core: enable using FI against projects without harnesses

src/fuzz_introspector/analysis.py

@@ -101,7 +101,8 @@ def load_data_files(self, parallelise=True, correlation_file=None):
         self.profiles = new_profiles
 
         logger.info("[+] Creating project profile")
-        self.proj_profile = project_profile.MergedProjectProfile(self.profiles)
+        self.proj_profile = project_profile.MergedProjectProfile(
+            self.profiles, self.language)
         self.proj_profile.coverage_url = self.coverage_url
 
         logger.info("[+] Refining profiles")

src/fuzz_introspector/data_loader.py

@@ -69,7 +69,7 @@ def read_fuzzer_data_file_to_profile(
 
     if not profile.has_entry_point():
         logger.info("Found no entrypoints")
-        return None
+
     logger.info("Returning profile")
     return profile
 
@@ -80,10 +80,10 @@ def _load_profile(data_file: str, language: str, manager, semaphore=None):
         semaphore.acquire()
 
     profile = read_fuzzer_data_file_to_profile(data_file, language)
-    logger.info('profile is none')
     if profile is not None:
         manager[data_file] = profile
-
+    else:
+        logger.error('profile is none')
     if semaphore is not None:
         semaphore.release()
 

src/fuzz_introspector/datatypes/fuzzer_profile.py

@@ -56,6 +56,8 @@ def __init__(self,
         self._target_lang = target_lang
         self.introspector_data_file = cfg_file
 
+        self.functions_reached_by_fuzzer: List[str] = []
+
         # Load calltree file
         self.fuzzer_callsite_calltree = cfg_load.data_file_read_calltree(
             cfg_file)
@@ -494,8 +496,6 @@ def _set_all_reached_functions(self) -> None:
                 self.functions_reached_by_fuzzer.append(entrypoint)
                 return
 
-        raise DataLoaderError("Can not identify entrypoint")
-
     def _set_all_unreached_functions(self) -> None:
         """Sets self.functions_unreached_by_fuzzer to all functions that are
         statically unreached. This is computed as the set difference between

src/fuzz_introspector/datatypes/project_profile.py

@@ -37,7 +37,8 @@ class MergedProjectProfile:
     digesting data from all the fuzzers in the project.
     """
 
-    def __init__(self, profiles: List[fuzzer_profile.FuzzerProfile]):
+    def __init__(self, profiles: List[fuzzer_profile.FuzzerProfile],
+                 language: str):
         self.name = None
         self.profiles = profiles
         self.all_functions: Dict[str,
@@ -49,6 +50,7 @@ def __init__(self, profiles: List[fuzzer_profile.FuzzerProfile]):
         self.coverage_url = "#"
         self.dst_to_fd_cache: Dict[str,
                                    function_profile.FunctionProfile] = dict()
+        self.language = language
 
         logger.info(
             f"Creating merged profile of {len(self.profiles)} profiles")
@@ -230,6 +232,8 @@ def target_lang(self):
         if len(set_of_targets) > 1:
             raise exceptions.AnalysisError(
                 "Project has fuzzers with multiple targets")
+        if not set_of_targets:
+            return self.language
         return set_of_targets.pop()
 
     @property

src/fuzz_introspector/frontends/oss_fuzz.py

@@ -66,6 +66,16 @@ def process_c_project(target_dir, entrypoint, out, module_only=False):
     logger.info('Creating base project.')
     project = frontend_c.Project(source_codes)
 
+    # We may not need to do this, but will do it while refactoring into
+    # the new frontends.
+    if not project.get_source_codes_with_harnesses():
+        target = os.path.join(out, 'fuzzerLogFile-0.data.yaml')
+        project.dump_module_logic(target, 'no-harness-in-project', target_dir)
+
+        with open(os.path.join(out, 'fuzzerLogFile-0.data'), 'w') as f:
+            f.write("Call tree\n")
+            f.write("====================================")
+
     if module_only:
         idx = 1
         target = os.path.join(out, 'report.yaml')