20002013 Ericsson AB. All Rights Reserved. The contents of this file are subject to the Erlang Public License, Version 1.1, (the "License"); you may not use this file except in compliance with the License. You should have received a copy of the Erlang Public License along with this software. If not, it can be retrieved online at http://www.erlang.org/. Software distributed under the License is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License for the specific language governing rights and limitations under the License. NIFs nif.xml

This is an example of how to solve the example problem by using NIFs. NIFs where introduced in R13B03 as an experimental feature. It is a simpler and more efficient way of calling C-code than using port drivers. NIFs are most suitable for synchronous functions like foo and bar in the example, that does some relatively short calculations without side effects and return the result.

NIFs

A NIF (Native Implemented Function) is a function that is implemented in C instead of Erlang. NIFs appear as any other functions to the callers. They belong to a module and are called like any other Erlang functions. The NIFs of a module are compiled and linked into a dynamic loadable shared library (SO in Unix, DLL in Windows). The NIF library must be loaded in runtime by the Erlang code of the module.

Since a NIF library is dynamically linked into the emulator process, this is the fastest way of calling C-code from Erlang (alongside port drivers). Calling NIFs requires no context switches. But it is also the least safe, because a crash in a NIF will bring the emulator down too.

Erlang Program

Even if all functions of a module will be NIFs, you still need an Erlang module for two reasons. First, the NIF library must be explicitly loaded by Erlang code in the same module. Second, all NIFs of a module must have an Erlang implementation as well. Normally these are minimal stub implementations that throw an exception. But it can also be used as fallback implementations for functions that do not have native implemenations on some architectures.

NIF libraries are loaded by calling erlang:load_nif/2, with the name of the shared library as argument. The second argument can be any term that will be passed on to the library and used for initialization.

We use the directive on_load to get function init to be automatically called when the module is loaded. If init returns anything other than ok, such when the loading of the NIF library fails in this example, the module will be unloaded and calls to functions within it will fail.

Loading the NIF library will override the stub implementations and cause calls to foo and bar to be dispatched to the NIF implementations instead.

NIF library code

The NIFs of the module are compiled and linked into a shared library. Each NIF is implemented as a normal C function. The macro ERL_NIF_INIT together with an array of structures defines the names, arity and function pointers of all the NIFs in the module. The header file erl_nif.h must be included. Since the library is a shared module, not a program, no main function should be present.

The function arguments passed to a NIF appears in an array argv, with argc as the length of the array and thus the arity of the function. The Nth argument of the function can be accessed as argv[N-1]. NIFs also takes an environment argument that serves as an opaque handle that is needed to be passed on to most API functions. The environment contains information about the calling Erlang process.

The first argument to ERL_NIF_INIT must be the name of the Erlang module as a C-identifier. It will be stringified by the macro. The second argument is the array of ErlNifFunc structures containing name, arity and function pointer of each NIF. The other arguments are pointers to callback functions that can be used to initialize the library. We do not use them is this simple example so we set them all to NULL.

Function arguments and return values are represented as values of type ERL_NIF_TERM. We use functions like enif_get_int and enif_make_int to convert between Erlang term and C-type. If the function argument argv[0] is not an integer then enif_get_int will return false, in which case we return by throwing a badarg-exception with enif_make_badarg.

Running the Example

1. Compile the C code.

unix> gcc -o complex6_nif.so -fpic -shared complex.c complex6_nif.c
windows> cl -LD -MD -Fe complex6_nif.dll complex.c complex6_nif.c

2. Start Erlang and compile the Erlang code.

> erl
Erlang R13B04 (erts-5.7.5) [64-bit] [smp:4:4] [rq:4] [async-threads:0] [kernel-poll:false]

Eshell V5.7.5  (abort with ^G)
1> c(complex6).
{ok,complex6}

3. Run the example.

3> complex6:foo(3).
4
4> complex6:bar(5).
10
5> complex6:foo("not an integer").
** exception error: bad argument
     in function  complex6:foo/1
        called as comlpex6:foo("not an integer")