Like a port program, the port communicates with an Erlang process. All communication goes through one Erlang process that is the connected process of the port driver. Terminating this process closes the port driver.

Before the port is created, the driver must be loaded. This is done with the function erl_dll:load_driver/1, with the name of the shared library as argument.

The port is then created using the BIF open_port/2, with the tuple {spawn, DriverName} as the first argument. The string SharedLib is the name of the port driver. The second argument is a list of options, none in this case:

-module(complex5).
-export([start/1, init/1]).

start(SharedLib) ->
    case erl_ddll:load_driver(".", SharedLib) of
        ok -> ok;
        {error, already_loaded} -> ok;
        _ -> exit({error, could_not_load_driver})
    end,
    spawn(?MODULE, init, [SharedLib]).

init(SharedLib) ->
  register(complex, self()),
  Port = open_port({spawn, SharedLib}, []),
  loop(Port).

Now complex5:foo/1 and complex5:bar/1 can be implemented. Both send a message to the complex process and receive the following reply:

foo(X) ->
    call_port({foo, X}).
bar(Y) ->
    call_port({bar, Y}).

call_port(Msg) ->
    complex ! {call, self(), Msg},
    receive
        {complex, Result} ->
            Result
    end.

The complex process performs the following:

loop(Port) ->
    receive
        {call, Caller, Msg} ->
            Port ! {self(), {command, encode(Msg)}},
            receive
                {Port, {data, Data}} ->
                    Caller ! {complex, decode(Data)}
            end,
            loop(Port)
    end.

Assuming that both the arguments and the results from the C functions are less than 256, a simple encoding/decoding scheme is employed. In this scheme, foo is represented by byte 1, bar is represented by 2, and the argument/result is represented by a single byte as well:

encode({foo, X}) -> [1, X];
encode({bar, Y}) -> [2, Y].

decode([Int]) -> Int.

The resulting Erlang program, including functions for stopping the port and detecting port failures, is as follows:

The C driver is a module that is compiled and linked into a shared library. It uses a driver structure and includes the header file erl_driver.h.

The driver structure is filled with the driver name and function pointers. It is returned from the special entry point, declared with the macro )]]>.

The functions for receiving and sending data are combined into a function, pointed out by the driver structure. The data sent into the port is given as arguments, and the replied data is sent with the C-function driver_output.

As the driver is a shared module, not a program, no main function is present. All function pointers are not used in this example, and the corresponding fields in the driver_entry structure are set to NULL.

All functions in the driver takes a handle (returned from start) that is just passed along by the Erlang process. This must in some way refer to the port driver instance.

The example_drv_start, is the only function that is called with a handle to the port instance, so this must be saved. It is customary to use an allocated driver-defined structure for this one, and to pass a pointer back as a reference.

It is not a good idea to use a global variable as the port driver can be spawned by multiple Erlang processes. This driver-structure is to be instantiated multiple times:

Step 1. Compile the C code:

unix> gcc -o example_drv.so -fpic -shared complex.c port_driver.c
windows> cl -LD -MD -Fe example_drv.dll complex.c port_driver.c

Step 2. Start Erlang and compile the Erlang code:

> erl
Erlang (BEAM) emulator version 5.1

Eshell V5.1 (abort with ^G)
1> c(complex5).
{ok,complex5}

Step 3. Run the example:

2> complex5:start("example_drv").
<0.34.0>
3> complex5:foo(3).
4
4> complex5:bar(5).
10
5> complex5:stop().
stop