wip pdb dump2config support

OmniBlade · Apr 9, 2024 · cb34531 · cb34531
1 parent 7bb83ba
commit cb34531
Show file tree

Hide file tree

Showing 4 changed files with 747 additions and 15 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -57,6 +57,13 @@ FetchContent_Declare(
 )
 FetchContent_MakeAvailable(json)
 
+FetchContent_Declare(
+    raw_pdb
+    GIT_REPOSITORY https://github.com/MolecularMatters/raw_pdb.git
+    GIT_TAG        5c29c02b897c884c7ead0b912b721346474e271e
+)
+FetchContent_MakeAvailable(raw_pdb)
+
 set(GIT_PRE_CONFIGURE_FILE "gitinfo.cpp.in")
 set(GIT_POST_CONFIGURE_FILE "${CMAKE_CURRENT_BINARY_DIR}/gitinfo.cpp")
 include(GitWatcher)
@@ -71,8 +78,10 @@ target_sources(unassemblize PRIVATE
     function.cpp
     function.h
     main.cpp
+    progdb.cpp
+    progdb.h
 )
-target_link_libraries(unassemblize PRIVATE Zydis LIEF::LIEF nlohmann_json)
+target_link_libraries(unassemblize PRIVATE Zydis LIEF::LIEF nlohmann_json raw_pdb)
 target_include_directories(unassemblize PRIVATE .)
 
 if(WINDOWS)

diff --git a/main.cpp b/main.cpp
@@ -11,6 +11,7 @@
  *            LICENSE
  */
 #include "function.h"
+#include "progdb.h"
 #include "gitinfo.h"
 #include <LIEF/LIEF.hpp>
 #include <getopt.h>
@@ -33,20 +34,22 @@ void print_help()
         "Usage:\n"
         "  unassemblize [OPTIONS] [INPUT]\n"
         "Options:\n"
-        "  -o --output     Filename for single file output. Default is program.S\n"
-        "  -f --format     Assembly output format.\n"
-        "  -c --config     Configuration file describing how to dissassemble the input\n"
-        "                  file and containing extra symbol info. Default: config.json\n"
-        "  -s --start      Starting address of a single function to dissassemble in\n"
-        "                  hexidecimal notation.\n"
-        "  -e --end        Ending address of a single function to dissassemble in\n"
-        "                  hexidecimal notation.\n"
-        "  -v --verbose    Verbose output on current state of the program.\n"
-        "  --section       Section to target for dissassembly, defaults to '.text'.\n"
-        "  --listsections  Prints a list of sections in the exe then exits.\n"
-        "  -d --dumpsyms   Dumps symbols stored in the executable to the config file.\n"
-        "                  then exits.\n"
-        "  -h --help       Displays this help.\n\n",
+        "  -o --output        Filename for single file output. Default is program.S\n"
+        "  -f --format        Assembly output format.\n"
+        "  -c --config        Configuration file describing how to dissassemble the input\n"
+        "                     file and containing extra symbol info. Default: config.json\n"
+        "  -s --start         Starting address of a single function to dissassemble in\n"
+        "                     hexidecimal notation.\n"
+        "  -e --end           Ending address of a single function to dissassemble in\n"
+        "                     hexidecimal notation.\n"
+        "  -v --verbose       Verbose output on current state of the program.\n"
+        "  --section          Section to target for dissassembly, defaults to '.text'.\n"
+        "  --listsections     Prints a list of sections in the exe then exits.\n"
+        "  -d --dumpsyms      Dumps symbols stored in the executable to the config file\n"
+        "                     then exits.\n"
+        "  -p --dumppdb       Dumps symbol data stored in a pdb file to a config file\n"
+        "                     then exits.\n"
+        "  -h --help          Displays this help.\n\n",
         revision,
         GitUncommittedChanges ? "~" : "",
         version);
@@ -70,11 +73,13 @@ int main(int argc, char **argv)
     const char *section_name = ".text";
     const char *output = "program.S";
     const char *config_file = "config.json";
+    const char *pdb_config_file = "pdb.json";
     const char *format_string = nullptr;
     uint64_t start_addr = 0;
     uint64_t end_addr = 0;
     bool print_secs = false;
     bool dump_syms = false;
+    bool dump_pdb = false;
     bool verbose = false;
 
     while (true) {
@@ -87,6 +92,7 @@ int main(int argc, char **argv)
             {"section", required_argument, nullptr, 1},
             {"listsections", no_argument, nullptr, 2},
             {"dumpsyms", no_argument, nullptr, 'd'},
+            {"dumppdb", no_argument, nullptr, 'p'},
             {"verbose", no_argument, nullptr, 'v'},
             {"help", no_argument, nullptr, 'h'},
             {nullptr, no_argument, nullptr, 0},
@@ -110,6 +116,9 @@ int main(int argc, char **argv)
             case 'd':
                 dump_syms = true;
                 break;
+            case 'p':
+                dump_pdb = true;
+                break;
             case 'o':
                 output = optarg;
                 break;
@@ -160,6 +169,7 @@ int main(int argc, char **argv)
     }
 
     unassemblize::Executable exe(argv[optind], format, verbose);
+    unassemblize::ProgramDatabase pdb(argv[optind], verbose);
 
     if (print_secs) {
         print_sections(exe);
@@ -171,6 +181,11 @@ int main(int argc, char **argv)
         return 0;
     }
 
+    if (dump_pdb) {
+        pdb.save_symbols_config(pdb_config_file);
+        return 0;
+    }
+
     exe.load_config(config_file);
 
     FILE *fp = nullptr;