IntelFsp2Pkg/Tools/ConfigEditor: Fixed data difference issue

Problem Description:
load_default_from_bin caused data difference
errors when loading fsp.bsf and fsp.fd,
due to multiple identical UPD signatures.
The original string search picked the first match,
which could be the wrong region.

Solution Description:
Updated GenYamlCfg.py to locate UPD data
by FFS GUID FSP_T, FSP_M, FSP_S_UPD_FFS_GUID
before searching UPD signatures.
This ensures correct region matching
and retains the old string search as a fallback.

Signed-off-by: Hsu JasonX <[email protected]>

Author: yunhsiax

IntelFsp2Pkg/Tools/ConfigEditor/GenYamlCfg.py

@@ -18,6 +18,42 @@
 from CommonUtility import value_to_bytearray, value_to_bytes, \
       bytes_to_value, get_bits_from_bytes, set_bits_to_bytes
 
+# FFS file structure constants
+FFS_GUID_SIZE = 16
+FFS_HEADER_SIZE = 24
+FFS_SEARCH_RANGE = 4096  # Search within 4KB after GUID
+
+# FSP UPD FFS GUID definitions
+FSP_T_UPD_FFS_GUID = '70BCF6A5-FFB1-47D8-B1AE-EFE5508E23EA'
+FSP_M_UPD_FFS_GUID = 'D5B86AEA-6AF7-40D4-8014-982301BC3D89'
+FSP_S_UPD_FFS_GUID = 'E3CD9B18-998C-4F76-B65E-98B154E5446F'
+
+def guid_str_to_bytes(guid_str):
+    """Convert GUID string to bytes in EFI GUID format (mixed endian)"""
+    # Remove dashes and convert to bytes
+    guid_str = guid_str.replace('-', '')
+
+    # EFI GUID format: first 3 fields are little-endian, last 2 are big-endian
+    # Parse: DWORD-WORD-WORD-BYTE[2]-BYTE[6]
+    data1 = bytes.fromhex(guid_str[0:8])[::-1]    # DWORD (4 bytes) - little endian
+    data2 = bytes.fromhex(guid_str[8:12])[::-1]   # WORD (2 bytes) - little endian
+    data3 = bytes.fromhex(guid_str[12:16])[::-1]  # WORD (2 bytes) - little endian
+    data4 = bytes.fromhex(guid_str[16:20])        # BYTE[2] (2 bytes) - big endian
+    data5 = bytes.fromhex(guid_str[20:32])        # BYTE[6] (6 bytes) - big endian
+
+    return data1 + data2 + data3 + data4 + data5
+
+# UPD signature to GUID mapping
+def get_upd_guid(upd_signature):
+    """Get FFS GUID for a given UPD signature based on suffix"""
+    if upd_signature.endswith('UPD_T'):
+        return FSP_T_UPD_FFS_GUID
+    elif upd_signature.endswith('UPD_M'):
+        return FSP_M_UPD_FFS_GUID
+    elif upd_signature.endswith('UPD_S'):
+        return FSP_S_UPD_FFS_GUID
+    return None
+
 # Generated file copyright header
 __copyright_tmp__ = """/** @file
 
@@ -1639,6 +1675,46 @@ def _get_cfg_segment(name, cfgs, level):
 
         return cfg_segs
 
+    def find_upd_in_ffs_by_guid(self, bin_data, upd_signature, ffs_guid):
+        """
+        Find UPD signature within a specific FFS file identified by GUID.
+
+        Args:
+            bin_data: Binary data to search in
+            upd_signature: UPD signature string (e.g., 'XXXUPD_M')
+            ffs_guid: GUID string of the FFS file to search in
+
+        Returns:
+            Position of UPD signature if found, -1 otherwise
+        """
+        if not ffs_guid:
+            return -1
+
+        # Convert GUID string to bytes for searching
+        guid_bytes = guid_str_to_bytes(ffs_guid)
+        sig_bytes = upd_signature.encode()
+
+        # Search for all occurrences of the GUID
+        search_pos = 0
+        while True:
+            guid_pos = bin_data.find(guid_bytes, search_pos)
+            if guid_pos < 0:
+                break
+
+            # Search for UPD signature within FFS_SEARCH_RANGE after GUID
+            search_start = guid_pos
+            search_end = min(guid_pos + FFS_SEARCH_RANGE, len(bin_data))
+
+            sig_pos = bin_data.find(sig_bytes, search_start, search_end)
+            if sig_pos >= 0:
+                print(f"Found {upd_signature} at offset 0x{sig_pos:X} within FFS GUID {ffs_guid}")
+                return sig_pos
+
+            # Continue searching for next occurrence of this GUID
+            search_pos = guid_pos + FFS_GUID_SIZE
+
+        return -1
+
     def import_tkinter(self):
         try:
             import tkinter
@@ -1656,16 +1732,28 @@ def get_bin_segment(self, bin_data):
             if key == 0:
                 bin_segs.append([seg[0], 0, len(bin_data)])
                 break
-            pos = bin_data.find(key)
-            if pos >= 0:
-                # ensure no other match for the key
-                next_pos = bin_data.find(key, pos + len(seg[0]))
-                if next_pos >= 0:
-                    if key == b'$SKLFSP$' or key == b'$BSWFSP$':
-                        string = ('Multiple matches for %s in '
-                                  'binary!\n\nA workaround applied to such '
-                                  'FSP 1.x binary to use second'
-                                  ' match instead of first match!' % key)
+
+            pos = -1
+
+            # Check if this UPD signature has a corresponding FFS GUID
+            ffs_guid = get_upd_guid(seg[0])
+            if ffs_guid:
+                pos = self.find_upd_in_ffs_by_guid(bin_data, seg[0], ffs_guid)
+                if pos >= 0:
+                    print(f"Using GUID-based search: found {seg[0]} at 0x{pos:X}")
+
+            # Fallback to original search if GUID-based search fails
+            if pos < 0:
+                pos = bin_data.find(key)
+                if pos >= 0:
+                    # Check for multiple matches
+                    next_pos = bin_data.find(key, pos + len(seg[0]))
+                    if next_pos >= 0:
+                        if key == b'$SKLFSP$' or key == b'$BSWFSP$':
+                            string = ('Warning: Multiple matches for %s in '
+                                      'binary!\n\nA workaround applied to such '
+                                      'FSP 1.x binary to use second'
+                                      ' match instead of first match!' % key)
                         if self._tk:
                             self._tk.messagebox.showwarning('Warning: ', string)
                         else:
@@ -1675,12 +1763,13 @@ def get_bin_segment(self, bin_data):
                         print("Warning: Multiple matches for '%s' "
                               "in binary, the 1st instance will be used !"
                               % seg[0])
+
+            if pos >= 0:
                 bin_segs.append([seg[0], pos, seg[2]])
                 self.binseg_dict[seg[0]] = pos
             else:
                 bin_segs.append([seg[0], -1, seg[2]])
                 self.binseg_dict[seg[0]] = -1
-                continue
 
         return bin_segs