• TUX



Perl 5 version 18.1 documentation
Recently read



perlcall - Perl calling conventions from C


The purpose of this document is to show you how to call Perl subroutines directly from C, i.e., how to write callbacks.

Apart from discussing the C interface provided by Perl for writing callbacks the document uses a series of examples to show how the interface actually works in practice. In addition some techniques for coding callbacks are covered.

Examples where callbacks are necessary include

  • An Error Handler

    You have created an XSUB interface to an application's C API.

    A fairly common feature in applications is to allow you to define a C function that will be called whenever something nasty occurs. What we would like is to be able to specify a Perl subroutine that will be called instead.

  • An Event-Driven Program

    The classic example of where callbacks are used is when writing an event driven program, such as for an X11 application. In this case you register functions to be called whenever specific events occur, e.g., a mouse button is pressed, the cursor moves into a window or a menu item is selected.

Although the techniques described here are applicable when embedding Perl in a C program, this is not the primary goal of this document. There are other details that must be considered and are specific to embedding Perl. For details on embedding Perl in C refer to perlembed.

Before you launch yourself head first into the rest of this document, it would be a good idea to have read the following two documents--perlxs and perlguts.


Although this stuff is easier to explain using examples, you first need be aware of a few important definitions.

Perl has a number of C functions that allow you to call Perl subroutines. They are

  1. I32 call_sv(SV* sv, I32 flags);
  2. I32 call_pv(char *subname, I32 flags);
  3. I32 call_method(char *methname, I32 flags);
  4. I32 call_argv(char *subname, I32 flags, char **argv);

The key function is call_sv. All the other functions are fairly simple wrappers which make it easier to call Perl subroutines in special cases. At the end of the day they will all call call_sv to invoke the Perl subroutine.

All the call_* functions have a flags parameter which is used to pass a bit mask of options to Perl. This bit mask operates identically for each of the functions. The settings available in the bit mask are discussed in FLAG VALUES.

Each of the functions will now be discussed in turn.

  • call_sv

    call_sv takes two parameters. The first, sv , is an SV*. This allows you to specify the Perl subroutine to be called either as a C string (which has first been converted to an SV) or a reference to a subroutine. The section, Using call_sv, shows how you can make use of call_sv.

  • call_pv

    The function, call_pv, is similar to call_sv except it expects its first parameter to be a C char* which identifies the Perl subroutine you want to call, e.g., call_pv("fred", 0) . If the subroutine you want to call is in another package, just include the package name in the string, e.g., "pkg::fred" .

  • call_method

    The function call_method is used to call a method from a Perl class. The parameter methname corresponds to the name of the method to be called. Note that the class that the method belongs to is passed on the Perl stack rather than in the parameter list. This class can be either the name of the class (for a static method) or a reference to an object (for a virtual method). See perlobj for more information on static and virtual methods and Using call_method for an example of using call_method.

  • call_argv

    call_argv calls the Perl subroutine specified by the C string stored in the subname parameter. It also takes the usual flags parameter. The final parameter, argv , consists of a NULL-terminated list of C strings to be passed as parameters to the Perl subroutine. See Using call_argv.

All the functions return an integer. This is a count of the number of items returned by the Perl subroutine. The actual items returned by the subroutine are stored on the Perl stack.

As a general rule you should always check the return value from these functions. Even if you are expecting only a particular number of values to be returned from the Perl subroutine, there is nothing to stop someone from doing something unexpected--don't say you haven't been warned.


The flags parameter in all the call_* functions is one of G_VOID, G_SCALAR, or G_ARRAY, which indicate the call context, OR'ed together with a bit mask of any combination of the other G_* symbols defined below.


Calls the Perl subroutine in a void context.

This flag has 2 effects:


It indicates to the subroutine being called that it is executing in a void context (if it executes wantarray the result will be the undefined value).


It ensures that nothing is actually returned from the subroutine.

The value returned by the call_* function indicates how many items have been returned by the Perl subroutine--in this case it will be 0.


Calls the Perl subroutine in a scalar context. This is the default context flag setting for all the call_* functions.

This flag has 2 effects:


It indicates to the subroutine being called that it is executing in a scalar context (if it executes wantarray the result will be false).


It ensures that only a scalar is actually returned from the subroutine. The subroutine can, of course, ignore the wantarray and return a list anyway. If so, then only the last element of the list will be returned.

The value returned by the call_* function indicates how many items have been returned by the Perl subroutine - in this case it will be either 0 or 1.

If 0, then you have specified the G_DISCARD flag.

If 1, then the item actually returned by the Perl subroutine will be stored on the Perl stack - the section Returning a Scalar shows how to access this value on the stack. Remember that regardless of how many items the Perl subroutine returns, only the last one will be accessible from the stack - think of the case where only one value is returned as being a list with only one element. Any other items that were returned will not exist by the time control returns from the call_* function. The section Returning a list in a scalar context shows an example of this behavior.


Calls the Perl subroutine in a list context.

As with G_SCALAR, this flag has 2 effects:


It indicates to the subroutine being called that it is executing in a list context (if it executes wantarray the result will be true).


It ensures that all items returned from the subroutine will be accessible when control returns from the call_* function.

The value returned by the call_* function indicates how many items have been returned by the Perl subroutine.

If 0, then you have specified the G_DISCARD flag.

If not 0, then it will be a count of the number of items returned by the subroutine. These items will be stored on the Perl stack. The section Returning a list of values gives an example of using the G_ARRAY flag and the mechanics of accessing the returned items from the Perl stack.


By default, the call_* functions place the items returned from by the Perl subroutine on the stack. If you are not interested in these items, then setting this flag will make Perl get rid of them automatically for you. Note that it is still possible to indicate a context to the Perl subroutine by using either G_SCALAR or G_ARRAY.

If you do not set this flag then it is very important that you make sure that any temporaries (i.e., parameters passed to the Perl subroutine and values returned from the subroutine) are disposed of yourself. The section Returning a Scalar gives details of how to dispose of these temporaries explicitly and the section Using Perl to dispose of temporaries discusses the specific circumstances where you can ignore the problem and let Perl deal with it for you.


Whenever a Perl subroutine is called using one of the call_* functions, it is assumed by default that parameters are to be passed to the subroutine. If you are not passing any parameters to the Perl subroutine, you can save a bit of time by setting this flag. It has the effect of not creating the @_ array for the Perl subroutine.

Although the functionality provided by this flag may seem straightforward, it should be used only if there is a good reason to do so. The reason for being cautious is that, even if you have specified the G_NOARGS flag, it is still possible for the Perl subroutine that has been called to think that you have passed it parameters.

In fact, what can happen is that the Perl subroutine you have called can access the @_ array from a previous Perl subroutine. This will occur when the code that is executing the call_* function has itself been called from another Perl subroutine. The code below illustrates this

  1. sub fred
  2. { print "@_\n" }
  3. sub joe
  4. { &fred }
  5. &joe(1,2,3);

This will print

  1. 1 2 3

What has happened is that fred accesses the @_ array which belongs to joe .


It is possible for the Perl subroutine you are calling to terminate abnormally, e.g., by calling die explicitly or by not actually existing. By default, when either of these events occurs, the process will terminate immediately. If you want to trap this type of event, specify the G_EVAL flag. It will put an eval { } around the subroutine call.

Whenever control returns from the call_* function you need to check the $@ variable as you would in a normal Perl script.

The value returned from the call_* function is dependent on what other flags have been specified and whether an error has occurred. Here are all the different cases that can occur:

  • If the call_* function returns normally, then the value returned is as specified in the previous sections.

  • If G_DISCARD is specified, the return value will always be 0.

  • If G_ARRAY is specified and an error has occurred, the return value will always be 0.

  • If G_SCALAR is specified and an error has occurred, the return value will be 1 and the value on the top of the stack will be undef. This means that if you have already detected the error by checking $@ and you want the program to continue, you must remember to pop the undef from the stack.

See Using G_EVAL for details on using G_EVAL.


Using the G_EVAL flag described above will always set $@ : clearing it if there was no error, and setting it to describe the error if there was an error in the called code. This is what you want if your intention is to handle possible errors, but sometimes you just want to trap errors and stop them interfering with the rest of the program.

This scenario will mostly be applicable to code that is meant to be called from within destructors, asynchronous callbacks, and signal handlers. In such situations, where the code being called has little relation to the surrounding dynamic context, the main program needs to be insulated from errors in the called code, even if they can't be handled intelligently. It may also be useful to do this with code for __DIE__ or __WARN__ hooks, and tie functions.

The G_KEEPERR flag is meant to be used in conjunction with G_EVAL in call_* functions that are used to implement such code, or with eval_sv . This flag has no effect on the call_* functions when G_EVAL is not used.

When G_KEEPERR is used, any error in the called code will terminate the call as usual, and the error will not propagate beyond the call (as usual for G_EVAL), but it will not go into $@ . Instead the error will be converted into a warning, prefixed with the string "\t(in cleanup)". This can be disabled using no warnings 'misc' . If there is no error, $@ will not be cleared.

Note that the G_KEEPERR flag does not propagate into inner evals; these may still set $@ .

The G_KEEPERR flag was introduced in Perl version 5.002.

See Using G_KEEPERR for an example of a situation that warrants the use of this flag.

Determining the Context

As mentioned above, you can determine the context of the currently executing subroutine in Perl with wantarray. The equivalent test can be made in C by using the GIMME_V macro, which returns G_ARRAY if you have been called in a list context, G_SCALAR if in a scalar context, or G_VOID if in a void context (i.e., the return value will not be used). An older version of this macro is called GIMME ; in a void context it returns G_SCALAR instead of G_VOID . An example of using the GIMME_V macro is shown in section Using GIMME_V.


Enough of the definition talk! Let's have a few examples.

Perl provides many macros to assist in accessing the Perl stack. Wherever possible, these macros should always be used when interfacing to Perl internals. We hope this should make the code less vulnerable to any changes made to Perl in the future.

Another point worth noting is that in the first series of examples I have made use of only the call_pv function. This has been done to keep the code simpler and ease you into the topic. Wherever possible, if the choice is between using call_pv and call_sv, you should always try to use call_sv. See Using call_sv for details.

No Parameters, Nothing Returned

This first trivial example will call a Perl subroutine, PrintUID, to print out the UID of the process.

  1. sub PrintUID
  2. {
  3. print "UID is $<\n";
  4. }

and here is a C function to call it

  1. static void
  2. call_PrintUID()
  3. {
  4. dSP;
  6. call_pv("PrintUID", G_DISCARD|G_NOARGS);
  7. }

Simple, eh?

A few points to note about this example:


Ignore dSP and PUSHMARK(SP) for now. They will be discussed in the next example.


We aren't passing any parameters to PrintUID so G_NOARGS can be specified.


We aren't interested in anything returned from PrintUID, so G_DISCARD is specified. Even if PrintUID was changed to return some value(s), having specified G_DISCARD will mean that they will be wiped by the time control returns from call_pv.


As call_pv is being used, the Perl subroutine is specified as a C string. In this case the subroutine name has been 'hard-wired' into the code.


Because we specified G_DISCARD, it is not necessary to check the value returned from call_pv. It will always be 0.

Passing Parameters

Now let's make a slightly more complex example. This time we want to call a Perl subroutine, LeftString , which will take 2 parameters--a string ($s) and an integer ($n). The subroutine will simply print the first $n characters of the string.

So the Perl subroutine would look like this:

  1. sub LeftString
  2. {
  3. my($s, $n) = @_;
  4. print substr($s, 0, $n), "\n";
  5. }

The C function required to call LeftString would look like this:

  1. static void
  2. call_LeftString(a, b)
  3. char * a;
  4. int b;
  5. {
  6. dSP;
  7. ENTER;
  10. XPUSHs(sv_2mortal(newSVpv(a, 0)));
  11. XPUSHs(sv_2mortal(newSViv(b)));
  12. PUTBACK;
  13. call_pv("LeftString", G_DISCARD);
  15. LEAVE;
  16. }

Here are a few notes on the C function call_LeftString.


Parameters are passed to the Perl subroutine using the Perl stack. This is the purpose of the code beginning with the line dSP and ending with the line PUTBACK . The dSP declares a local copy of the stack pointer. This local copy should always be accessed as SP .


If you are going to put something onto the Perl stack, you need to know where to put it. This is the purpose of the macro dSP --it declares and initializes a local copy of the Perl stack pointer.

All the other macros which will be used in this example require you to have used this macro.

The exception to this rule is if you are calling a Perl subroutine directly from an XSUB function. In this case it is not necessary to use the dSP macro explicitly--it will be declared for you automatically.


Any parameters to be pushed onto the stack should be bracketed by the PUSHMARK and PUTBACK macros. The purpose of these two macros, in this context, is to count the number of parameters you are pushing automatically. Then whenever Perl is creating the @_ array for the subroutine, it knows how big to make it.

The PUSHMARK macro tells Perl to make a mental note of the current stack pointer. Even if you aren't passing any parameters (like the example shown in the section No Parameters, Nothing Returned) you must still call the PUSHMARK macro before you can call any of the call_* functions--Perl still needs to know that there are no parameters.

The PUTBACK macro sets the global copy of the stack pointer to be the same as our local copy. If we didn't do this, call_pv wouldn't know where the two parameters we pushed were--remember that up to now all the stack pointer manipulation we have done is with our local copy, not the global copy.


Next, we come to XPUSHs. This is where the parameters actually get pushed onto the stack. In this case we are pushing a string and an integer.

See XSUBs and the Argument Stack in perlguts for details on how the XPUSH macros work.


Because we created temporary values (by means of sv_2mortal() calls) we will have to tidy up the Perl stack and dispose of mortal SVs.

This is the purpose of

  1. ENTER;

at the start of the function, and

  2. LEAVE;

at the end. The ENTER /SAVETMPS pair creates a boundary for any temporaries we create. This means that the temporaries we get rid of will be limited to those which were created after these calls.

The FREETMPS /LEAVE pair will get rid of any values returned by the Perl subroutine (see next example), plus it will also dump the mortal SVs we have created. Having ENTER /SAVETMPS at the beginning of the code makes sure that no other mortals are destroyed.

Think of these macros as working a bit like { and } in Perl to limit the scope of local variables.

See the section Using Perl to Dispose of Temporaries for details of an alternative to using these macros.


Finally, LeftString can now be called via the call_pv function. The only flag specified this time is G_DISCARD. Because we are passing 2 parameters to the Perl subroutine this time, we have not specified G_NOARGS.

Returning a Scalar

Now for an example of dealing with the items returned from a Perl subroutine.

Here is a Perl subroutine, Adder, that takes 2 integer parameters and simply returns their sum.

  1. sub Adder
  2. {
  3. my($a, $b) = @_;
  4. $a + $b;
  5. }

Because we are now concerned with the return value from Adder, the C function required to call it is now a bit more complex.

  1. static void
  2. call_Adder(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. int count;
  8. ENTER;
  11. XPUSHs(sv_2mortal(newSViv(a)));
  12. XPUSHs(sv_2mortal(newSViv(b)));
  13. PUTBACK;
  14. count = call_pv("Adder", G_SCALAR);
  15. SPAGAIN;
  16. if (count != 1)
  17. croak("Big trouble\n");
  18. printf ("The sum of %d and %d is %d\n", a, b, POPi);
  19. PUTBACK;
  21. LEAVE;
  22. }

Points to note this time are


The only flag specified this time was G_SCALAR. That means that the @_ array will be created and that the value returned by Adder will still exist after the call to call_pv.


The purpose of the macro SPAGAIN is to refresh the local copy of the stack pointer. This is necessary because it is possible that the memory allocated to the Perl stack has been reallocated during the call_pv call.

If you are making use of the Perl stack pointer in your code you must always refresh the local copy using SPAGAIN whenever you make use of the call_* functions or any other Perl internal function.


Although only a single value was expected to be returned from Adder, it is still good practice to check the return code from call_pv anyway.

Expecting a single value is not quite the same as knowing that there will be one. If someone modified Adder to return a list and we didn't check for that possibility and take appropriate action the Perl stack would end up in an inconsistent state. That is something you really don't want to happen ever.


The POPi macro is used here to pop the return value from the stack. In this case we wanted an integer, so POPi was used.

Here is the complete list of POP macros available, along with the types they return.

  1. POPs SV
  2. POPp pointer
  3. POPn double
  4. POPi integer
  5. POPl long

The final PUTBACK is used to leave the Perl stack in a consistent state before exiting the function. This is necessary because when we popped the return value from the stack with POPi it updated only our local copy of the stack pointer. Remember, PUTBACK sets the global stack pointer to be the same as our local copy.

Returning a List of Values

Now, let's extend the previous example to return both the sum of the parameters and the difference.

Here is the Perl subroutine

  1. sub AddSubtract
  2. {
  3. my($a, $b) = @_;
  4. ($a+$b, $a-$b);
  5. }

and this is the C function

  1. static void
  2. call_AddSubtract(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. int count;
  8. ENTER;
  11. XPUSHs(sv_2mortal(newSViv(a)));
  12. XPUSHs(sv_2mortal(newSViv(b)));
  13. PUTBACK;
  14. count = call_pv("AddSubtract", G_ARRAY);
  15. SPAGAIN;
  16. if (count != 2)
  17. croak("Big trouble\n");
  18. printf ("%d - %d = %d\n", a, b, POPi);
  19. printf ("%d + %d = %d\n", a, b, POPi);
  20. PUTBACK;
  22. LEAVE;
  23. }

If call_AddSubtract is called like this

  1. call_AddSubtract(7, 4);

then here is the output

  1. 7 - 4 = 3
  2. 7 + 4 = 11



We wanted list context, so G_ARRAY was used.


Not surprisingly POPi is used twice this time because we were retrieving 2 values from the stack. The important thing to note is that when using the POP* macros they come off the stack in reverse order.

Returning a List in a Scalar Context

Say the Perl subroutine in the previous section was called in a scalar context, like this

  1. static void
  2. call_AddSubScalar(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. int count;
  8. int i;
  9. ENTER;
  12. XPUSHs(sv_2mortal(newSViv(a)));
  13. XPUSHs(sv_2mortal(newSViv(b)));
  14. PUTBACK;
  15. count = call_pv("AddSubtract", G_SCALAR);
  16. SPAGAIN;
  17. printf ("Items Returned = %d\n", count);
  18. for (i = 1; i <= count; ++i)
  19. printf ("Value %d = %d\n", i, POPi);
  20. PUTBACK;
  22. LEAVE;
  23. }

The other modification made is that call_AddSubScalar will print the number of items returned from the Perl subroutine and their value (for simplicity it assumes that they are integer). So if call_AddSubScalar is called

  1. call_AddSubScalar(7, 4);

then the output will be

  1. Items Returned = 1
  2. Value 1 = 3

In this case the main point to note is that only the last item in the list is returned from the subroutine. AddSubtract actually made it back to call_AddSubScalar.

Returning Data from Perl via the Parameter List

It is also possible to return values directly via the parameter list--whether it is actually desirable to do it is another matter entirely.

The Perl subroutine, Inc, below takes 2 parameters and increments each directly.

  1. sub Inc
  2. {
  3. ++ $_[0];
  4. ++ $_[1];
  5. }

and here is a C function to call it.

  1. static void
  2. call_Inc(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. int count;
  8. SV * sva;
  9. SV * svb;
  10. ENTER;
  12. sva = sv_2mortal(newSViv(a));
  13. svb = sv_2mortal(newSViv(b));
  15. XPUSHs(sva);
  16. XPUSHs(svb);
  17. PUTBACK;
  18. count = call_pv("Inc", G_DISCARD);
  19. if (count != 0)
  20. croak ("call_Inc: expected 0 values from 'Inc', got %d\n",
  21. count);
  22. printf ("%d + 1 = %d\n", a, SvIV(sva));
  23. printf ("%d + 1 = %d\n", b, SvIV(svb));
  25. LEAVE;
  26. }

To be able to access the two parameters that were pushed onto the stack after they return from call_pv it is necessary to make a note of their addresses--thus the two variables sva and svb .

The reason this is necessary is that the area of the Perl stack which held them will very likely have been overwritten by something else by the time control returns from call_pv.

Using G_EVAL

Now an example using G_EVAL. Below is a Perl subroutine which computes the difference of its 2 parameters. If this would result in a negative result, the subroutine calls die.

  1. sub Subtract
  2. {
  3. my ($a, $b) = @_;
  4. die "death can be fatal\n" if $a < $b;
  5. $a - $b;
  6. }

and some C to call it

  1. static void
  2. call_Subtract(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. int count;
  8. ENTER;
  11. XPUSHs(sv_2mortal(newSViv(a)));
  12. XPUSHs(sv_2mortal(newSViv(b)));
  13. PUTBACK;
  14. count = call_pv("Subtract", G_EVAL|G_SCALAR);
  15. SPAGAIN;
  16. /* Check the eval first */
  17. if (SvTRUE(ERRSV))
  18. {
  19. printf ("Uh oh - %s\n", SvPV_nolen(ERRSV));
  20. POPs;
  21. }
  22. else
  23. {
  24. if (count != 1)
  25. croak("call_Subtract: wanted 1 value from 'Subtract', got %d\n",
  26. count);
  27. printf ("%d - %d = %d\n", a, b, POPi);
  28. }
  29. PUTBACK;
  31. LEAVE;
  32. }

If call_Subtract is called thus

  1. call_Subtract(4, 5)

the following will be printed

  1. Uh oh - death can be fatal



We want to be able to catch the die so we have used the G_EVAL flag. Not specifying this flag would mean that the program would terminate immediately at the die statement in the subroutine Subtract.


The code

  1. if (SvTRUE(ERRSV))
  2. {
  3. printf ("Uh oh - %s\n", SvPV_nolen(ERRSV));
  4. POPs;
  5. }

is the direct equivalent of this bit of Perl

  1. print "Uh oh - $@\n" if $@;

PL_errgv is a perl global of type GV * that points to the symbol table entry containing the error. ERRSV therefore refers to the C equivalent of $@ .


Note that the stack is popped using POPs in the block where SvTRUE(ERRSV) is true. This is necessary because whenever a call_* function invoked with G_EVAL|G_SCALAR returns an error, the top of the stack holds the value undef. Because we want the program to continue after detecting this error, it is essential that the stack be tidied up by removing the undef.


Consider this rather facetious example, where we have used an XS version of the call_Subtract example above inside a destructor:

  1. package Foo;
  2. sub new { bless {}, $_[0] }
  3. sub Subtract {
  4. my($a,$b) = @_;
  5. die "death can be fatal" if $a < $b;
  6. $a - $b;
  7. }
  8. sub DESTROY { call_Subtract(5, 4); }
  9. sub foo { die "foo dies"; }
  10. package main;
  11. {
  12. my $foo = Foo->new;
  13. eval { $foo->foo };
  14. }
  15. print "Saw: $@" if $@; # should be, but isn't

This example will fail to recognize that an error occurred inside the eval {} . Here's why: the call_Subtract code got executed while perl was cleaning up temporaries when exiting the outer braced block, and because call_Subtract is implemented with call_pv using the G_EVAL flag, it promptly reset $@ . This results in the failure of the outermost test for $@ , and thereby the failure of the error trap.

Appending the G_KEEPERR flag, so that the call_pv call in call_Subtract reads:

  1. count = call_pv("Subtract", G_EVAL|G_SCALAR|G_KEEPERR);

will preserve the error and restore reliable error handling.

Using call_sv

In all the previous examples I have 'hard-wired' the name of the Perl subroutine to be called from C. Most of the time though, it is more convenient to be able to specify the name of the Perl subroutine from within the Perl script.

Consider the Perl code below

  1. sub fred
  2. {
  3. print "Hello there\n";
  4. }
  5. CallSubPV("fred");

Here is a snippet of XSUB which defines CallSubPV.

  1. void
  2. CallSubPV(name)
  3. char * name
  4. CODE:
  6. call_pv(name, G_DISCARD|G_NOARGS);

That is fine as far as it goes. The thing is, the Perl subroutine can be specified as only a string, however, Perl allows references to subroutines and anonymous subroutines. This is where call_sv is useful.

The code below for CallSubSV is identical to CallSubPV except that the name parameter is now defined as an SV* and we use call_sv instead of call_pv.

  1. void
  2. CallSubSV(name)
  3. SV * name
  4. CODE:
  6. call_sv(name, G_DISCARD|G_NOARGS);

Because we are using an SV to call fred the following can all be used:

  1. CallSubSV("fred");
  2. CallSubSV(\&fred);
  3. $ref = \&fred;
  4. CallSubSV($ref);
  5. CallSubSV( sub { print "Hello there\n" } );

As you can see, call_sv gives you much greater flexibility in how you can specify the Perl subroutine.

You should note that, if it is necessary to store the SV (name in the example above) which corresponds to the Perl subroutine so that it can be used later in the program, it not enough just to store a copy of the pointer to the SV. Say the code above had been like this:

  1. static SV * rememberSub;
  2. void
  3. SaveSub1(name)
  4. SV * name
  5. CODE:
  6. rememberSub = name;
  7. void
  8. CallSavedSub1()
  9. CODE:
  11. call_sv(rememberSub, G_DISCARD|G_NOARGS);

The reason this is wrong is that, by the time you come to use the pointer rememberSub in CallSavedSub1 , it may or may not still refer to the Perl subroutine that was recorded in SaveSub1 . This is particularly true for these cases:

  1. SaveSub1(\&fred);
  2. CallSavedSub1();
  3. SaveSub1( sub { print "Hello there\n" } );
  4. CallSavedSub1();

By the time each of the SaveSub1 statements above has been executed, the SV*s which corresponded to the parameters will no longer exist. Expect an error message from Perl of the form

  1. Can't use an undefined value as a subroutine reference at ...

for each of the CallSavedSub1 lines.

Similarly, with this code

  1. $ref = \&fred;
  2. SaveSub1($ref);
  3. $ref = 47;
  4. CallSavedSub1();

you can expect one of these messages (which you actually get is dependent on the version of Perl you are using)

  1. Not a CODE reference at ...
  2. Undefined subroutine &main::47 called ...

The variable $ref may have referred to the subroutine fred whenever the call to SaveSub1 was made but by the time CallSavedSub1 gets called it now holds the number 47 . Because we saved only a pointer to the original SV in SaveSub1 , any changes to $ref will be tracked by the pointer rememberSub . This means that whenever CallSavedSub1 gets called, it will attempt to execute the code which is referenced by the SV* rememberSub . In this case though, it now refers to the integer 47 , so expect Perl to complain loudly.

A similar but more subtle problem is illustrated with this code:

  1. $ref = \&fred;
  2. SaveSub1($ref);
  3. $ref = \&joe;
  4. CallSavedSub1();

This time whenever CallSavedSub1 gets called it will execute the Perl subroutine joe (assuming it exists) rather than fred as was originally requested in the call to SaveSub1 .

To get around these problems it is necessary to take a full copy of the SV. The code below shows SaveSub2 modified to do that.

  1. static SV * keepSub = (SV*)NULL;
  2. void
  3. SaveSub2(name)
  4. SV * name
  5. CODE:
  6. /* Take a copy of the callback */
  7. if (keepSub == (SV*)NULL)
  8. /* First time, so create a new SV */
  9. keepSub = newSVsv(name);
  10. else
  11. /* Been here before, so overwrite */
  12. SvSetSV(keepSub, name);
  13. void
  14. CallSavedSub2()
  15. CODE:
  17. call_sv(keepSub, G_DISCARD|G_NOARGS);

To avoid creating a new SV every time SaveSub2 is called, the function first checks to see if it has been called before. If not, then space for a new SV is allocated and the reference to the Perl subroutine name is copied to the variable keepSub in one operation using newSVsv . Thereafter, whenever SaveSub2 is called, the existing SV, keepSub , is overwritten with the new value using SvSetSV .

Using call_argv

Here is a Perl subroutine which prints whatever parameters are passed to it.

  1. sub PrintList
  2. {
  3. my(@list) = @_;
  4. foreach (@list) { print "$_\n" }
  5. }

And here is an example of call_argv which will call PrintList.

  1. static char * words[] = {"alpha", "beta", "gamma", "delta", NULL};
  2. static void
  3. call_PrintList()
  4. {
  5. dSP;
  6. call_argv("PrintList", G_DISCARD, words);
  7. }

Note that it is not necessary to call PUSHMARK in this instance. This is because call_argv will do it for you.

Using call_method

Consider the following Perl code:

  1. {
  2. package Mine;
  3. sub new
  4. {
  5. my($type) = shift;
  6. bless [@_]
  7. }
  8. sub Display
  9. {
  10. my ($self, $index) = @_;
  11. print "$index: $$self[$index]\n";
  12. }
  13. sub PrintID
  14. {
  15. my($class) = @_;
  16. print "This is Class $class version 1.0\n";
  17. }
  18. }

It implements just a very simple class to manage an array. Apart from the constructor, new , it declares methods, one static and one virtual. The static method, PrintID , prints out simply the class name and a version number. The virtual method, Display , prints out a single element of the array. Here is an all-Perl example of using it.

  1. $a = Mine->new('red', 'green', 'blue');
  2. $a->Display(1);
  3. Mine->PrintID;

will print

  1. 1: green
  2. This is Class Mine version 1.0

Calling a Perl method from C is fairly straightforward. The following things are required:

  • A reference to the object for a virtual method or the name of the class for a static method

  • The name of the method

  • Any other parameters specific to the method

Here is a simple XSUB which illustrates the mechanics of calling both the PrintID and Display methods from C.

  1. void
  2. call_Method(ref, method, index)
  3. SV * ref
  4. char * method
  5. int index
  6. CODE:
  8. XPUSHs(ref);
  9. XPUSHs(sv_2mortal(newSViv(index)));
  10. PUTBACK;
  11. call_method(method, G_DISCARD);
  12. void
  13. call_PrintID(class, method)
  14. char * class
  15. char * method
  16. CODE:
  18. XPUSHs(sv_2mortal(newSVpv(class, 0)));
  19. PUTBACK;
  20. call_method(method, G_DISCARD);

So the methods PrintID and Display can be invoked like this:

  1. $a = Mine->new('red', 'green', 'blue');
  2. call_Method($a, 'Display', 1);
  3. call_PrintID('Mine', 'PrintID');

The only thing to note is that, in both the static and virtual methods, the method name is not passed via the stack--it is used as the first parameter to call_method.


Here is a trivial XSUB which prints the context in which it is currently executing.

  1. void
  2. PrintContext()
  3. CODE:
  4. I32 gimme = GIMME_V;
  5. if (gimme == G_VOID)
  6. printf ("Context is Void\n");
  7. else if (gimme == G_SCALAR)
  8. printf ("Context is Scalar\n");
  9. else
  10. printf ("Context is Array\n");

And here is some Perl to test it.

  1. PrintContext;
  2. $a = PrintContext;
  3. @a = PrintContext;

The output from that will be

  1. Context is Void
  2. Context is Scalar
  3. Context is Array

Using Perl to Dispose of Temporaries

In the examples given to date, any temporaries created in the callback (i.e., parameters passed on the stack to the call_* function or values returned via the stack) have been freed by one of these methods:

  • Specifying the G_DISCARD flag with call_*

  • Explicitly using the ENTER /SAVETMPS --FREETMPS /LEAVE pairing

There is another method which can be used, namely letting Perl do it for you automatically whenever it regains control after the callback has terminated. This is done by simply not using the

  1. ENTER;
  3. ...
  5. LEAVE;

sequence in the callback (and not, of course, specifying the G_DISCARD flag).

If you are going to use this method you have to be aware of a possible memory leak which can arise under very specific circumstances. To explain these circumstances you need to know a bit about the flow of control between Perl and the callback routine.

The examples given at the start of the document (an error handler and an event driven program) are typical of the two main sorts of flow control that you are likely to encounter with callbacks. There is a very important distinction between them, so pay attention.

In the first example, an error handler, the flow of control could be as follows. You have created an interface to an external library. Control can reach the external library like this

  1. perl --> XSUB --> external library

Whilst control is in the library, an error condition occurs. You have previously set up a Perl callback to handle this situation, so it will get executed. Once the callback has finished, control will drop back to Perl again. Here is what the flow of control will be like in that situation

  1. perl --> XSUB --> external library
  2. ...
  3. error occurs
  4. ...
  5. external library --> call_* --> perl
  6. |
  7. perl <-- XSUB <-- external library <-- call_* <----+

After processing of the error using call_* is completed, control reverts back to Perl more or less immediately.

In the diagram, the further right you go the more deeply nested the scope is. It is only when control is back with perl on the extreme left of the diagram that you will have dropped back to the enclosing scope and any temporaries you have left hanging around will be freed.

In the second example, an event driven program, the flow of control will be more like this

  1. perl --> XSUB --> event handler
  2. ...
  3. event handler --> call_* --> perl
  4. |
  5. event handler <-- call_* <----+
  6. ...
  7. event handler --> call_* --> perl
  8. |
  9. event handler <-- call_* <----+
  10. ...
  11. event handler --> call_* --> perl
  12. |
  13. event handler <-- call_* <----+

In this case the flow of control can consist of only the repeated sequence

  1. event handler --> call_* --> perl

for practically the complete duration of the program. This means that control may never drop back to the surrounding scope in Perl at the extreme left.

So what is the big problem? Well, if you are expecting Perl to tidy up those temporaries for you, you might be in for a long wait. For Perl to dispose of your temporaries, control must drop back to the enclosing scope at some stage. In the event driven scenario that may never happen. This means that, as time goes on, your program will create more and more temporaries, none of which will ever be freed. As each of these temporaries consumes some memory your program will eventually consume all the available memory in your system--kapow!

So here is the bottom line--if you are sure that control will revert back to the enclosing Perl scope fairly quickly after the end of your callback, then it isn't absolutely necessary to dispose explicitly of any temporaries you may have created. Mind you, if you are at all uncertain about what to do, it doesn't do any harm to tidy up anyway.

Strategies for Storing Callback Context Information

Potentially one of the trickiest problems to overcome when designing a callback interface can be figuring out how to store the mapping between the C callback function and the Perl equivalent.

To help understand why this can be a real problem first consider how a callback is set up in an all C environment. Typically a C API will provide a function to register a callback. This will expect a pointer to a function as one of its parameters. Below is a call to a hypothetical function register_fatal which registers the C function to get called when a fatal error occurs.

  1. register_fatal(cb1);

The single parameter cb1 is a pointer to a function, so you must have defined cb1 in your code, say something like this

  1. static void
  2. cb1()
  3. {
  4. printf ("Fatal Error\n");
  5. exit(1);
  6. }

Now change that to call a Perl subroutine instead

  1. static SV * callback = (SV*)NULL;
  2. static void
  3. cb1()
  4. {
  5. dSP;
  7. /* Call the Perl sub to process the callback */
  8. call_sv(callback, G_DISCARD);
  9. }
  10. void
  11. register_fatal(fn)
  12. SV * fn
  13. CODE:
  14. /* Remember the Perl sub */
  15. if (callback == (SV*)NULL)
  16. callback = newSVsv(fn);
  17. else
  18. SvSetSV(callback, fn);
  19. /* register the callback with the external library */
  20. register_fatal(cb1);

where the Perl equivalent of register_fatal and the callback it registers, pcb1 , might look like this

  1. # Register the sub pcb1
  2. register_fatal(\&pcb1);
  3. sub pcb1
  4. {
  5. die "I'm dying...\n";
  6. }

The mapping between the C callback and the Perl equivalent is stored in the global variable callback .

This will be adequate if you ever need to have only one callback registered at any time. An example could be an error handler like the code sketched out above. Remember though, repeated calls to register_fatal will replace the previously registered callback function with the new one.

Say for example you want to interface to a library which allows asynchronous file i/o. In this case you may be able to register a callback whenever a read operation has completed. To be of any use we want to be able to call separate Perl subroutines for each file that is opened. As it stands, the error handler example above would not be adequate as it allows only a single callback to be defined at any time. What we require is a means of storing the mapping between the opened file and the Perl subroutine we want to be called for that file.

Say the i/o library has a function asynch_read which associates a C function ProcessRead with a file handle fh --this assumes that it has also provided some routine to open the file and so obtain the file handle.

  1. asynch_read(fh, ProcessRead)

This may expect the C ProcessRead function of this form

  1. void
  2. ProcessRead(fh, buffer)
  3. int fh;
  4. char * buffer;
  5. {
  6. ...
  7. }

To provide a Perl interface to this library we need to be able to map between the fh parameter and the Perl subroutine we want called. A hash is a convenient mechanism for storing this mapping. The code below shows a possible implementation

  1. static HV * Mapping = (HV*)NULL;
  2. void
  3. asynch_read(fh, callback)
  4. int fh
  5. SV * callback
  6. CODE:
  7. /* If the hash doesn't already exist, create it */
  8. if (Mapping == (HV*)NULL)
  9. Mapping = newHV();
  10. /* Save the fh -> callback mapping */
  11. hv_store(Mapping, (char*)&fh, sizeof(fh), newSVsv(callback), 0);
  12. /* Register with the C Library */
  13. asynch_read(fh, asynch_read_if);

and asynch_read_if could look like this

  1. static void
  2. asynch_read_if(fh, buffer)
  3. int fh;
  4. char * buffer;
  5. {
  6. dSP;
  7. SV ** sv;
  8. /* Get the callback associated with fh */
  9. sv = hv_fetch(Mapping, (char*)&fh , sizeof(fh), FALSE);
  10. if (sv == (SV**)NULL)
  11. croak("Internal error...\n");
  13. XPUSHs(sv_2mortal(newSViv(fh)));
  14. XPUSHs(sv_2mortal(newSVpv(buffer, 0)));
  15. PUTBACK;
  16. /* Call the Perl sub */
  17. call_sv(*sv, G_DISCARD);
  18. }

For completeness, here is asynch_close . This shows how to remove the entry from the hash Mapping .

  1. void
  2. asynch_close(fh)
  3. int fh
  4. CODE:
  5. /* Remove the entry from the hash */
  6. (void) hv_delete(Mapping, (char*)&fh, sizeof(fh), G_DISCARD);
  7. /* Now call the real asynch_close */
  8. asynch_close(fh);

So the Perl interface would look like this

  1. sub callback1
  2. {
  3. my($handle, $buffer) = @_;
  4. }
  5. # Register the Perl callback
  6. asynch_read($fh, \&callback1);
  7. asynch_close($fh);

The mapping between the C callback and Perl is stored in the global hash Mapping this time. Using a hash has the distinct advantage that it allows an unlimited number of callbacks to be registered.

What if the interface provided by the C callback doesn't contain a parameter which allows the file handle to Perl subroutine mapping? Say in the asynchronous i/o package, the callback function gets passed only the buffer parameter like this

  1. void
  2. ProcessRead(buffer)
  3. char * buffer;
  4. {
  5. ...
  6. }

Without the file handle there is no straightforward way to map from the C callback to the Perl subroutine.

In this case a possible way around this problem is to predefine a series of C functions to act as the interface to Perl, thus

  1. #define MAX_CB 3
  2. #define NULL_HANDLE -1
  3. typedef void (*FnMap)();
  4. struct MapStruct {
  5. FnMap Function;
  6. SV * PerlSub;
  7. int Handle;
  8. };
  9. static void fn1();
  10. static void fn2();
  11. static void fn3();
  12. static struct MapStruct Map [MAX_CB] =
  13. {
  14. { fn1, NULL, NULL_HANDLE },
  15. { fn2, NULL, NULL_HANDLE },
  16. { fn3, NULL, NULL_HANDLE }
  17. };
  18. static void
  19. Pcb(index, buffer)
  20. int index;
  21. char * buffer;
  22. {
  23. dSP;
  25. XPUSHs(sv_2mortal(newSVpv(buffer, 0)));
  26. PUTBACK;
  27. /* Call the Perl sub */
  28. call_sv(Map[index].PerlSub, G_DISCARD);
  29. }
  30. static void
  31. fn1(buffer)
  32. char * buffer;
  33. {
  34. Pcb(0, buffer);
  35. }
  36. static void
  37. fn2(buffer)
  38. char * buffer;
  39. {
  40. Pcb(1, buffer);
  41. }
  42. static void
  43. fn3(buffer)
  44. char * buffer;
  45. {
  46. Pcb(2, buffer);
  47. }
  48. void
  49. array_asynch_read(fh, callback)
  50. int fh
  51. SV * callback
  52. CODE:
  53. int index;
  54. int null_index = MAX_CB;
  55. /* Find the same handle or an empty entry */
  56. for (index = 0; index < MAX_CB; ++index)
  57. {
  58. if (Map[index].Handle == fh)
  59. break;
  60. if (Map[index].Handle == NULL_HANDLE)
  61. null_index = index;
  62. }
  63. if (index == MAX_CB && null_index == MAX_CB)
  64. croak ("Too many callback functions registered\n");
  65. if (index == MAX_CB)
  66. index = null_index;
  67. /* Save the file handle */
  68. Map[index].Handle = fh;
  69. /* Remember the Perl sub */
  70. if (Map[index].PerlSub == (SV*)NULL)
  71. Map[index].PerlSub = newSVsv(callback);
  72. else
  73. SvSetSV(Map[index].PerlSub, callback);
  74. asynch_read(fh, Map[index].Function);
  75. void
  76. array_asynch_close(fh)
  77. int fh
  78. CODE:
  79. int index;
  80. /* Find the file handle */
  81. for (index = 0; index < MAX_CB; ++ index)
  82. if (Map[index].Handle == fh)
  83. break;
  84. if (index == MAX_CB)
  85. croak ("could not close fh %d\n", fh);
  86. Map[index].Handle = NULL_HANDLE;
  87. SvREFCNT_dec(Map[index].PerlSub);
  88. Map[index].PerlSub = (SV*)NULL;
  89. asynch_close(fh);

In this case the functions fn1 , fn2 , and fn3 are used to remember the Perl subroutine to be called. Each of the functions holds a separate hard-wired index which is used in the function Pcb to access the Map array and actually call the Perl subroutine.

There are some obvious disadvantages with this technique.

Firstly, the code is considerably more complex than with the previous example.

Secondly, there is a hard-wired limit (in this case 3) to the number of callbacks that can exist simultaneously. The only way to increase the limit is by modifying the code to add more functions and then recompiling. None the less, as long as the number of functions is chosen with some care, it is still a workable solution and in some cases is the only one available.

To summarize, here are a number of possible methods for you to consider for storing the mapping between C and the Perl callback

Ignore the problem - Allow only 1 callback

For a lot of situations, like interfacing to an error handler, this may be a perfectly adequate solution.

Create a sequence of callbacks - hard wired limit

If it is impossible to tell from the parameters passed back from the C callback what the context is, then you may need to create a sequence of C callback interface functions, and store pointers to each in an array.

Use a parameter to map to the Perl callback

A hash is an ideal mechanism to store the mapping between C and Perl.

Alternate Stack Manipulation

Although I have made use of only the POP* macros to access values returned from Perl subroutines, it is also possible to bypass these macros and read the stack using the ST macro (See perlxs for a full description of the ST macro).

Most of the time the POP* macros should be adequate; the main problem with them is that they force you to process the returned values in sequence. This may not be the most suitable way to process the values in some cases. What we want is to be able to access the stack in a random order. The ST macro as used when coding an XSUB is ideal for this purpose.

The code below is the example given in the section Returning a List of Values recoded to use ST instead of POP* .

  1. static void
  2. call_AddSubtract2(a, b)
  3. int a;
  4. int b;
  5. {
  6. dSP;
  7. I32 ax;
  8. int count;
  9. ENTER;
  12. XPUSHs(sv_2mortal(newSViv(a)));
  13. XPUSHs(sv_2mortal(newSViv(b)));
  14. PUTBACK;
  15. count = call_pv("AddSubtract", G_ARRAY);
  16. SPAGAIN;
  17. SP -= count;
  18. ax = (SP - PL_stack_base) + 1;
  19. if (count != 2)
  20. croak("Big trouble\n");
  21. printf ("%d + %d = %d\n", a, b, SvIV(ST(0)));
  22. printf ("%d - %d = %d\n", a, b, SvIV(ST(1)));
  23. PUTBACK;
  25. LEAVE;
  26. }



Notice that it was necessary to define the variable ax . This is because the ST macro expects it to exist. If we were in an XSUB it would not be necessary to define ax as it is already defined for us.


The code

  2. SP -= count;
  3. ax = (SP - PL_stack_base) + 1;

sets the stack up so that we can use the ST macro.


Unlike the original coding of this example, the returned values are not accessed in reverse order. So ST(0) refers to the first value returned by the Perl subroutine and ST(count-1) refers to the last.

Creating and Calling an Anonymous Subroutine in C

As we've already shown, call_sv can be used to invoke an anonymous subroutine. However, our example showed a Perl script invoking an XSUB to perform this operation. Let's see how it can be done inside our C code:

  1. ...
  2. SV *cvrv = eval_pv("sub { print 'You will not find me cluttering any namespace!' }", TRUE);
  3. ...
  4. call_sv(cvrv, G_VOID|G_NOARGS);

eval_pv is used to compile the anonymous subroutine, which will be the return value as well (read more about eval_pv in eval_pv in perlapi). Once this code reference is in hand, it can be mixed in with all the previous examples we've shown.


Sometimes you need to invoke the same subroutine repeatedly. This usually happens with a function that acts on a list of values, such as Perl's built-in sort(). You can pass a comparison function to sort(), which will then be invoked for every pair of values that needs to be compared. The first() and reduce() functions from List::Util follow a similar pattern.

In this case it is possible to speed up the routine (often quite substantially) by using the lightweight callback API. The idea is that the calling context only needs to be created and destroyed once, and the sub can be called arbitrarily many times in between.

It is usual to pass parameters using global variables (typically $_ for one parameter, or $a and $b for two parameters) rather than via @_. (It is possible to use the @_ mechanism if you know what you're doing, though there is as yet no supported API for it. It's also inherently slower.)

The pattern of macro calls is like this:

  1. dMULTICALL; /* Declare local variables */
  2. I32 gimme = G_SCALAR; /* context of the call: G_SCALAR,
  3. * G_ARRAY, or G_VOID */
  4. PUSH_MULTICALL(cv); /* Set up the context for calling cv,
  5. and set local vars appropriately */
  6. /* loop */ {
  7. /* set the value(s) af your parameter variables */
  8. MULTICALL; /* Make the actual call */
  9. } /* end of loop */
  10. POP_MULTICALL; /* Tear down the calling context */

For some concrete examples, see the implementation of the first() and reduce() functions of List::Util 1.18. There you will also find a header file that emulates the multicall API on older versions of perl.


perlxs, perlguts, perlembed


Paul Marquess

Special thanks to the following people who assisted in the creation of the document.

Jeff Okamoto, Tim Bunce, Nick Gianniotis, Steve Kelem, Gurusamy Sarathy and Larry Wall.


Version 1.3, 14th Apr 1997