diff --git a/docs/index.rst b/docs/index.rst
index fa7469f..6b90173 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -15,7 +15,7 @@ Overall structure
 For the runtime program, you always need *pb.h* for type declarations.
 Depending on whether you want to encode, decode, or both, you also need *pb_encode.h/c* or *pb_decode.h/c*.
 
-If your *.proto* file encodes submessages or other fields using pointers, you must compile *pb_decode.c* with a preprocessor macro named *MALLOC_HEADER* that is the name of a header with definitions (either as functions or macros) for *malloc()*, *realloc()* and *free()*.  For a typical hosted configuration, this should be *<stdlib.h>*.
+If your *.proto* file encodes submessages or other fields using pointers, you must compile *pb_decode.c* with a preprocessor macro named *MALLOC_HEADER* that is the name of a header with definitions (either as functions or macros) for *calloc()*, *realloc()* and *free()*.  For a typical hosted configuration, this should be *<stdlib.h>*.
 
 The high-level encoding and decoding functions take an array of *pb_field_t* structures, which describes the fields of a message structure. Usually you want these autogenerated from a *.proto* file. The tool script *nanopb_generator.py* accomplishes this.
 
diff --git a/pb_decode.c b/pb_decode.c
index 8931887..e4d0b3e 100644
--- a/pb_decode.c
+++ b/pb_decode.c
@@ -661,10 +661,11 @@ bool checkreturn pb_dec_submessage(pb_istream_t *stream, const pb_field_t *field
 #ifdef MALLOC_HEADER
         if (*(void**)dest == NULL)
         {
-            void *object = malloc(msg->size);
+            void *object = calloc(1, msg->size);
             if (!object)
                 return false;
             *(void**)dest = object;
+            pb_message_set_to_defaults(msg, object);
             dest = object;
         } else {
             dest = *(void**)dest;
diff --git a/tests/decode_unittests.c b/tests/decode_unittests.c
index 352146c..6ac03ed 100644
--- a/tests/decode_unittests.c
+++ b/tests/decode_unittests.c
@@ -287,12 +287,12 @@ int main()
         
         memset(&dest, 0, sizeof(dest));
 #ifdef MALLOC_HEADER
-        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x2A\x01\x65\x32\x01\x66\x3A\x00"
-                  "\x42\x01\x63\x4A\x01\x64"),
+        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x1A\x00\x2A\x01\x65\x32\x01\x66"
+                  "\x3A\x00\x42\x01\x63\x4A\x01\x64"),
               pb_decode(&s, PointerContainer_msg, &dest)))
         TEST(0 == strcmp(dest.text, "a"))
         TEST(dest.blob.size == 1 && dest.blob.bytes[0] == 'b')
-        TEST(dest.submsg == NULL)
+        TEST(dest.submsg != NULL && dest.submsg->data == 10)
         TEST(dest.rtext_count == 1 && (0 == strcmp(dest.rtext[0], "e")))
         TEST(dest.rblob_count == 1 && dest.rblob[0].size == 1 &&
              dest.rblob[0].bytes[0] == 'f')
@@ -301,8 +301,8 @@ int main()
         TEST(dest.oblob.size == 1 && dest.oblob.bytes[0] == 'd')
         TEST(pb_clean(PointerContainer_msg, &dest));
 #else
-        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x2A\x01\x65\x32\x01\x66\x3A\x00"
-                  "\x42\x01\x63\x4A\x01\x64"),
+        TEST((s = S("\x0A\x01\x61\x12\x01\x62\x2A\x01\x65\x32\x01\x66\x01\x66"
+                  "\x3A\x00\x42\x01\x63\x4A\x01\x64"),
               !pb_decode(&s, PointerContainer_msg, &dest)))
 #endif
     }
diff --git a/tests/encode_unittests.c b/tests/encode_unittests.c
index d3b3151..bf332e4 100644
--- a/tests/encode_unittests.c
+++ b/tests/encode_unittests.c
@@ -273,7 +273,7 @@ int main()
         uint8_t buffer[128];
         pb_ostream_t s;
         PointerContainer msg;
-        PointerContainer msg2;
+        DefaultContainer msg2;
         IntegerArray msg3;
         
         COMMENT("Test pb_encode with pointer fields.")
@@ -285,7 +285,7 @@ int main()
         msg.blob.bytes = (uint8_t*)"b";
         msg.submsg = &msg2;
         TEST(WRITES(pb_encode(&s, PointerContainer_msg, &msg),
-                    "\x0A\x01\x61\x12\x01\x62"))
+                    "\x0A\x01\x61\x12\x01\x62\x1A\x00"))
         
         memset(&msg3, 0, sizeof(msg3));
         msg.rtext_count = 1;
@@ -296,7 +296,7 @@ int main()
         msg.rsubmsg_count = 1;
         msg.rsubmsg[0] = &msg3;
         TEST(WRITES(pb_encode(&s, PointerContainer_msg, &msg),
-                    "\x0A\x01\x61\x12\x01\x62"
+                    "\x0A\x01\x61\x12\x01\x62\x1A\x00"
                     "\x2A\x01\x65\x32\x01\x66\x3A\x00"));
         
         PointerContainer_set(msg, otext);
@@ -305,7 +305,7 @@ int main()
         msg.oblob.size = 1;
         msg.oblob.bytes = (uint8_t*)"d";
         TEST(WRITES(pb_encode(&s, PointerContainer_msg, &msg),
-                    "\x0A\x01\x61\x12\x01\x62"
+                    "\x0A\x01\x61\x12\x01\x62\x1A\x00"
                     "\x2A\x01\x65\x32\x01\x66\x3A\x00"
                     "\x42\x01\x63\x4A\x01\x64"));
     }
diff --git a/tests/unittestproto.proto b/tests/unittestproto.proto
index 89d571e..84ec970 100644
--- a/tests/unittestproto.proto
+++ b/tests/unittestproto.proto
@@ -27,10 +27,14 @@ message CallbackContainerContainer {
     required CallbackContainer submsg = 1;
 }
 
+message DefaultContainer {
+    optional int32 data = 1 [default = 10];
+}
+
 message PointerContainer {
     required string text = 1 [(nanopb).pointer = true];
     required bytes blob = 2 [(nanopb).pointer = true];
-    optional PointerContainer submsg = 3 [(nanopb).pointer = true];
+    required DefaultContainer submsg = 3 [(nanopb).pointer = true];
     // This should be rejected:
     // required int32 data = 4 [(nanopb).pointer = true];
     repeated string rtext = 5 [(nanopb).pointer = true, (nanopb).max_count = 10];