diff --git a/core/iwasm/fast-jit/cg/x86-64/jit_codegen_x86_64.cpp b/core/iwasm/fast-jit/cg/x86-64/jit_codegen_x86_64.cpp
index 89b12da00..4714c5d9b 100644
--- a/core/iwasm/fast-jit/cg/x86-64/jit_codegen_x86_64.cpp
+++ b/core/iwasm/fast-jit/cg/x86-64/jit_codegen_x86_64.cpp
@@ -696,9 +696,16 @@ mov_r_to_m(x86::Assembler &a, uint32 bytes_dst, uint32 kind_dst,
  * @return new stream
  */
 static bool
-mov_imm_to_m(x86::Assembler &a, x86::Mem &m_dst, Imm imm_src)
+mov_imm_to_m(x86::Assembler &a, x86::Mem &m_dst, Imm imm_src, uint32 bytes_dst)
 {
-    a.mov(m_dst, imm_src);
+    if (bytes_dst == 8) {
+        /* As there is no instruction `MOV m64, imm64`, we use
+           two instructions to implement it */
+        a.mov(regs_i64[REG_I64_FREE_IDX], imm_src);
+        a.mov(m_dst, regs_i64[REG_I64_FREE_IDX]);
+    }
+    else
+        a.mov(m_dst, imm_src);
     return true;
 }
 
@@ -931,7 +938,7 @@ st_imm_to_base_imm_offset_imm(x86::Assembler &a, uint32 bytes_dst,
     x86::Mem m((uintptr_t)(base + offset), bytes_dst);
     Imm imm;
     imm_set_value(imm, data_src, bytes_dst);
-    return mov_imm_to_m(a, m, imm);
+    return mov_imm_to_m(a, m, imm, bytes_dst);
 }
 
 /**
@@ -954,7 +961,7 @@ st_imm_to_base_imm_offset_r(x86::Assembler &a, uint32 bytes_dst, void *data_src,
     x86::Mem m(regs_i64[reg_no_offset], base, bytes_dst);
     Imm imm;
     imm_set_value(imm, data_src, bytes_dst);
-    return mov_imm_to_m(a, m, imm);
+    return mov_imm_to_m(a, m, imm, bytes_dst);
 }
 
 /**
@@ -977,7 +984,7 @@ st_imm_to_base_r_offset_imm(x86::Assembler &a, uint32 bytes_dst, void *data_src,
     x86::Mem m(regs_i64[reg_no_base], offset, bytes_dst);
     Imm imm;
     imm_set_value(imm, data_src, bytes_dst);
-    return mov_imm_to_m(a, m, imm);
+    return mov_imm_to_m(a, m, imm, bytes_dst);
 }
 
 /**
@@ -1001,7 +1008,7 @@ st_imm_to_base_r_offset_r(x86::Assembler &a, uint32 bytes_dst, void *data_src,
     x86::Mem m(regs_i64[reg_no_base], regs_i64[reg_no_offset], 0, 0, bytes_dst);
     Imm imm;
     imm_set_value(imm, data_src, bytes_dst);
-    return mov_imm_to_m(a, m, imm);
+    return mov_imm_to_m(a, m, imm, bytes_dst);
 }
 
 /**
diff --git a/core/iwasm/fast-jit/fe/jit_emit_control.c b/core/iwasm/fast-jit/fe/jit_emit_control.c
index 407a78544..f638c0901 100644
--- a/core/iwasm/fast-jit/fe/jit_emit_control.c
+++ b/core/iwasm/fast-jit/fe/jit_emit_control.c
@@ -454,7 +454,7 @@ handle_op_end(JitCompContext *cc, uint8 **p_frame_ip, bool is_block_polymorphic)
                     jit_basic_block_label(block->basic_block_end);
             }
             else if (insn->opcode == JIT_OP_BNE) {
-                *(jit_insn_opnd(insn, 1)) =
+                *(jit_insn_opnd(insn, 2)) =
                     jit_basic_block_label(block->basic_block_end);
             }
             else {
diff --git a/core/iwasm/interpreter/wasm_interp_classic.c b/core/iwasm/interpreter/wasm_interp_classic.c
index 11e8fd7a4..a38500513 100644
--- a/core/iwasm/interpreter/wasm_interp_classic.c
+++ b/core/iwasm/interpreter/wasm_interp_classic.c
@@ -837,8 +837,14 @@ jit_invoke_native(WASMExecEnv *exec_env, uint32 func_idx,
     WASMModuleInstance *module_inst =
         (WASMModuleInstance *)exec_env->module_inst;
     WASMFunctionInstance *cur_func = module_inst->functions + func_idx;
+    uint32 *sp_org;
 
+    sp_org = prev_frame->sp;
     wasm_interp_call_func_native(module_inst, exec_env, cur_func, prev_frame);
+    /* Restore the stack pointer of previous frame as the caller in
+       jitted code will just read the return value and won't decrease
+       the stack pointer */
+    prev_frame->sp = sp_org;
 
     return wasm_get_exception(module_inst) ? false : true;
 }