fix: improved efficiency of docker mount in software plugin

fkie-cad · Nov 27, 2024 · 00dc1f3 · 00dc1f3
1 parent ea694e5
commit 00dc1f3
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 8 deletions.
diff --git a/src/plugins/analysis/software_components/code/software_components.py b/src/plugins/analysis/software_components/code/software_components.py
@@ -87,7 +87,7 @@ def get_version_for_component(self, result, file_object: FileObject):
                     'function_name': result['meta']['_version_function'],
                 }
             versions.update(
-                extract_data_from_ghidra(file_object.binary, input_data, config.backend.docker_mount_base_dir)
+                extract_data_from_ghidra(file_object.file_path, input_data, config.backend.docker_mount_base_dir)
             )
         if '' in versions and len(versions) > 1:  # if there are actual version results, remove the "empty" result
             versions.remove('')

diff --git a/src/plugins/analysis/software_components/internal/resolve_version_format_string.py b/src/plugins/analysis/software_components/internal/resolve_version_format_string.py
@@ -12,27 +12,27 @@
 
 from helperFunctions.docker import run_docker_container
 
-CONTAINER_TARGET_PATH = '/work'
+CONTAINER_TARGET_PATH = Path('/work')
+INPUT_PATH = CONTAINER_TARGET_PATH / 'ghidra_input'
 DOCKER_IMAGE = 'fact/format_string_resolver'
 DOCKER_OUTPUT_FILE = 'ghidra_output.json'
 TIMEOUT = 300
 KEY_FILE = 'key_file'
 
 
-def extract_data_from_ghidra(input_file_data: bytes, input_data: dict, path: str) -> list[str]:
+def extract_data_from_ghidra(file_path: str, input_data: dict, path: str) -> list[str]:
     with TemporaryDirectory(prefix='FSR_', dir=path) as tmp_dir:
         tmp_dir_path = Path(tmp_dir)
-        ghidra_input_file = tmp_dir_path / 'ghidra_input'
         (tmp_dir_path / KEY_FILE).write_text(json.dumps(input_data))
-        ghidra_input_file.write_bytes(input_file_data)
         with suppress(DockerException, TimeoutError):
             proc = run_docker_container(
                 DOCKER_IMAGE,
                 logging_label='FSR',
                 timeout=TIMEOUT,
-                command=f'/work/ghidra_input {CONTAINER_TARGET_PATH}',
+                command=f'{INPUT_PATH} {CONTAINER_TARGET_PATH}',
                 mounts=[
-                    Mount(CONTAINER_TARGET_PATH, tmp_dir, type='bind'),
+                    Mount(str(CONTAINER_TARGET_PATH), tmp_dir, type='bind'),
+                    Mount(str(INPUT_PATH), file_path, type='bind', read_only=True),
                 ],
             )
             if 'Traceback' in proc.stderr:

diff --git a/src/plugins/analysis/software_components/test/test_resolve_version_format_string.py b/src/plugins/analysis/software_components/test/test_resolve_version_format_string.py
@@ -25,7 +25,7 @@
 )
 def test_extract_data_from_ghidra(backend_config, test_file, input_data, expected_output):
     test_file = Path(__file__).parent / 'data' / test_file
-    result = extract_data_from_ghidra(test_file.read_bytes(), input_data, str(backend_config.docker_mount_base_dir))
+    result = extract_data_from_ghidra(str(test_file), input_data, str(backend_config.docker_mount_base_dir))
     assert len(result) == 1
     assert result == expected_output