diff --git a/edlclient/Library/firehose.py b/edlclient/Library/firehose.py
index 8618612..8cf6122 100755
--- a/edlclient/Library/firehose.py
+++ b/edlclient/Library/firehose.py
@@ -1356,10 +1356,11 @@ class firehose(metaclass=LogBase):
 
         def cmd_patch_multiple(lun, start_sector, byte_offset, patch_data):
             offset = 0
-            size_each_patch = 4
+            size_each_patch = 8 if len(patch_data) % 8 == 0 else 4
+            unpack_fmt = "<I" if size_each_patch == 4 else "<Q"
             write_size = len(patch_data)
             for i in range(0, write_size, size_each_patch):
-                pdata_subset = int(unpack("<I", patch_data[offset:offset+size_each_patch])[0])
+                pdata_subset = int(unpack(unpack_fmt, patch_data[offset:offset+size_each_patch])[0])
                 self.cmd_patch( lun, start_sector, byte_offset + offset, pdata_subset, size_each_patch, True)
                 offset += size_each_patch
             return True
@@ -1479,7 +1480,7 @@ class firehose(metaclass=LogBase):
                                 _, lun_b, gpt_data_b, guid_gpt_b = resp
                                 backup_gpt_data_b, backup_guid_gpt_b = self.get_gpt(lun_b, 0, 0 , 0, guid_gpt_b.header.backup_lba)
 
-                            if not check_gpt_hdr:
+                            if not check_gpt_hdr and partitionname_a[:3] != "xbl": # xbl partition don't need check consistency
                                 sts, gpt_data_a, backup_gpt_data_a = ensure_gpt_hdr_consistency(guid_gpt_a, backup_guid_gpt_a, gpt_data_a, backup_gpt_data_a)
                                 if not sts:
                                     return False