binman: Add a new 'image-pos' property

At present each entry has an offset within its parent section. This is
useful for figuring out how entries relate to one another. However it
is sometimes necessary to locate an entry within an image, regardless
of which sections it is nested inside.

Add a new 'image-pos' property to provide this information. Also add
some documentation for the -u option binman provides, which updates the
device tree with final entry information.

Since the image position is a better symbol to use for the position of
U-Boot as obtained by SPL, update the SPL symbols to use this instead of
offset, which might be incorrect if hierarchical sections are used.

Signed-off-by: Simon Glass <sjg@chromium.org>

tools/binman/entry.py

@@ -64,6 +64,7 @@ class Entry(object):
         self.pad_before = 0
         self.pad_after = 0
         self.offset_unset = False
+        self.image_pos = None
         if read_node:
             self.ReadNode()
 
@@ -133,7 +134,7 @@ class Entry(object):
 
     def AddMissingProperties(self):
         """Add new properties to the device tree as needed for this entry"""
-        for prop in ['offset', 'size']:
+        for prop in ['offset', 'size', 'image-pos']:
             if not prop in self._node.props:
                 self._node.AddZeroProp(prop)
 
@@ -141,6 +142,7 @@ class Entry(object):
         """Set the value of device-tree properties calculated by binman"""
         self._node.SetInt('offset', self.offset)
         self._node.SetInt('size', self.size)
+        self._node.SetInt('image-pos', self.image_pos)
 
     def ProcessFdt(self, fdt):
         return True
@@ -265,6 +267,14 @@ class Entry(object):
         self.offset = pos
         self.size = size
 
+    def SetImagePos(self, image_pos):
+        """Set the position in the image
+
+        Args:
+            image_pos: Position of this entry in the image
+        """
+        self.image_pos = image_pos + self.offset
+
     def ProcessContents(self):
         pass