linux-ramdump-parser-v2: Correct sizes for binary logbuf

The sizes for reading some of the structures of the binary logbuf
do not match exactly. Switch to explicitly sized functions due to
size changes across architectures.

Change-Id: Iae09a069f5cb641c61a63d864ec59d0c8359e65b

linux-ramdump-parser-v2/parsers/dmesg.py

@@ -33,7 +33,7 @@ class Dmesg(RamParser):
         len_offset = ramdump.field_offset('struct log', 'len')
 
         msg = logbuf + idx
-        msg_len = ramdump.read_word(msg + len_offset)
+        msg_len = ramdump.read_u16(msg + len_offset)
         if (msg_len == 0):
             return logbuf
         else:
@@ -43,7 +43,7 @@ class Dmesg(RamParser):
         len_offset = ramdump.field_offset('struct log', 'len')
         msg = idx
 
-        msg_len = ramdump.read_halfword(msg + len_offset)
+        msg_len = ramdump.read_u16(msg + len_offset)
         if (msg_len == 0):
             self.wrap_cnt += 1
             return logbuf
@@ -59,14 +59,14 @@ class Dmesg(RamParser):
         text_len_offset = ramdump.field_offset('struct log', 'text_len')
         log_size = ramdump.sizeof('struct log')
 
-        first_idx = ramdump.read_word(first_idx_addr)
-        last_idx = ramdump.read_word(last_idx_addr)
+        first_idx = ramdump.read_u32(first_idx_addr)
+        last_idx = ramdump.read_u32(last_idx_addr)
 
         curr_idx = logbuf_addr + first_idx
 
         while curr_idx != logbuf_addr + last_idx and self.wrap_cnt < 2:
             timestamp = ramdump.read_dword(curr_idx + time_offset)
-            text_len = ramdump.read_halfword(curr_idx + text_len_offset)
+            text_len = ramdump.read_u16(curr_idx + text_len_offset)
             text_str = ramdump.read_cstring(curr_idx + log_size, text_len)
             for partial in text_str.split('\n'):
                 f = '[{0:>5}.{1:0>6d}] {2}'.format(