The Mono runtime can be embedded into C and C++ applications. Your C/C++ code can invoke managed code running in the Mono/.NET world and you can also surface your internal application APIs to Mono and .NET.
For an overview of how to embed Mono into your application and the strategies that you can use to embed Mono, check the Mono website's Embedding Mono page.
This page is the companion API reference for the above guide.
The simplest way of embedding Mono is illustrated here:
int main (int argc, char *argv) { /* * Load the default Mono configuration file, this is needed * if you are planning on using the dllmaps defined on the * system configuration */ mono_config_parse (NULL); /* * mono_jit_init() creates a domain: each assembly is * loaded and run in a MonoDomain. */ MonoDomain *domain = mono_jit_init ("startup.exe"); /* * Optionally, add an internal call that your startup.exe * code can call, this will bridge startup.exe to Mono */ mono_add_internal_call ("Sample::GetMessage", getMessage); /* * Open the executable, and run the Main method declared * in the executable */ MonoAssembly *assembly = mono_domain_assembly_open (domain, "startup.exe"); if (!assembly) exit (2); /* * mono_jit_exec() will run the Main() method in the assembly. * The return value needs to be looked up from * System.Environment.ExitCode. */ mono_jit_exec (domain, assembly, argc, argv); } /* The C# signature for this method is: string GetMessage () in class Sample */ MonoString* getMessage () { return mono_string_new (mono_domain_get (), "Hello, world"); }
mono_jit_init
mono_jit_parse_options
Process the command line options in argv as done by the runtime executable.
This should be called before mono_jit_init
.
mono_jit_exec
assembly | reference to an assembly |
argc | argument count |
argv | argument vector |
mono_set_dirs
assembly_dir | the base directory for assemblies |
config_dir | the base directory for configuration files |
This routine is used internally and by developers embedding the runtime into their own applications.
There are a number of cases to consider: Mono as a system-installed package that is available on the location preconfigured or Mono in a relocated location.
If you are using a system-installed Mono, you can pass NULL
to both parameters. If you are not, you should compute both
directory values and call this routine.
The values for a given PREFIX are:
assembly_dir: PREFIX/lib config_dir: PREFIX/etc
Notice that embedders that use Mono in a relocated way must compute the location at runtime, as they will be in control of where Mono is installed.
mono_parse_default_optimizations
mono_runtime_set_main_args
argc | number of arguments from the command line |
argv | array of strings from the command line |
mono_runtime_run_main
.mono_jit_cleanup
mono_jit_set_trace_options
options | string representing the trace options |
TRUE
if the options were parsed and set correctly, FALSE
otherwise.
The Mono runtime provides two mechanisms to expose C code to the CIL universe: internal calls and native C code. Internal calls are tightly integrated with the runtime, and have the least overhead, as they use the same data types that the runtime uses.
The other option is to use the Platform Invoke (P/Invoke) to call C code from the CIL universe, using the standard P/Invoke mechanisms.
To register an internal call, use this call you use the mono_add_internal_call routine.
mono_add_internal_call
name | method specification to surface to the managed world |
method | pointer to a C method to invoke when the method is called |
This method surfaces the C function pointed by method as a method that has been surfaced in managed code with the method specified in name as an internal call.
Internal calls are surfaced to all app domains loaded and they are accessibly by a type with the specified name.
You must provide a fully qualified type name, that is namespaces and type name, followed by a colon and the method name, with an optional signature to bind.
For example, the following are all valid declarations:
MyApp.Services.ScriptService:Accelerate
MyApp.Services.ScriptService:Slowdown(int,bool)
You use method parameters in cases where there might be more than one surface method to managed code. That way you can register different internal calls for different method overloads.
The internal calls are invoked with no marshalling. This means that .NET
types like System.String
are exposed as MonoString*
parameters. This is
different than the way that strings are surfaced in P/Invoke.
For more information on how the parameters are marshalled, see the Mono Embedding page.
See the Method Description reference for more information on the format of method descriptions.
Unlike internal calls, Platform/Invoke is easier to use and more portable. It allows you to share code with Windows and .NET that have a different setup for internal calls to their own runtime.
Usually P/Invoke declarations reference external libraries like:
[DllImport ("opengl")] void glBegin (GLEnum mode)
Mono extends P/Invoke to support looking up symbols not in an external library, but looking up those symbols into the same address space as your program, to do this, use the special library name "__Internal". This will direct Mono to lookup the method in your own process.
There are situations where the host operating system does not support looking up symbols on the process address space. For situations like this you can use the mono_dl_register_library.
Managed objects are represented as MonoObject* types. Those objects that the runtime consumes directly have more specific C definitions (for example strings are of type MonoString *, delegates are of type MonoDelegate* but they are still MonoObject *s).
As of Mono 1.2.x types defined in mscorlib.dll do not have their fields reordered in any way. But other libraries might have their fields reordered. In these cases, Managed structures and objects have the same layout in the C# code as they do in the unmanaged world.
Structures defined outside corlib must have a specific StructLayout definition, and have it set as sequential if you plan on accessing these fields directly from C code.
Important Internal calls do not provide support for marshalling structures. This means that any API calls that take a structure (excluding the system types like int32, int64, etc) must be passed as a pointer, in C# this means passing the value as a "ref" or "out" parameter.
Certain features of the Mono runtime, like DLL mapping, are available through a configuration file that is loaded at runtime. The default Mono implementation loads the configuration file from $sysconfig/mono/config (typically this is /etc/mono/config).
See the mono-config(5) man page for more details on what goes in this file.
The following APIs expose this functionality:
mono_config_cleanup
mono_config_is_server_mode
mono_config_parse
filename | the filename to load the configuration variables from. |
NULL
filename to parse the default config files
(or the file in the MONO_CONFIG
env var).mono_config_parse_memory
buffer | a pointer to an string XML representation of the configuration |
mono_config_set_server_mode
mono_config_string_for_assembly_file
mono_get_config_dir
mono_get_machine_config
mono_register_machine_config
mono_set_config_dir
These are not recommended ways of initializing Mono, they are done internally by mono_jit_init, but are here to explain what happens internally.
mono_runtime_exec_managed_code
domain | Application domain |
main_func | function to invoke from the execution thread |
main_args | parameter to the main_func |
main_func is called back from the thread with main_args as the
parameter. The callback function is expected to start Main
eventually. This function then waits for all managed threads to
finish.
It is not necessary anymore to execute managed code in a subthread,
so this function should not be used anymore by default: just
execute the code and then call mono_thread_manage().
mono_runtime_exec_main
mono_init
Creates the initial application domain and initializes the mono_defaults structure.
This function is guaranteed to not run any IL code. The runtime is initialized using the default runtime version.
mono_init_from_assembly
domain_name | name to give to the initial domain |
filename | filename to load on startup |
Used by the runtime, users should use mono_jit_init instead.
Creates the initial application domain and initializes the mono_defaults structure. This function is guaranteed to not run any IL code. The runtime is initialized using the runtime version required by the provided executable. The version is determined by looking at the exe configuration file and the version PE field)
mono_init_version
Used by the runtime, users should use mono_jit_init
instead.
Creates the initial application domain and initializes the mono_defaults
structure.
This function is guaranteed to not run any IL code. The runtime is initialized using the provided rutime version.
mono_jit_exec
assembly | reference to an assembly |
argc | argument count |
argv | argument vector |
mono_jit_set_aot_mode
mono_set_break_policy
policy_callback | the new callback function |
Allow embedders to decide whether to actually obey breakpoint instructions
(both break IL instructions and Debugger.Break
method calls), for example
to not allow an app to be aborted by a perfectly valid IL opcode when executing
untrusted or semi-trusted code.
policy_callback will be called every time a break point instruction needs to
be inserted with the method argument being the method that calls Debugger.Break
or has the IL break
instruction. The callback should return MONO_BREAK_POLICY_NEVER
if it wants the breakpoint to not be effective in the given method.
MONO_BREAK_POLICY_ALWAYS
is the default.
mono_get_runtime_build_info
VERSION (FULL_VERSION BUILD_DATE)
and build date is optional.mono_set_signal_chaining
Enable/disable signal chaining. This should be called before mono_jit_init
.
If signal chaining is enabled, the runtime saves the original signal handlers before
installing its own handlers, and calls the original ones in the following cases:
- a SIGSEGV
/ SIGABRT
signal received while executing native (i.e. not JITted) code.
- SIGPROF
- SIGFPE
- SIGQUIT
- SIGUSR2
Signal chaining only works on POSIX platforms.