Require ELF closure verification when staging Linux speech runtime

tools/stage_speech_worker.py

@@ -1,6 +1,7 @@
 """Verify and atomically stage the isolated normal speech runtime; no inference."""
 import argparse
 import hashlib
+import importlib.util
 import json
 from pathlib import Path
 import re
@@ -28,6 +29,15 @@ def input_path(root, name):
     return Path(root) / 'tools' / (name if '/' in name else 'speech-runtime/' + name)
 
 
+def verify_linux_closure(directory):
+    # Load our pinned builder, never code from the candidate runtime directory.
+    path = ROOT / 'tools/speech-runtime/build.py'
+    spec = importlib.util.spec_from_file_location('searchi_speech_build', path)
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    module.verify_linux_closure(directory)
+
+
 def inventory(directory):
     path = Path(directory) / 'runtime-manifest.json'
     if path.is_symlink() or not path.is_file() or not 0 < path.stat().st_size <= 512 * 1024:
@@ -84,6 +94,8 @@ def verify_directory(directory, root=ROOT, platform='windows', architecture='x86
                          'VULKAN-LICENSE.txt', 'MSVC-REDISTRIBUTION.txt'} | CRT_LIBRARIES
     if platform == 'windows' and not windows_libraries.issubset(names):
         raise ValueError('Speech runtime lacks required shared libraries')
+    if platform == 'linux':
+        verify_linux_closure(directory)
     records = {'speech-runtime/' + record['path']: record for record in manifest['files']}
     manifest_path = directory / 'runtime-manifest.json'
     records['speech-runtime/runtime-manifest.json'] = {

tools/test_stage_speech_worker.py

@@ -3,12 +3,71 @@ from pathlib import Path
 import shutil
 import tempfile
 import unittest
+from unittest import mock
 
 import stage_speech_worker as speech
 import test_package_windows as windows
 
 
 class SpeechStagingTests(unittest.TestCase):
+    def linux_fixture(self, directory):
+        root, resources, _, _ = windows.PackageTests().fixture(directory)
+        runtime = resources / 'speech-runtime'
+        manifest_path = runtime / 'runtime-manifest.json'
+        manifest = json.loads(manifest_path.read_text())
+        for record in manifest['files']:
+            (runtime / record['path']).unlink()
+        manifest['files'] = []
+        for name in ('searchi-speech-worker', 'libggml-cpu-x64.so', 'libggml-vulkan.so'):
+            path = runtime / name
+            path.write_bytes(b'ELF fixture inspected by mocked readelf')
+            manifest['files'].append({'path': name, 'bytes': path.stat().st_size,
+                                      'sha256': speech.digest(path)})
+        manifest_path.write_text(json.dumps(manifest))
+        return root, runtime
+
+    def test_linux_hash_valid_inventory_requires_elf_closure(self):
+        with tempfile.TemporaryDirectory() as directory:
+            root, runtime = self.linux_fixture(directory)
+            with mock.patch.object(speech, 'identity', return_value=('linux', 'x86_64', None)), \
+                    mock.patch.object(speech, 'verify_linux_closure') as closure:
+                records = speech.verify_directory(runtime, root, platform='linux')
+                closure.assert_called_once_with(runtime)
+                self.assertIn('speech-runtime/searchi-speech-worker', records)
+
+    def test_linux_closure_failure_and_missing_readelf_are_not_bypassed(self):
+        for error in (ValueError('Unpackaged Linux speech dependency'),
+                      FileNotFoundError('readelf is required')):
+            with self.subTest(error=type(error)), tempfile.TemporaryDirectory() as directory:
+                root, runtime = self.linux_fixture(directory)
+                with mock.patch.object(speech, 'identity', return_value=('linux', 'x86_64', None)), \
+                        mock.patch.object(speech, 'verify_linux_closure', side_effect=error):
+                    with self.assertRaises(type(error)):
+                        speech.verify_directory(runtime, root, platform='linux')
+
+    def test_windows_does_not_require_elf_tools(self):
+        with tempfile.TemporaryDirectory() as directory:
+            root, resources, _, _ = windows.PackageTests().fixture(directory)
+            with mock.patch.object(speech, 'verify_linux_closure') as closure:
+                speech.verify_directory(resources / 'speech-runtime', root)
+                closure.assert_not_called()
+
+    def test_loaded_builder_checks_every_linux_elf_and_rejects_build_tree_runpath(self):
+        with tempfile.TemporaryDirectory() as directory:
+            _, runtime = self.linux_fixture(directory)
+            valid = '(RUNPATH) Library runpath: [$ORIGIN]\n(NEEDED) Shared library: [libc.so.6]\n'
+            with mock.patch('subprocess.check_output', return_value=valid) as readelf:
+                speech.verify_linux_closure(runtime)
+                self.assertEqual(readelf.call_count, 3)
+                self.assertEqual({call.args[0][2] for call in readelf.call_args_list},
+                                 {str(runtime / name) for name in
+                                  ('searchi-speech-worker', 'libggml-cpu-x64.so', 'libggml-vulkan.so')})
+                self.assertTrue(all(call.args[0][:2] == ['readelf', '--dynamic']
+                                    for call in readelf.call_args_list))
+            with mock.patch('subprocess.check_output', return_value=valid.replace('$ORIGIN', '/build/whisper')):
+                with self.assertRaisesRegex(ValueError, 'origin-only RUNPATH'):
+                    speech.verify_linux_closure(runtime)
+
     def test_replacement_prunes_obsolete_declared_code_and_keeps_closed_inventory(self):
         with tempfile.TemporaryDirectory() as directory:
             root, resources, _, _ = windows.PackageTests().fixture(directory)