Add SWU extraction

sig/scripts/build-swu.sh

@@ -0,0 +1,131 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+if [[ $# -ne 2 ]]; then
+    printf 'Usage: %s input-dir output.swu\n' "$0" >&2
+    exit 1
+fi
+
+indir=$1
+output=$2
+
+python3 - "$indir" "$output" <<'PY'
+import json
+import hashlib
+import pathlib
+import stat
+import sys
+
+
+HEADER_LEN = 110
+MANIFEST = ".swu-manifest.json"
+
+
+def align4_len(value):
+    return (-value) % 4
+
+
+def fail(message):
+    print(message, file=sys.stderr)
+    sys.exit(1)
+
+
+def safe_path(root, name):
+    if name.startswith("/"):
+        fail(f"absolute archive path is not supported: {name}")
+    path = (root / name).resolve()
+    root_resolved = root.resolve()
+    if path != root_resolved and root_resolved not in path.parents:
+        fail(f"archive path escapes input directory: {name}")
+    return path
+
+
+def checksum(payload):
+    return sum(payload) & 0xFFFFFFFF
+
+
+def header_for(item, payload):
+    name_bytes = item["name"].encode("utf-8") + b"\0"
+    values = [
+        item["ino"],
+        item["mode"],
+        item["uid"],
+        item["gid"],
+        item["nlink"],
+        item["mtime"],
+        len(payload),
+        item["devmajor"],
+        item["devminor"],
+        item["rdevmajor"],
+        item["rdevminor"],
+        len(name_bytes),
+        checksum(payload),
+    ]
+    return b"070702" + b"".join(f"{value & 0xFFFFFFFF:08X}".encode("ascii") for value in values)
+
+
+def append_entry(out, item, payload):
+    name_bytes = item["name"].encode("utf-8") + b"\0"
+    out.extend(header_for(item, payload))
+    if len(out) % 4 != 2:  # 110-byte newc/crc header always ends at +2 mod 4.
+        fail("internal cpio alignment error before filename")
+    out.extend(name_bytes)
+    out.extend(b"\0" * align4_len(len(out)))
+    out.extend(payload)
+    out.extend(b"\0" * align4_len(len(out)))
+
+
+if len(sys.argv) != 3:
+    fail("Usage: build-swu.sh input-dir output.swu")
+
+root = pathlib.Path(sys.argv[1])
+output = pathlib.Path(sys.argv[2])
+manifest_path = root / MANIFEST
+
+if not root.is_dir():
+    fail(f"input directory does not exist: {root}")
+if not manifest_path.is_file():
+    fail(f"missing manifest: {manifest_path}")
+
+manifest = json.loads(manifest_path.read_text())
+if manifest.get("format") != "svr4-crc-cpio":
+    fail("manifest was not created from a CRC SWU archive")
+
+entry_names = [item["name"] for item in manifest["entries"]]
+if "cpio_item_md5" in entry_names:
+    lines = []
+    for item in manifest["entries"]:
+        name = item["name"]
+        if name == "cpio_item_md5":
+            break
+        path = safe_path(root, name)
+        if stat.S_IFMT(item["mode"]) == stat.S_IFREG:
+            lines.append(f"{hashlib.md5(path.read_bytes()).hexdigest()}  {name}\n")
+    safe_path(root, "cpio_item_md5").write_text("".join(lines))
+
+out = bytearray()
+for item in manifest["entries"]:
+    path = safe_path(root, item["name"])
+    file_type = stat.S_IFMT(item["mode"])
+    if file_type == stat.S_IFDIR:
+        if not path.is_dir():
+            fail(f"missing directory entry: {item['name']}")
+        payload = b""
+    elif file_type == stat.S_IFREG:
+        if not path.is_file():
+            fail(f"missing file entry: {item['name']}")
+        payload = path.read_bytes()
+    else:
+        fail(f"unsupported cpio entry type for {item['name']}: mode {item['mode']:o}")
+    append_entry(out, item, payload)
+
+trailer = manifest["trailer"]
+if trailer.get("name") != "TRAILER!!!":
+    fail("manifest trailer entry is invalid")
+append_entry(out, trailer, b"")
+out.extend(b"\0" * int(manifest.get("final_padding", 0)))
+
+output.parent.mkdir(parents=True, exist_ok=True)
+output.write_bytes(out)
+print(f"wrote {output}")
+PY

sig/scripts/extract-swu.sh

@@ -0,0 +1,153 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+if [[ $# -ne 2 ]]; then
+    printf 'Usage: %s input.swu output-dir\n' "$0" >&2
+    exit 1
+fi
+
+input=$1
+outdir=$2
+
+python3 - "$input" "$outdir" <<'PY'
+import json
+import os
+import pathlib
+import stat
+import sys
+
+
+HEADER_LEN = 110
+MANIFEST = ".swu-manifest.json"
+
+
+def align4(value):
+    return (value + 3) & ~3
+
+
+def fail(message):
+    print(message, file=sys.stderr)
+    sys.exit(1)
+
+
+def safe_path(root, name):
+    if name.startswith("/"):
+        fail(f"absolute archive path is not supported: {name}")
+    path = (root / name).resolve()
+    root_resolved = root.resolve()
+    if path != root_resolved and root_resolved not in path.parents:
+        fail(f"archive path escapes output directory: {name}")
+    return path
+
+
+if len(sys.argv) != 3:
+    fail("Usage: extract-swu.sh input.swu output-dir")
+
+input_path = pathlib.Path(sys.argv[1])
+out_root = pathlib.Path(sys.argv[2])
+
+if not input_path.is_file():
+    fail(f"input does not exist: {input_path}")
+
+if out_root.exists() and any(out_root.iterdir()):
+    fail(f"output directory is not empty: {out_root}")
+out_root.mkdir(parents=True, exist_ok=True)
+
+data = input_path.read_bytes()
+offset = 0
+entries = []
+trailer = None
+
+while True:
+    if offset + HEADER_LEN > len(data):
+        fail("truncated cpio header")
+
+    header_offset = offset
+    header = data[offset:offset + HEADER_LEN]
+    offset += HEADER_LEN
+    fields = [header[i:i + 8].decode("ascii") for i in range(6, HEADER_LEN, 8)]
+    magic = header[:6].decode("ascii")
+    if magic != "070702":
+        fail(f"unsupported cpio magic at offset {header_offset}: {magic!r}")
+
+    values = [int(field, 16) for field in fields]
+    (
+        ino,
+        mode,
+        uid,
+        gid,
+        nlink,
+        mtime,
+        filesize,
+        devmajor,
+        devminor,
+        rdevmajor,
+        rdevminor,
+        namesize,
+        check,
+    ) = values
+
+    if offset + namesize > len(data):
+        fail("truncated cpio filename")
+    raw_name = data[offset:offset + namesize]
+    offset += namesize
+    if not raw_name.endswith(b"\0"):
+        fail("cpio filename is missing NUL terminator")
+    name = raw_name[:-1].decode("utf-8")
+    offset = align4(offset)
+
+    if offset + filesize > len(data):
+        fail(f"truncated cpio payload for {name}")
+    payload = data[offset:offset + filesize]
+    offset += filesize
+    offset = align4(offset)
+
+    item = {
+        "name": name,
+        "ino": ino,
+        "mode": mode,
+        "uid": uid,
+        "gid": gid,
+        "nlink": nlink,
+        "mtime": mtime,
+        "devmajor": devmajor,
+        "devminor": devminor,
+        "rdevmajor": rdevmajor,
+        "rdevminor": rdevminor,
+    }
+
+    if name == "TRAILER!!!":
+        trailer = item
+        break
+
+    entries.append(item)
+    path = safe_path(out_root, name)
+    file_type = stat.S_IFMT(mode)
+    if file_type == stat.S_IFDIR:
+        path.mkdir(parents=True, exist_ok=True)
+    elif file_type == stat.S_IFREG:
+        path.parent.mkdir(parents=True, exist_ok=True)
+        path.write_bytes(payload)
+    else:
+        fail(f"unsupported cpio entry type for {name}: mode {mode:o}")
+
+    os.chmod(path, stat.S_IMODE(mode))
+    os.utime(path, (mtime, mtime), follow_symlinks=False)
+
+if trailer is None:
+    fail("archive has no TRAILER!!! entry")
+
+final_padding = len(data) - offset
+if data[offset:] != b"\0" * final_padding:
+    fail("archive has non-zero bytes after TRAILER!!!")
+
+manifest = {
+    "format": "svr4-crc-cpio",
+    "block_size": 512,
+    "final_padding": final_padding,
+    "entries": entries,
+    "trailer": trailer,
+}
+(out_root / MANIFEST).write_text(json.dumps(manifest, indent=2) + "\n")
+print(f"extracted {len(entries)} entries to {out_root}")
+PY