py/modstruct: Fix struct.unpack with unaligned offset of native type.

With this patch alignment is done relative to the start of the buffer that
is being unpacked, not the raw pointer value, as per CPython.

Fixes issue #3314.

5 files changed, 24 insertions, 6 deletions

diff --git a/extmod/moductypes.c b/extmod/moductypes.c
index 9b46371f3..d4c7611df 100644
--- a/extmod/moductypes.c
+++ b/extmod/moductypes.c
@@ -299,7 +299,7 @@ STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(uctypes_struct_sizeof_obj, 1, 2, ucty
 static inline mp_obj_t get_unaligned(uint val_type, byte *p, int big_endian) {
     char struct_type = big_endian ? '>' : '<';
     static const char type2char[16] = "BbHhIiQq------fd";
-    return mp_binary_get_val(struct_type, type2char[val_type], &p);
+    return mp_binary_get_val(struct_type, type2char[val_type], p, &p);
 }
 
 static inline void set_unaligned(uint val_type, byte *p, int big_endian, mp_obj_t val) {
diff --git a/py/binary.c b/py/binary.c
index a142776c3..9810e4660 100644
--- a/py/binary.c
+++ b/py/binary.c
@@ -185,14 +185,14 @@ long long mp_binary_get_int(mp_uint_t size, bool is_signed, bool big_endian, con
 }
 
 #define is_signed(typecode) (typecode > 'Z')
-mp_obj_t mp_binary_get_val(char struct_type, char val_type, byte **ptr) {
+mp_obj_t mp_binary_get_val(char struct_type, char val_type, byte *p_base, byte **ptr) {
     byte *p = *ptr;
     mp_uint_t align;
 
     size_t size = mp_binary_get_size(struct_type, val_type, &align);
     if (struct_type == '@') {
-        // Make pointer aligned
-        p = (byte*)MP_ALIGN(p, (size_t)align);
+        // Align p relative to p_base
+        p = p_base + (uintptr_t)MP_ALIGN(p - p_base, (size_t)align);
         #if MP_ENDIANNESS_LITTLE
         struct_type = '<';
         #else
diff --git a/py/binary.h b/py/binary.h
index 71182042f..092b72288 100644
--- a/py/binary.h
+++ b/py/binary.h
@@ -38,7 +38,7 @@ size_t mp_binary_get_size(char struct_type, char val_type, mp_uint_t *palign);
 mp_obj_t mp_binary_get_val_array(char typecode, void *p, mp_uint_t index);
 void mp_binary_set_val_array(char typecode, void *p, mp_uint_t index, mp_obj_t val_in);
 void mp_binary_set_val_array_from_int(char typecode, void *p, mp_uint_t index, mp_int_t val);
-mp_obj_t mp_binary_get_val(char struct_type, char val_type, byte **ptr);
+mp_obj_t mp_binary_get_val(char struct_type, char val_type, byte *p_base, byte **ptr);
 void mp_binary_set_val(char struct_type, char val_type, mp_obj_t val_in, byte **ptr);
 long long mp_binary_get_int(mp_uint_t size, bool is_signed, bool big_endian, const byte *src);
 void mp_binary_set_int(mp_uint_t val_sz, bool big_endian, byte *dest, mp_uint_t val);
diff --git a/py/modstruct.c b/py/modstruct.c
index 8617a8e0d..957e4917b 100644
--- a/py/modstruct.c
+++ b/py/modstruct.c
@@ -146,6 +146,7 @@ STATIC mp_obj_t struct_unpack_from(size_t n_args, const mp_obj_t *args) {
         }
         p += offset;
     }
+    byte *p_base = p;
 
     // Check that the input buffer is big enough to unpack all the values
     if (p + total_sz > end_p) {
@@ -164,7 +165,7 @@ STATIC mp_obj_t struct_unpack_from(size_t n_args, const mp_obj_t *args) {
             res->items[i++] = item;
         } else {
             while (cnt--) {
-                item = mp_binary_get_val(fmt_type, *fmt, &p);
+                item = mp_binary_get_val(fmt_type, *fmt, p_base, &p);
                 res->items[i++] = item;
             }
         }
diff --git a/tests/basics/struct_endian.py b/tests/basics/struct_endian.py
new file mode 100644
index 000000000..ae3243824
--- /dev/null
+++ b/tests/basics/struct_endian.py
@@ -0,0 +1,17 @@
+# test ustruct and endian specific things
+
+try:
+    import ustruct as struct
+except:
+    try:
+        import struct
+    except ImportError:
+        print("SKIP")
+        raise SystemExit
+
+# unpack/unpack_from with unaligned native type
+buf = b'0123456789'
+print(struct.unpack('h', memoryview(buf)[1:3]))
+print(struct.unpack_from('i', buf, 1))
+print(struct.unpack_from('@i', buf, 1))
+print(struct.unpack_from('@ii', buf, 1))