core: find where functions are declared in header files (#1624)

core: find where functions are declared in headfiles

Signed-off-by: David Korczynski <david@adalogics.com>

src/fuzz_introspector/analysis.py

@@ -909,11 +909,17 @@ def correlate_introspector_func_to_debug_information(if_func,
 
 def correlate_introspection_functions_to_debug_info(all_functions_json_report,
                                                     debug_all_functions,
-                                                    proj_lang):
+                                                    proj_lang,
+                                                    report_dict=dict()):
     """Correlates function data collected by debug information to function
     data collected by LLVMs module, and uses the correlated data to generate
     function signatures for each function based on debug information."""
 
+    # Find header files
+    normalized_paths = set()
+    for header_file in report_dict.get('all_files_in_project', []):
+        normalized_paths.add(os.path.normpath(header_file['source_file']))
+
     # A lot of look-ups are needed when matching LLVM functions to debug
     # functions. Start with creating two indexes to make these look-ups
     # faster.
@@ -924,6 +930,21 @@ def correlate_introspection_functions_to_debug_info(all_functions_json_report,
         df['source']['source_file'] = os.path.normpath(df['source'].get(
             'source_file', ''))
 
+        # Find the header file of this debug function.
+        possible_header_files = set()
+        for header_src_file in normalized_paths:
+            if not header_src_file.endswith(".h"):
+                continue
+            if not os.path.isfile(header_src_file):
+                continue
+            with open(header_src_file, 'r') as header_file_fd:
+                content = header_file_fd.read()
+            name = df.get('name', 'TOTALLYRANDOMNOTFUNCNAME123')
+            for line_idx, line in enumerate(content.split("\n")):
+                if f'{name}(' in line:
+                    possible_header_files.add(header_src_file)
+        df['possible-header-files'] = list(possible_header_files)
+
         # Append debug function to name-index.
         entry_list1 = debug_dict_by_name.get(df.get('name', ''), [])
         entry_list1.append(df)
@@ -939,6 +960,7 @@ def correlate_introspection_functions_to_debug_info(all_functions_json_report,
     for dl3 in debug_dict_by_filename:
         print("%s ------- %d" % (dl3, len(debug_dict_by_filename[dl3])))
 
+    # Now correlate signatures
     for if_func in all_functions_json_report:
         func_sig, correlated_debug_function = correlate_introspector_func_to_debug_information(
             if_func, debug_all_functions, debug_dict_by_name,

src/fuzz_introspector/debug_info.py

@@ -305,6 +305,7 @@ def load_debug_report(debug_files):
         'all_global_variables': list(all_global_variables.values()),
         'all_types': list(all_types.values())
     }
+
     return report_dict
 
 

src/fuzz_introspector/html_report.py

@@ -763,7 +763,7 @@ def create_html_report(introspection_proj: analysis.IntrospectionProject,
     # Correlate debug info to introspector functions
     analysis.correlate_introspection_functions_to_debug_info(
         all_functions_json_report, introspection_proj.debug_all_functions,
-        proj_profile.target_lang)
+        proj_profile.target_lang, introspection_proj.debug_report)
 
     # Write various stats and all-functions data to summary.json
     proj_profile.write_stats_to_summary_file()