feat(handler): Add support for HP BDL format

A BDL file contains multiple IPKG packages

The header contains useful informations, such as the brand, device_id, version of the BDL file, etc,

After that header, there is a table of content that has the offset and the size of each IPKG in the BDL file.

This repo https://github.com/tylerwhall/hpbdl/ contains information that was useful

Author: Antoine Pecoraro

tests/integration/archive/hp/bdl/__input__/sample.bdl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:240533e4188e0af859989fca729c4a0847de3d5688906c4a22d36f843afdf271
+size 4043

tests/integration/archive/hp/bdl/__output__/sample.bdl_extract/ipkg000

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c99dd4e215b508395aed42ebfee1e87d64b6bb5551c165934fe1a759e40093ee
+size 1345

tests/integration/archive/hp/bdl/__output__/sample.bdl_extract/ipkg000_extract/sample.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6478ea19e2122817e1986151847d3590e114ffd7c5d65d8d91cccea316c1cf2
+size 17

tests/integration/archive/hp/bdl/__output__/sample.bdl_extract/ipkg001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c99dd4e215b508395aed42ebfee1e87d64b6bb5551c165934fe1a759e40093ee
+size 1345

tests/integration/archive/hp/bdl/__output__/sample.bdl_extract/ipkg001_extract/sample.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6478ea19e2122817e1986151847d3590e114ffd7c5d65d8d91cccea316c1cf2
+size 17

unblob/handlers/__init__.py

@@ -2,6 +2,7 @@
 from .archive import ar, arc, arj, cab, cpio, dmg, rar, sevenzip, stuffit, tar, zip
 from .archive.dlink import encrpted_img, shrs
 from .archive.engeniustech import engenius
+from .archive.hp import bdl
 from .archive.instar import bneg
 from .archive.netgear import chk, trx
 from .archive.qnap import qnap_nas
@@ -62,6 +63,7 @@
     hdr.HDR2Handler,
     qnap_nas.QnapHandler,
     bneg.BNEGHandler,
+    bdl.HPBDLHandler,
     sparse.SparseHandler,
     ar.ARHandler,
     arc.ARCHandler,

unblob/handlers/archive/hp/bdl.py

@@ -0,0 +1,104 @@
+import io
+from pathlib import Path
+from typing import Optional
+
+from dissect.cstruct import Instance
+from structlog import get_logger
+
+from unblob.extractor import carve_chunk_to_file
+from unblob.file_utils import Endian, File, InvalidInputFormat, StructParser, snull
+from unblob.models import Chunk, Extractor, HexString, StructHandler, ValidChunk
+
+logger = get_logger()
+
+C_DEFINITIONS = r"""
+    typedef struct bdl_toc_entry {
+        uint64 offset;
+        uint64 size;
+    } bdl_toc_entry_t;
+
+    typedef struct bdl_header {
+        char magic[4];
+        uint16 major;
+        uint16 minor;
+        uint32 toc_offset;
+        char unknown[4];
+        uint32 toc_entries;
+        uint32 unknowns_2[3];
+        char release[256];
+        char brand[256];
+        char device_id[256];
+        char unknown_3[9];
+        char version[256];
+        char revision[256];
+    } bdl_header_t;
+"""
+
+
+def is_valid_header(header: Instance) -> bool:
+    if header.toc_offset == 0 or header.toc_entries == 0:
+        return False
+    try:
+        snull(header.release).decode("utf-8")
+        snull(header.brand).decode("utf-8")
+        snull(header.device_id).decode("utf-8")
+        snull(header.version).decode("utf-8")
+        snull(header.revision).decode("utf-8")
+    except UnicodeDecodeError:
+        return False
+    return True
+
+
+class HPBDLExtractor(Extractor):
+    def __init__(self):
+        self._struct_parser = StructParser(C_DEFINITIONS)
+
+    def extract(self, inpath: Path, outdir: Path):
+        entries = []
+        with File.from_path(inpath) as file:
+            header = self._struct_parser.parse("bdl_header_t", file, Endian.LITTLE)
+            file.seek(header.toc_offset, io.SEEK_SET)
+            for i in range(header.toc_entries):
+                entry = self._struct_parser.parse(
+                    "bdl_toc_entry_t", file, Endian.LITTLE
+                )
+                entries.append(
+                    (
+                        outdir.joinpath(outdir.joinpath(Path(f"ipkg{i:03}"))),
+                        Chunk(
+                            start_offset=entry.offset,
+                            end_offset=entry.offset + entry.size,
+                        ),
+                    )
+                )
+
+            for carve_path, chunk in entries:
+                carve_chunk_to_file(
+                    file=file,
+                    chunk=chunk,
+                    carve_path=carve_path,
+                )
+
+
+class HPBDLHandler(StructHandler):
+    NAME = "bdl"
+
+    PATTERNS = [HexString("69 62 64 6C 01 00 01 00")]
+
+    C_DEFINITIONS = C_DEFINITIONS
+    HEADER_STRUCT = "bdl_header_t"
+    EXTRACTOR = HPBDLExtractor()
+
+    def calculate_chunk(self, file: File, start_offset: int) -> Optional[ValidChunk]:
+        header = self.parse_header(file, endian=Endian.LITTLE)
+
+        if not is_valid_header(header):
+            raise InvalidInputFormat("Invalid BDL header.")
+
+        file.seek(start_offset + header.toc_offset, io.SEEK_SET)
+        end_offset = -1
+        for _ in range(header.toc_entries):
+            entry = self._struct_parser.parse("bdl_toc_entry_t", file, Endian.LITTLE)
+            end_offset = max(end_offset, start_offset + entry.offset + entry.size)
+
+        return ValidChunk(start_offset=start_offset, end_offset=end_offset)