Fix handling of LZ4 legacy frames.

There is no end magic after the compressed content, but we can detect it
given that except for the last block, all blocks are 8MB uncompressed.

Inspired by vmlinux-to-elf (see
https://github.com/marin-m/vmlinux-to-elf/blob/master/vmlinux_to_elf/utils/lz4_legacy.py).

Author: qkaiser

poetry.lock

@@ -28,6 +28,7 @@ pluggy = "^1.0.0"
 file-magic = "^0.4.0"
 hyperscan = "^0.2.0"
 lark = "^1.1.2"
+lz4 = "^4.0.0"
 
 
 [tool.poetry.dev-dependencies]

pyproject.toml

@@ -4,6 +4,7 @@
 import io
 from typing import Optional
 
+from lz4.block import decompress
 from structlog import get_logger
 
 from unblob.extractors import Command
@@ -28,6 +29,7 @@
     FRAME_SIZE_LEN
 ) = BLOCK_CHECKSUM_LEN = CONTENT_CHECKSUM_LEN = MAGIC_LEN = DICTID_LEN = 4
 FLG_LEN = BD_LEN = HC_LEN = 1
+MAX_LEGACY_BLOCK_SIZE = 8 * 1024 * 1024  # 8 MB
 
 
 class FLG:
@@ -89,7 +91,12 @@ def calculate_chunk(self, file: File, start_offset: int) -> Optional[ValidChunk]
                 file.seek(-4, io.SEEK_CUR)
                 break
 
-            file.seek(block_compressed_size, io.SEEK_CUR)
+            compressed_block = file.read(block_compressed_size)
+            uncompressed_block = decompress(compressed_block, MAX_LEGACY_BLOCK_SIZE)
+
+            # See 'fixed block size' in https://android.googlesource.com/platform/external/lz4/+/HEAD/doc/lz4_Frame_format.md#legacy-frame
+            if len(uncompressed_block) < MAX_LEGACY_BLOCK_SIZE:
+                break
 
         end_offset = file.tell()
         return ValidChunk(start_offset=start_offset, end_offset=end_offset)