Add AOT_INTRINSIC_FLAG_F32_CONST for xtensa frontend, because espressif xtensa llvm backend does not support float-point immediate for now

Author: zhanheng1

core/iwasm/aot/aot_intrinsic.c

@@ -877,6 +877,7 @@ aot_intrinsic_fill_capability_flags(AOTCompContext *comp_ctx)
         add_f64_common_intrinsics(comp_ctx);
         add_i64_common_intrinsics(comp_ctx);
         add_common_float_integer_conversion(comp_ctx);
+	add_intrinsic_capability(comp_ctx, AOT_INTRINSIC_FLAG_F32_CONST);
     }
     else {
         /*

core/iwasm/compilation/aot_emit_compare.c

@@ -159,8 +159,7 @@ aot_compile_op_f32_compare(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
     POP_F32(lhs);
 
     if (comp_ctx->disable_llvm_intrinsics
-        && aot_intrinsic_check_capability(comp_ctx, "f32_cmp")
-        && strcmp(comp_ctx->target_arch, "xtensa")) {
+        && aot_intrinsic_check_capability(comp_ctx, "f32_cmp")) {
         LLVMTypeRef param_types[3];
         LLVMValueRef opcond = LLVMConstInt(I32_TYPE, cond, true);
         param_types[0] = I32_TYPE;