + MEMORY64_FLAG) {
-        return false;
-    }
-    else if ((mem_flag & SHARED_MEMORY_FLAG)
-             && !(mem_flag & MAX_PAGE_COUNT_FLAG)) {
-        return false;
-    }
-
-    return true;
-}
-
 static bool
 load_memory_import(const uint8 **p_buf, const uint8 *buf_end,
                    WASMModule *parent_module, const char *sub_module_name,
@@ -766,7 +735,9 @@ load_memory_import(const uint8 **p_buf, const uint8 *buf_end,
     uint32 declare_max_page_count = 0;
 
     read_leb_uint32(p, p_end, mem_flag);
-    bh_assert(check_memory_flag(mem_flag));
+    if (!wasm_memory_check_flags(mem_flag, error_buf, error_buf_size, false)) {
+        return false;
+    }
 
 #if WASM_ENABLE_APP_FRAMEWORK == 0
     is_memory64 = mem_flag & MEMORY64_FLAG;
@@ -796,7 +767,6 @@ load_memory_import(const uint8 **p_buf, const uint8 *buf_end,
     memory->num_bytes_per_page = DEFAULT_NUM_BYTES_PER_PAGE;
 
     *p_buf = p;
-    (void)check_memory_flag;
     return true;
 }
 
@@ -891,7 +861,10 @@ load_memory(const uint8 **p_buf, const uint8 *buf_end, WASMMemory *memory,
     read_leb_uint32(p, p_end, memory->flags);
     bh_assert(p - p_org <= 1);
     (void)p_org;
-    bh_assert(check_memory_flag(memory->flags));
+    if (!wasm_memory_check_flags(memory->flags, error_buf, error_buf_size,
+                                 false)) {
+        return false;
+    }
 
 #if WASM_ENABLE_APP_FRAMEWORK == 0
     is_memory64 = memory->flags & MEMORY64_FLAG;
@@ -916,7 +889,6 @@ load_memory(const uint8 **p_buf, const uint8 *buf_end, WASMMemory *memory,
     memory->num_bytes_per_page = DEFAULT_NUM_BYTES_PER_PAGE;
 
     *p_buf = p;
-    (void)check_memory_flag;
     return true;
 }
 
diff --git a/core/iwasm/interpreter/wasm_runtime.c b/core/iwasm/interpreter/wasm_runtime.c
index c0e325b5..0a952224 100644
--- a/core/iwasm/interpreter/wasm_runtime.c
+++ b/core/iwasm/interpreter/wasm_runtime.c
@@ -1408,19 +1408,23 @@ execute_malloc_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
 #endif
     WASMExecEnv *exec_env_created = NULL;
     WASMModuleInstanceCommon *module_inst_old = NULL;
-    uint32 argv[3], argc;
+    union {
+        uint32 u32[3];
+        uint64 u64;
+    } argv;
+    uint32 argc;
     bool ret;
 #if WASM_ENABLE_MEMORY64 != 0
     bool is_memory64 = module_inst->memories[0]->is_memory64;
     if (is_memory64) {
         argc = 2;
-        PUT_I64_TO_ADDR(&argv[0], size);
+        PUT_I64_TO_ADDR(&argv.u64, size);
     }
     else
 #endif
     {
         argc = 1;
-        argv[0] = (uint32)size;
+        argv.u32[0] = (uint32)size;
     }
 
     /* if __retain is exported, then this module is compiled by
@@ -1431,7 +1435,7 @@ execute_malloc_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
         /* the malloc function from assemblyscript is:
             function __new(size: usize, id: u32)
             id = 0 means this is an ArrayBuffer object */
-        argv[argc] = 0;
+        argv.u32[argc] = 0;
         argc++;
     }
 
@@ -1472,10 +1476,10 @@ execute_malloc_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
         }
     }
 
-    ret = wasm_call_function(exec_env, malloc_func, argc, argv);
+    ret = wasm_call_function(exec_env, malloc_func, argc, argv.u32);
 
     if (retain_func && ret)
-        ret = wasm_call_function(exec_env, retain_func, 1, argv);
+        ret = wasm_call_function(exec_env, retain_func, 1, argv.u32);
 
     if (module_inst_old)
         /* Restore the existing exec_env's module inst */
@@ -1487,11 +1491,11 @@ execute_malloc_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
     if (ret) {
 #if WASM_ENABLE_MEMORY64 != 0
         if (is_memory64)
-            *p_result = GET_I64_FROM_ADDR(&argv[0]);
+            *p_result = GET_I64_FROM_ADDR(&argv.u64);
         else
 #endif
         {
-            *p_result = argv[0];
+            *p_result = argv.u32[0];
         }
     }
     return ret;
@@ -1506,18 +1510,22 @@ execute_free_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
 #endif
     WASMExecEnv *exec_env_created = NULL;
     WASMModuleInstanceCommon *module_inst_old = NULL;
-    uint32 argv[2], argc;
+    union {
+        uint32 u32[2];
+        uint64 u64;
+    } argv;
+    uint32 argc;
     bool ret;
 
 #if WASM_ENABLE_MEMORY64 != 0
     if (module_inst->memories[0]->is_memory64) {
-        PUT_I64_TO_ADDR(&argv[0], offset);
+        PUT_I64_TO_ADDR(&argv.u64, offset);
         argc = 2;
     }
     else
 #endif
     {
-        argv[0] = (uint32)offset;
+        argv.u32[0] = (uint32)offset;
         argc = 1;
     }
 
@@ -1558,7 +1566,7 @@ execute_free_function(WASMModuleInstance *module_inst, WASMExecEnv *exec_env,
         }
     }
 
-    ret = wasm_call_function(exec_env, free_func, argc, argv);
+    ret = wasm_call_function(exec_env, free_func, argc, argv.u32);
 
     if (module_inst_old)
         /* Restore the existing exec_env's module inst */
@@ -4176,7 +4184,7 @@ fail:
 #if WASM_ENABLE_BULK_MEMORY != 0
 bool
 llvm_jit_memory_init(WASMModuleInstance *module_inst, uint32 seg_index,
-                     uint32 offset, uint32 len, uint32 dst)
+                     uint32 offset, uint32 len, size_t dst)
 {
     WASMMemoryInstance *memory_inst;
     WASMModule *module;
@@ -4211,7 +4219,7 @@ llvm_jit_memory_init(WASMModuleInstance *module_inst, uint32 seg_index,
         (WASMModuleInstanceCommon *)module_inst, (uint64)dst);
 
     SHARED_MEMORY_LOCK(memory_inst);
-    bh_memcpy_s(maddr, (uint32)(memory_inst->memory_data_size - dst),
+    bh_memcpy_s(maddr, CLAMP_U64_TO_U32(memory_inst->memory_data_size - dst),
                 data + offset, len);
     SHARED_MEMORY_UNLOCK(memory_inst);
     return true;
diff --git a/core/iwasm/interpreter/wasm_runtime.h b/core/iwasm/interpreter/wasm_runtime.h
index e60d4f7c..8ff34ee3 100644
--- a/core/iwasm/interpreter/wasm_runtime.h
+++ b/core/iwasm/interpreter/wasm_runtime.h
@@ -760,7 +760,7 @@ llvm_jit_invoke_native(WASMExecEnv *exec_env, uint32 func_idx, uint32 argc,
 #if WASM_ENABLE_BULK_MEMORY != 0
 bool
 llvm_jit_memory_init(WASMModuleInstance *module_inst, uint32 seg_index,
-                     uint32 offset, uint32 len, uint32 dst);
+                     uint32 offset, uint32 len, size_t dst);
 
 bool
 llvm_jit_data_drop(WASMModuleInstance *module_inst, uint32 seg_index);
diff --git a/doc/build_wamr.md b/doc/build_wamr.md
index e8b6ad47..f1c32f44 100644
--- a/doc/build_wamr.md
+++ b/doc/build_wamr.md
@@ -79,7 +79,7 @@ cmake -DWAMR_BUILD_PLATFORM=linux -DWAMR_BUILD_TARGET=ARM
 #### **Enable memory64 feature**
 - **WAMR_BUILD_MEMORY64**=1/0, default to disable if not set
 
-> Note: Currently, the memory64 feature is only supported in classic interpreter running mode.
+> Note: Currently, the memory64 feature is only supported in classic interpreter running mode and AOT mode.
 
 #### **Enable thread manager**
 - **WAMR_BUILD_THREAD_MGR**=1/0, default to disable if not set
diff --git a/product-mini/platforms/alios-things/aos.mk b/product-mini/platforms/alios-things/aos.mk
index 3f25cb98..947a4a91 100644
--- a/product-mini/platforms/alios-things/aos.mk
+++ b/product-mini/platforms/alios-things/aos.mk
@@ -111,6 +111,7 @@ $(NAME)_SOURCES := ${SHARED_ROOT}/platform/alios/alios_platform.c \
                    ${IWASM_ROOT}/common/wasm_runtime_common.c \
                    ${IWASM_ROOT}/common/wasm_native.c \
                    ${IWASM_ROOT}/common/wasm_exec_env.c \
+                   ${IWASM_ROOT}/common/wasm_loader_common.c \
                    ${IWASM_ROOT}/common/wasm_memory.c \
                    ${IWASM_ROOT}/common/wasm_c_api.c \
                    ${IWASM_ROOT}/common/arch/${INVOKE_NATIVE} \
diff --git a/product-mini/platforms/nuttx/wamr.mk b/product-mini/platforms/nuttx/wamr.mk
index 7aac0e35..75bd69be 100644
--- a/product-mini/platforms/nuttx/wamr.mk
+++ b/product-mini/platforms/nuttx/wamr.mk
@@ -451,6 +451,7 @@ CSRCS += nuttx_platform.c \
          wasm_runtime_common.c \
          wasm_native.c \
          wasm_exec_env.c \
+         wasm_loader_common.c \
          wasm_memory.c \
          wasm_c_api.c
 
diff --git a/tests/wamr-test-suites/spec-test-script/runtest.py b/tests/wamr-test-suites/spec-test-script/runtest.py
index 13229d97..3c32481f 100755
--- a/tests/wamr-test-suites/spec-test-script/runtest.py
+++ b/tests/wamr-test-suites/spec-test-script/runtest.py
@@ -1119,9 +1119,6 @@ def compile_wasm_to_aot(wasm_tempfile, aot_tempfile, runner, opts, r, output = '
         cmd.append("--enable-gc")
         cmd.append("--enable-tail-call")
 
-    if opts.memory64:
-        cmd.append("--enable-memory64")
-
     if output == 'object':
         cmd.append("--format=object")
     elif output == 'ir':
@@ -1134,9 +1131,10 @@ def compile_wasm_to_aot(wasm_tempfile, aot_tempfile, runner, opts, r, output = '
 
     # Bounds checks is disabled by default for 64-bit targets, to
     # use the hardware based bounds checks. But it is not supported
-    # in QEMU with NuttX.
-    # Enable bounds checks explicitly for all targets if running in QEMU.
-    if opts.qemu:
+    # in QEMU with NuttX and in memory64 mode.
+    # Enable bounds checks explicitly for all targets if running in QEMU or all targets
+    # running in memory64 mode.
+    if opts.qemu or opts.memory64:
         cmd.append("--bounds-checks=1")
 
     # RISCV64 requires -mcmodel=medany, which can be set by --size-level=1
diff --git a/tests/wamr-test-suites/spec-test-script/thread_proposal_remove_memory64_flag_case.patch b/tests/wamr-test-suites/spec-test-script/thread_proposal_remove_memory64_flag_case.patch
new file mode 100644
index 00000000..0ed07e3d
--- /dev/null
+++ b/tests/wamr-test-suites/spec-test-script/thread_proposal_remove_memory64_flag_case.patch
@@ -0,0 +1,20 @@
+diff --git a/test/core/binary.wast b/test/core/binary.wast
+index b9fa438c..08ecee27 100644
+--- a/test/core/binary.wast
++++ b/test/core/binary.wast
+@@ -621,15 +621,6 @@
+ )
+
+ ;; Malformed memory limits flag
+-(assert_malformed
+-  (module binary
+-      "\00asm" "\01\00\00\00"
+-      "\05\03\01"                           ;; memory section with one entry
+-      "\04"                                 ;; malformed memory limits flag
+-      "\00"                                 ;; min 0
+-  )
+-  "malformed limits flags"
+-)
+ (assert_malformed
+   (module binary
+       "\00asm" "\01\00\00\00"
diff --git a/tests/wamr-test-suites/test_wamr.sh b/tests/wamr-test-suites/test_wamr.sh
index b9b298a5..0b27df46 100755
--- a/tests/wamr-test-suites/test_wamr.sh
+++ b/tests/wamr-test-suites/test_wamr.sh
@@ -409,6 +409,19 @@ function setup_wabt()
     fi
 }
 
+function compile_reference_interpreter()
+{
+    echo "compile the reference intepreter"
+    pushd interpreter
+    make
+    if [ $? -ne 0 ]
+    then
+        echo "Failed to compile the reference interpreter"
+        exit 1
+    fi
+    popd
+}
+
 # TODO: with iwasm only
 function spec_test()
 {
@@ -457,6 +470,7 @@ function spec_test()
 
         git apply ../../spec-test-script/thread_proposal_ignore_cases.patch
         git apply ../../spec-test-script/thread_proposal_fix_atomic_case.patch
+        git apply ../../spec-test-script/thread_proposal_remove_memory64_flag_case.patch
     fi
 
     if [ ${ENABLE_EH} == 1 ]; then
@@ -500,10 +514,7 @@ function spec_test()
             git apply ../../spec-test-script/gc_nuttx_tail_call.patch
         fi
 
-        echo "compile the reference intepreter"
-        pushd interpreter
-        make
-        popd
+        compile_reference_interpreter
     fi
 
     # update memory64 cases
@@ -519,14 +530,11 @@ function spec_test()
         git restore . && git clean -ffd .
         # Reset to commit: "Merge remote-tracking branch 'upstream/main' into merge2"
         git reset --hard 48e69f394869c55b7bbe14ac963c09f4605490b6
-        git checkout 044d0d2e77bdcbe891f7e0b9dd2ac01d56435f0b -- test/core/elem.wast
+        git checkout 044d0d2e77bdcbe891f7e0b9dd2ac01d56435f0b -- test/core/elem.wast test/core/data.wast
         git apply ../../spec-test-script/ignore_cases.patch
         git apply ../../spec-test-script/memory64.patch
 
-        echo "compile the reference intepreter"
-        pushd interpreter
-        make
-        popd
+        compile_reference_interpreter
     fi
 
     popd
diff --git a/wamr-compiler/CMakeLists.txt b/wamr-compiler/CMakeLists.txt
index 6a3f9752..245b4a03 100644
--- a/wamr-compiler/CMakeLists.txt
+++ b/wamr-compiler/CMakeLists.txt
@@ -46,6 +46,7 @@ add_definitions(-DWASM_ENABLE_DUMP_CALL_STACK=1)
 add_definitions(-DWASM_ENABLE_PERF_PROFILING=1)
 add_definitions(-DWASM_ENABLE_LOAD_CUSTOM_SECTION=1)
 add_definitions(-DWASM_ENABLE_MODULE_INST_CONTEXT=1)
+add_definitions(-DWASM_ENABLE_MEMORY64=1)
 
 add_definitions(-DWASM_ENABLE_GC=1)