17.7 Compiling and injecting code in gdb

gdb supports on-demand compilation and code injection into programs running under gdb. GCC 5.0 or higher built with libcc1.so must be installed for this functionality to be enabled. This functionality is implemented with the following commands.

compile code source-code

compile code -raw – source-code

Compile source-code with the compiler language found as the current language in gdb (see Languages). If compilation and injection is not supported with the current language specified in gdb, or the compiler does not support this feature, an error message will be printed. If source-code compiles and links successfully, gdb will load the object-code emitted, and execute it within the context of the currently selected inferior. It is important to note that the compiled code is executed immediately. After execution, the compiled code is removed from gdb and any new types or variables you have defined will be deleted.

The command allows you to specify source-code in two ways. The simplest method is to provide a single line of code to the command. E.g.:

          compile code printf ("hello world\n");

If you specify options on the command line as well as source code, they may conflict. The ‘--’ delimiter can be used to separate options from actual source code. E.g.:

          compile code -r -- printf ("hello world\n");

Alternatively you can enter source code as multiple lines of text. To enter this mode, invoke the ‘compile code’ command without any text following the command. This will start the multiple-line editor and allow you to type as many lines of source code as required. When you have completed typing, enter ‘end’ on its own line to exit the editor.

          compile code
          >printf ("hello\n");
          >printf ("world\n");
          >end

Specifying ‘-raw’, prohibits gdb from wrapping the provided source-code in a callable scope. In this case, you must specify the entry point of the code by defining a function named _gdb_expr_. The ‘-raw’ code cannot access variables of the inferior. Using ‘-raw’ option may be needed for example when source-code requires ‘#include’ lines which may conflict with inferior symbols otherwise.

compile file filename

compile file -raw filename

Like compile code, but take the source code from filename.

          compile file /home/user/example.c

compile print expr

compile print /f expr

Compile and execute expr with the compiler language found as the current language in gdb (see Languages). By default the value of expr is printed in a format appropriate to its data type; you can choose a different format by specifying ‘/f’, where f is a letter specifying the format; see Output Formats.

compile print

compile print /f

Alternatively you can enter the expression (source code producing it) as multiple lines of text. To enter this mode, invoke the ‘compile print’ command without any text following the command. This will start the multiple-line editor.

The process of compiling and injecting the code can be inspected using:

set debug compile

Turns on or off display of gdb process of compiling and injecting the code. The default is off.

show debug compile

Displays the current state of displaying gdb process of compiling and injecting the code.

17.7.1 Compilation options for the compile command

gdb needs to specify the right compilation options for the code to be injected, in part to make its ABI compatible with the inferior and in part to make the injected code compatible with gdb's injecting process.

The options used, in increasing precedence:

target architecture and OS options (gdbarch)

These options depend on target processor type and target operating system, usually they specify at least 32-bit (-m32) or 64-bit (-m64) compilation option.

compilation options recorded in the target

gcc (since version 4.7) stores the options used for compilation into DW_AT_producer part of DWARF debugging information according to the gcc option -grecord-gcc-switches. One has to explicitly specify -g during inferior compilation otherwise gcc produces no DWARF. This feature is only relevant for platforms where -g produces DWARF by default, otherwise one may try to enforce DWARF by using -gdwarf-4.

compilation options set by set compile-args

You can override compilation options using the following command:

set compile-args

Set compilation options used for compiling and injecting code with the compile commands. These options override any conflicting ones from the target architecture and/or options stored during inferior compilation.

show compile-args

Displays the current state of compilation options override. This does not show all the options actually used during compilation, use set debug compile for that.

17.7.2 Caveats when using the compile command

There are a few caveats to keep in mind when using the compile command. As the caveats are different per language, the table below highlights specific issues on a per language basis.

C code examples and caveats

When the language in gdb is set to ‘C’, the compiler will attempt to compile the source code with a ‘C’ compiler. The source code provided to the compile command will have much the same access to variables and types as it normally would if it were part of the program currently being debugged in gdb.

Below is a sample program that forms the basis of the examples that follow. This program has been compiled and loaded into gdb, much like any other normal debugging session.

          void function1 (void)
          {
             int i = 42;
             printf ("function 1\n");
          }
          
          void function2 (void)
          {
             int j = 12;
             function1 ();
          }
          
          int main(void)
          {
             int k = 6;
             int *p;
             function2 ();
             return 0;
          }

For the purposes of the examples in this section, the program above has been compiled, loaded into gdb, stopped at the function main, and gdb is awaiting input from the user.

To access variables and types for any program in gdb, the program must be compiled and packaged with debug information. The compile command is not an exception to this rule. Without debug information, you can still use the compile command, but you will be very limited in what variables and types you can access.

So with that in mind, the example above has been compiled with debug information enabled. The compile command will have access to all variables and types (except those that may have been optimized out). Currently, as gdb has stopped the program in the main function, the compile command would have access to the variable k. You could invoke the compile command and type some source code to set the value of k. You can also read it, or do anything with that variable you would normally do in C. Be aware that changes to inferior variables in the compile command are persistent. In the following example:

          compile code k = 3;

the variable k is now 3. It will retain that value until something else in the example program changes it, or another compile command changes it.

Normal scope and access rules apply to source code compiled and injected by the compile command. In the example, the variables j and k are not accessible yet, because the program is currently stopped in the main function, where these variables are not in scope. Therefore, the following command

          compile code j = 3;

will result in a compilation error message.

Once the program is continued, execution will bring these variables in scope, and they will become accessible; then the code you specify via the compile command will be able to access them.

You can create variables and types with the compile command as part of your source code. Variables and types that are created as part of the compile command are not visible to the rest of the program for the duration of its run. This example is valid:

          compile code int ff = 5; printf ("ff is %d\n", ff);

However, if you were to type the following into gdb after that command has completed:

          compile code printf ("ff is %d\n'', ff);

a compiler error would be raised as the variable ff no longer exists. Object code generated and injected by the compile command is removed when its execution ends. Caution is advised when assigning to program variables values of variables created by the code submitted to the compile command. This example is valid:

          compile code int ff = 5; k = ff;

The value of the variable ff is assigned to k. The variable k does not require the existence of ff to maintain the value it has been assigned. However, pointers require particular care in assignment. If the source code compiled with the compile command changed the address of a pointer in the example program, perhaps to a variable created in the compile command, that pointer would point to an invalid location when the command exits. The following example would likely cause issues with your debugged program:

          compile code int ff = 5; p = &ff;

In this example, p would point to ff when the compile command is executing the source code provided to it. However, as variables in the (example) program persist with their assigned values, the variable p would point to an invalid location when the command exists. A general rule should be followed in that you should either assign NULL to any assigned pointers, or restore a valid location to the pointer before the command exits.

Similar caution must be exercised with any structs, unions, and typedefs defined in compile command. Types defined in the compile command will no longer be available in the next compile command. Therefore, if you cast a variable to a type defined in the compile command, care must be taken to ensure that any future need to resolve the type can be achieved.

          (gdb) compile code static struct a { int a; } v = { 42 }; argv = &v;
          (gdb) compile code printf ("%d\n", ((struct a *) argv)->a);
          gdb command line:1:36: error: dereferencing pointer to incomplete type ‘struct a’
          Compilation failed.
          (gdb) compile code struct a { int a; }; printf ("%d\n", ((struct a *) argv)->a);
          42

Variables that have been optimized away by the compiler are not accessible to the code submitted to the compile command. Access to those variables will generate a compiler error which gdb will print to the console.

17.7.3 Compiler search for the compile command

gdb needs to find gcc for the inferior being debugged which may not be obvious for remote targets of different architecture than where gdb is running. Environment variable PATH (PATH from shell that executed gdb, not the one set by gdb command set environment). See Environment. PATH on gdb host is searched for gcc binary matching the target architecture and operating system.

Specifically PATH is searched for binaries matching regular expression arch(-[^-]*)?-os-gcc according to the inferior target being debugged. arch is processor name — multiarch is supported, so for example both i386 and x86_64 targets look for pattern (x86_64|i.86) and both s390 and s390x targets look for pattern s390x?. os is currently supported only for pattern linux(-gnu)?.