3 perlcall - Perl calling conventions from C
7 The purpose of this document is to show you how to call Perl subroutines
8 directly from C, i.e., how to write I<callbacks>.
10 Apart from discussing the C interface provided by Perl for writing
11 callbacks the document uses a series of examples to show how the
12 interface actually works in practice. In addition some techniques for
13 coding callbacks are covered.
15 Examples where callbacks are necessary include
19 =item * An Error Handler
21 You have created an XSUB interface to an application's C API.
23 A fairly common feature in applications is to allow you to define a C
24 function that will be called whenever something nasty occurs. What we
25 would like is to be able to specify a Perl subroutine that will be
28 =item * An Event-Driven Program
30 The classic example of where callbacks are used is when writing an
31 event driven program, such as for an X11 application. In this case
32 you register functions to be called whenever specific events occur,
33 e.g., a mouse button is pressed, the cursor moves into a window or a
34 menu item is selected.
38 Although the techniques described here are applicable when embedding
39 Perl in a C program, this is not the primary goal of this document.
40 There are other details that must be considered and are specific to
41 embedding Perl. For details on embedding Perl in C refer to
44 Before you launch yourself head first into the rest of this document,
45 it would be a good idea to have read the following two documents--L<perlxs>
48 =head1 THE CALL_ FUNCTIONS
50 Although this stuff is easier to explain using examples, you first need
51 be aware of a few important definitions.
53 Perl has a number of C functions that allow you to call Perl
56 I32 call_sv(SV* sv, I32 flags);
57 I32 call_pv(char *subname, I32 flags);
58 I32 call_method(char *methname, I32 flags);
59 I32 call_argv(char *subname, I32 flags, char **argv);
61 The key function is I<call_sv>. All the other functions are
62 fairly simple wrappers which make it easier to call Perl subroutines in
63 special cases. At the end of the day they will all call I<call_sv>
64 to invoke the Perl subroutine.
66 All the I<call_*> functions have a C<flags> parameter which is
67 used to pass a bit mask of options to Perl. This bit mask operates
68 identically for each of the functions. The settings available in the
69 bit mask are discussed in L</FLAG VALUES>.
71 Each of the functions will now be discussed in turn.
77 I<call_sv> takes two parameters. The first, C<sv>, is an SV*.
78 This allows you to specify the Perl subroutine to be called either as a
79 C string (which has first been converted to an SV) or a reference to a
80 subroutine. The section, L</Using call_sv>, shows how you can make
85 The function, I<call_pv>, is similar to I<call_sv> except it
86 expects its first parameter to be a C char* which identifies the Perl
87 subroutine you want to call, e.g., C<call_pv("fred", 0)>. If the
88 subroutine you want to call is in another package, just include the
89 package name in the string, e.g., C<"pkg::fred">.
93 The function I<call_method> is used to call a method from a Perl
94 class. The parameter C<methname> corresponds to the name of the method
95 to be called. Note that the class that the method belongs to is passed
96 on the Perl stack rather than in the parameter list. This class can be
97 either the name of the class (for a static method) or a reference to an
98 object (for a virtual method). See L<perlobj> for more information on
99 static and virtual methods and L</Using call_method> for an example
100 of using I<call_method>.
104 I<call_argv> calls the Perl subroutine specified by the C string
105 stored in the C<subname> parameter. It also takes the usual C<flags>
106 parameter. The final parameter, C<argv>, consists of a NULL-terminated
107 list of C strings to be passed as parameters to the Perl subroutine.
108 See L</Using call_argv>.
112 All the functions return an integer. This is a count of the number of
113 items returned by the Perl subroutine. The actual items returned by the
114 subroutine are stored on the Perl stack.
116 As a general rule you should I<always> check the return value from
117 these functions. Even if you are expecting only a particular number of
118 values to be returned from the Perl subroutine, there is nothing to
119 stop someone from doing something unexpected--don't say you haven't
124 The C<flags> parameter in all the I<call_*> functions is one of C<G_VOID>,
125 C<G_SCALAR>, or C<G_ARRAY>, which indicate the call context, OR'ed together
126 with a bit mask of any combination of the other G_* symbols defined below.
130 =for apidoc AmnUh||G_VOID
132 Calls the Perl subroutine in a void context.
134 This flag has 2 effects:
140 It indicates to the subroutine being called that it is executing in
141 a void context (if it executes I<wantarray> the result will be the
146 It ensures that nothing is actually returned from the subroutine.
150 The value returned by the I<call_*> function indicates how many
151 items have been returned by the Perl subroutine--in this case it will
157 =for apidoc AmnUh||G_SCALAR
159 Calls the Perl subroutine in a scalar context. This is the default
160 context flag setting for all the I<call_*> functions.
162 This flag has 2 effects:
168 It indicates to the subroutine being called that it is executing in a
169 scalar context (if it executes I<wantarray> the result will be false).
173 It ensures that only a scalar is actually returned from the subroutine.
174 The subroutine can, of course, ignore the I<wantarray> and return a
175 list anyway. If so, then only the last element of the list will be
180 The value returned by the I<call_*> function indicates how many
181 items have been returned by the Perl subroutine - in this case it will
184 If 0, then you have specified the G_DISCARD flag.
186 If 1, then the item actually returned by the Perl subroutine will be
187 stored on the Perl stack - the section L</Returning a Scalar> shows how
188 to access this value on the stack. Remember that regardless of how
189 many items the Perl subroutine returns, only the last one will be
190 accessible from the stack - think of the case where only one value is
191 returned as being a list with only one element. Any other items that
192 were returned will not exist by the time control returns from the
193 I<call_*> function. The section L</Returning a List in Scalar
194 Context> shows an example of this behavior.
199 =for apidoc AmnUh||G_ARRAY
201 Calls the Perl subroutine in a list context.
203 As with G_SCALAR, this flag has 2 effects:
209 It indicates to the subroutine being called that it is executing in a
210 list context (if it executes I<wantarray> the result will be true).
214 It ensures that all items returned from the subroutine will be
215 accessible when control returns from the I<call_*> function.
219 The value returned by the I<call_*> function indicates how many
220 items have been returned by the Perl subroutine.
222 If 0, then you have specified the G_DISCARD flag.
224 If not 0, then it will be a count of the number of items returned by
225 the subroutine. These items will be stored on the Perl stack. The
226 section L</Returning a List of Values> gives an example of using the
227 G_ARRAY flag and the mechanics of accessing the returned items from the
232 =for apidoc AmnUh||G_DISCARD
234 By default, the I<call_*> functions place the items returned from
235 by the Perl subroutine on the stack. If you are not interested in
236 these items, then setting this flag will make Perl get rid of them
237 automatically for you. Note that it is still possible to indicate a
238 context to the Perl subroutine by using either G_SCALAR or G_ARRAY.
240 If you do not set this flag then it is I<very> important that you make
241 sure that any temporaries (i.e., parameters passed to the Perl
242 subroutine and values returned from the subroutine) are disposed of
243 yourself. The section L</Returning a Scalar> gives details of how to
244 dispose of these temporaries explicitly and the section L</Using Perl to
245 Dispose of Temporaries> discusses the specific circumstances where you
246 can ignore the problem and let Perl deal with it for you.
250 =for apidoc AmnUh||G_NOARGS
252 Whenever a Perl subroutine is called using one of the I<call_*>
253 functions, it is assumed by default that parameters are to be passed to
254 the subroutine. If you are not passing any parameters to the Perl
255 subroutine, you can save a bit of time by setting this flag. It has
256 the effect of not creating the C<@_> array for the Perl subroutine.
258 Although the functionality provided by this flag may seem
259 straightforward, it should be used only if there is a good reason to do
260 so. The reason for being cautious is that, even if you have specified
261 the G_NOARGS flag, it is still possible for the Perl subroutine that
262 has been called to think that you have passed it parameters.
264 In fact, what can happen is that the Perl subroutine you have called
265 can access the C<@_> array from a previous Perl subroutine. This will
266 occur when the code that is executing the I<call_*> function has
267 itself been called from another Perl subroutine. The code below
282 What has happened is that C<fred> accesses the C<@_> array which
288 =for apidoc AmnUh||G_EVAL
290 It is possible for the Perl subroutine you are calling to terminate
291 abnormally, e.g., by calling I<die> explicitly or by not actually
292 existing. By default, when either of these events occurs, the
293 process will terminate immediately. If you want to trap this
294 type of event, specify the G_EVAL flag. It will put an I<eval { }>
295 around the subroutine call.
297 Whenever control returns from the I<call_*> function you need to
298 check the C<$@> variable as you would in a normal Perl script.
300 The value returned from the I<call_*> function is dependent on
301 what other flags have been specified and whether an error has
302 occurred. Here are all the different cases that can occur:
308 If the I<call_*> function returns normally, then the value
309 returned is as specified in the previous sections.
313 If G_DISCARD is specified, the return value will always be 0.
317 If G_ARRAY is specified I<and> an error has occurred, the return value
322 If G_SCALAR is specified I<and> an error has occurred, the return value
323 will be 1 and the value on the top of the stack will be I<undef>. This
324 means that if you have already detected the error by checking C<$@> and
325 you want the program to continue, you must remember to pop the I<undef>
330 See L</Using G_EVAL> for details on using G_EVAL.
334 =for apidoc AmnUh||G_KEEPERR
336 Using the G_EVAL flag described above will always set C<$@>: clearing
337 it if there was no error, and setting it to describe the error if there
338 was an error in the called code. This is what you want if your intention
339 is to handle possible errors, but sometimes you just want to trap errors
340 and stop them interfering with the rest of the program.
342 This scenario will mostly be applicable to code that is meant to be called
343 from within destructors, asynchronous callbacks, and signal handlers.
344 In such situations, where the code being called has little relation to the
345 surrounding dynamic context, the main program needs to be insulated from
346 errors in the called code, even if they can't be handled intelligently.
347 It may also be useful to do this with code for C<__DIE__> or C<__WARN__>
348 hooks, and C<tie> functions.
350 The G_KEEPERR flag is meant to be used in conjunction with G_EVAL in
351 I<call_*> functions that are used to implement such code, or with
352 C<eval_sv>. This flag has no effect on the C<call_*> functions when
355 When G_KEEPERR is used, any error in the called code will terminate the
356 call as usual, and the error will not propagate beyond the call (as usual
357 for G_EVAL), but it will not go into C<$@>. Instead the error will be
358 converted into a warning, prefixed with the string "\t(in cleanup)".
359 This can be disabled using C<no warnings 'misc'>. If there is no error,
360 C<$@> will not be cleared.
362 Note that the G_KEEPERR flag does not propagate into inner evals; these
365 The G_KEEPERR flag was introduced in Perl version 5.002.
367 See L</Using G_KEEPERR> for an example of a situation that warrants the
370 =head2 Determining the Context
372 As mentioned above, you can determine the context of the currently
373 executing subroutine in Perl with I<wantarray>. The equivalent test
374 can be made in C by using the C<GIMME_V> macro, which returns
375 C<G_ARRAY> if you have been called in a list context, C<G_SCALAR> if
376 in a scalar context, or C<G_VOID> if in a void context (i.e., the
377 return value will not be used). An older version of this macro is
378 called C<GIMME>; in a void context it returns C<G_SCALAR> instead of
379 C<G_VOID>. An example of using the C<GIMME_V> macro is shown in
380 section L</Using GIMME_V>.
384 Enough of the definition talk! Let's have a few examples.
386 Perl provides many macros to assist in accessing the Perl stack.
387 Wherever possible, these macros should always be used when interfacing
388 to Perl internals. We hope this should make the code less vulnerable
389 to any changes made to Perl in the future.
391 Another point worth noting is that in the first series of examples I
392 have made use of only the I<call_pv> function. This has been done
393 to keep the code simpler and ease you into the topic. Wherever
394 possible, if the choice is between using I<call_pv> and
395 I<call_sv>, you should always try to use I<call_sv>. See
396 L</Using call_sv> for details.
398 =head2 No Parameters, Nothing Returned
400 This first trivial example will call a Perl subroutine, I<PrintUID>, to
401 print out the UID of the process.
408 and here is a C function to call it
416 call_pv("PrintUID", G_DISCARD|G_NOARGS);
421 A few points to note about this example:
427 Ignore C<dSP> and C<PUSHMARK(SP)> for now. They will be discussed in
432 We aren't passing any parameters to I<PrintUID> so G_NOARGS can be
437 We aren't interested in anything returned from I<PrintUID>, so
438 G_DISCARD is specified. Even if I<PrintUID> was changed to
439 return some value(s), having specified G_DISCARD will mean that they
440 will be wiped by the time control returns from I<call_pv>.
444 As I<call_pv> is being used, the Perl subroutine is specified as a
445 C string. In this case the subroutine name has been 'hard-wired' into the
450 Because we specified G_DISCARD, it is not necessary to check the value
451 returned from I<call_pv>. It will always be 0.
455 =head2 Passing Parameters
457 Now let's make a slightly more complex example. This time we want to
458 call a Perl subroutine, C<LeftString>, which will take 2 parameters--a
459 string ($s) and an integer ($n). The subroutine will simply
460 print the first $n characters of the string.
462 So the Perl subroutine would look like this:
467 print substr($s, 0, $n), "\n";
470 The C function required to call I<LeftString> would look like this:
473 call_LeftString(a, b)
484 PUSHs(sv_2mortal(newSVpv(a, 0)));
485 PUSHs(sv_2mortal(newSViv(b)));
488 call_pv("LeftString", G_DISCARD);
494 Here are a few notes on the C function I<call_LeftString>.
500 Parameters are passed to the Perl subroutine using the Perl stack.
501 This is the purpose of the code beginning with the line C<dSP> and
502 ending with the line C<PUTBACK>. The C<dSP> declares a local copy
503 of the stack pointer. This local copy should B<always> be accessed
508 If you are going to put something onto the Perl stack, you need to know
509 where to put it. This is the purpose of the macro C<dSP>--it declares
510 and initializes a I<local> copy of the Perl stack pointer.
512 All the other macros which will be used in this example require you to
513 have used this macro.
515 The exception to this rule is if you are calling a Perl subroutine
516 directly from an XSUB function. In this case it is not necessary to
517 use the C<dSP> macro explicitly--it will be declared for you
522 Any parameters to be pushed onto the stack should be bracketed by the
523 C<PUSHMARK> and C<PUTBACK> macros. The purpose of these two macros, in
524 this context, is to count the number of parameters you are
525 pushing automatically. Then whenever Perl is creating the C<@_> array for the
526 subroutine, it knows how big to make it.
528 The C<PUSHMARK> macro tells Perl to make a mental note of the current
529 stack pointer. Even if you aren't passing any parameters (like the
530 example shown in the section L</No Parameters, Nothing Returned>) you
531 must still call the C<PUSHMARK> macro before you can call any of the
532 I<call_*> functions--Perl still needs to know that there are no
535 The C<PUTBACK> macro sets the global copy of the stack pointer to be
536 the same as our local copy. If we didn't do this, I<call_pv>
537 wouldn't know where the two parameters we pushed were--remember that
538 up to now all the stack pointer manipulation we have done is with our
539 local copy, I<not> the global copy.
543 Next, we come to EXTEND and PUSHs. This is where the parameters
544 actually get pushed onto the stack. In this case we are pushing a
545 string and an integer.
547 Alternatively you can use the XPUSHs() macro, which combines a
548 C<EXTEND(SP, 1)> and C<PUSHs()>. This is less efficient if you're
549 pushing multiple values.
551 See L<perlguts/"XSUBs and the Argument Stack"> for details
552 on how the PUSH macros work.
556 Because we created temporary values (by means of sv_2mortal() calls)
557 we will have to tidy up the Perl stack and dispose of mortal SVs.
559 This is the purpose of
564 at the start of the function, and
569 at the end. The C<ENTER>/C<SAVETMPS> pair creates a boundary for any
570 temporaries we create. This means that the temporaries we get rid of
571 will be limited to those which were created after these calls.
573 The C<FREETMPS>/C<LEAVE> pair will get rid of any values returned by
574 the Perl subroutine (see next example), plus it will also dump the
575 mortal SVs we have created. Having C<ENTER>/C<SAVETMPS> at the
576 beginning of the code makes sure that no other mortals are destroyed.
578 Think of these macros as working a bit like C<{> and C<}> in Perl
579 to limit the scope of local variables.
581 See the section L</Using Perl to Dispose of Temporaries> for details of
582 an alternative to using these macros.
586 Finally, I<LeftString> can now be called via the I<call_pv> function.
587 The only flag specified this time is G_DISCARD. Because we are passing
588 2 parameters to the Perl subroutine this time, we have not specified
593 =head2 Returning a Scalar
595 Now for an example of dealing with the items returned from a Perl
598 Here is a Perl subroutine, I<Adder>, that takes 2 integer parameters
599 and simply returns their sum.
607 Because we are now concerned with the return value from I<Adder>, the C
608 function required to call it is now a bit more complex.
623 PUSHs(sv_2mortal(newSViv(a)));
624 PUSHs(sv_2mortal(newSViv(b)));
627 count = call_pv("Adder", G_SCALAR);
632 croak("Big trouble\n");
634 printf ("The sum of %d and %d is %d\n", a, b, POPi);
641 Points to note this time are
647 The only flag specified this time was G_SCALAR. That means that the C<@_>
648 array will be created and that the value returned by I<Adder> will
649 still exist after the call to I<call_pv>.
653 The purpose of the macro C<SPAGAIN> is to refresh the local copy of the
654 stack pointer. This is necessary because it is possible that the memory
655 allocated to the Perl stack has been reallocated during the
658 If you are making use of the Perl stack pointer in your code you must
659 always refresh the local copy using SPAGAIN whenever you make use
660 of the I<call_*> functions or any other Perl internal function.
664 Although only a single value was expected to be returned from I<Adder>,
665 it is still good practice to check the return code from I<call_pv>
668 Expecting a single value is not quite the same as knowing that there
669 will be one. If someone modified I<Adder> to return a list and we
670 didn't check for that possibility and take appropriate action the Perl
671 stack would end up in an inconsistent state. That is something you
672 I<really> don't want to happen ever.
676 The C<POPi> macro is used here to pop the return value from the stack.
677 In this case we wanted an integer, so C<POPi> was used.
680 Here is the complete list of POP macros available, along with the types
685 POPpbytex pointer to bytes (PV)
688 POPu unsigned integer (UV)
692 Since these macros have side-effects don't use them as arguments to
693 macros that may evaluate their argument several times, for example:
695 /* Bad idea, don't do this */
697 const char *s = SvPV(POPs, len);
699 Instead, use a temporary:
703 const char *s = SvPV(sv, len);
705 or a macro that guarantees it will evaluate its arguments only once:
708 const char *s = SvPVx(POPs, len);
712 The final C<PUTBACK> is used to leave the Perl stack in a consistent
713 state before exiting the function. This is necessary because when we
714 popped the return value from the stack with C<POPi> it updated only our
715 local copy of the stack pointer. Remember, C<PUTBACK> sets the global
716 stack pointer to be the same as our local copy.
721 =head2 Returning a List of Values
723 Now, let's extend the previous example to return both the sum of the
724 parameters and the difference.
726 Here is the Perl subroutine
734 and this is the C function
737 call_AddSubtract(a, b)
749 PUSHs(sv_2mortal(newSViv(a)));
750 PUSHs(sv_2mortal(newSViv(b)));
753 count = call_pv("AddSubtract", G_ARRAY);
758 croak("Big trouble\n");
760 printf ("%d - %d = %d\n", a, b, POPi);
761 printf ("%d + %d = %d\n", a, b, POPi);
768 If I<call_AddSubtract> is called like this
770 call_AddSubtract(7, 4);
772 then here is the output
783 We wanted list context, so G_ARRAY was used.
787 Not surprisingly C<POPi> is used twice this time because we were
788 retrieving 2 values from the stack. The important thing to note is that
789 when using the C<POP*> macros they come off the stack in I<reverse>
794 =head2 Returning a List in Scalar Context
796 Say the Perl subroutine in the previous section was called in a scalar
800 call_AddSubScalar(a, b)
813 PUSHs(sv_2mortal(newSViv(a)));
814 PUSHs(sv_2mortal(newSViv(b)));
817 count = call_pv("AddSubtract", G_SCALAR);
821 printf ("Items Returned = %d\n", count);
823 for (i = 1; i <= count; ++i)
824 printf ("Value %d = %d\n", i, POPi);
831 The other modification made is that I<call_AddSubScalar> will print the
832 number of items returned from the Perl subroutine and their value (for
833 simplicity it assumes that they are integer). So if
834 I<call_AddSubScalar> is called
836 call_AddSubScalar(7, 4);
838 then the output will be
843 In this case the main point to note is that only the last item in the
844 list is returned from the subroutine. I<AddSubtract> actually made it back to
845 I<call_AddSubScalar>.
848 =head2 Returning Data from Perl via the Parameter List
850 It is also possible to return values directly via the parameter
851 list--whether it is actually desirable to do it is another matter entirely.
853 The Perl subroutine, I<Inc>, below takes 2 parameters and increments
862 and here is a C function to call it.
877 sva = sv_2mortal(newSViv(a));
878 svb = sv_2mortal(newSViv(b));
886 count = call_pv("Inc", G_DISCARD);
889 croak ("call_Inc: expected 0 values from 'Inc', got %d\n",
892 printf ("%d + 1 = %d\n", a, SvIV(sva));
893 printf ("%d + 1 = %d\n", b, SvIV(svb));
899 To be able to access the two parameters that were pushed onto the stack
900 after they return from I<call_pv> it is necessary to make a note
901 of their addresses--thus the two variables C<sva> and C<svb>.
903 The reason this is necessary is that the area of the Perl stack which
904 held them will very likely have been overwritten by something else by
905 the time control returns from I<call_pv>.
912 Now an example using G_EVAL. Below is a Perl subroutine which computes
913 the difference of its 2 parameters. If this would result in a negative
914 result, the subroutine calls I<die>.
920 die "death can be fatal\n" if $a < $b;
925 and some C to call it
941 PUSHs(sv_2mortal(newSViv(a)));
942 PUSHs(sv_2mortal(newSViv(b)));
945 count = call_pv("Subtract", G_EVAL|G_SCALAR);
949 /* Check the eval first */
953 printf ("Uh oh - %s\n", SvPV_nolen(err_tmp));
959 croak("call_Subtract: wanted 1 value from 'Subtract', got %d\n",
962 printf ("%d - %d = %d\n", a, b, POPi);
970 If I<call_Subtract> is called thus
974 the following will be printed
976 Uh oh - death can be fatal
984 We want to be able to catch the I<die> so we have used the G_EVAL
985 flag. Not specifying this flag would mean that the program would
986 terminate immediately at the I<die> statement in the subroutine
996 printf ("Uh oh - %s\n", SvPV_nolen(err_tmp));
1000 is the direct equivalent of this bit of Perl
1002 print "Uh oh - $@\n" if $@;
1004 C<PL_errgv> is a perl global of type C<GV *> that points to the symbol
1005 table entry containing the error. C<ERRSV> therefore refers to the C
1006 equivalent of C<$@>. We use a local temporary, C<err_tmp>, since
1007 C<ERRSV> is a macro that calls a function, and C<SvTRUE(ERRSV)> would
1008 end up calling that function multiple times.
1010 =for apidoc AmnUh|GV *|PL_errgv
1014 Note that the stack is popped using C<POPs> in the block where
1015 C<SvTRUE(err_tmp)> is true. This is necessary because whenever a
1016 I<call_*> function invoked with G_EVAL|G_SCALAR returns an error,
1017 the top of the stack holds the value I<undef>. Because we want the
1018 program to continue after detecting this error, it is essential that
1019 the stack be tidied up by removing the I<undef>.
1024 =head2 Using G_KEEPERR
1026 Consider this rather facetious example, where we have used an XS
1027 version of the call_Subtract example above inside a destructor:
1030 sub new { bless {}, $_[0] }
1033 die "death can be fatal" if $a < $b;
1036 sub DESTROY { call_Subtract(5, 4); }
1037 sub foo { die "foo dies"; }
1044 print "Saw: $@" if $@; # should be, but isn't
1046 This example will fail to recognize that an error occurred inside the
1047 C<eval {}>. Here's why: the call_Subtract code got executed while perl
1048 was cleaning up temporaries when exiting the outer braced block, and because
1049 call_Subtract is implemented with I<call_pv> using the G_EVAL
1050 flag, it promptly reset C<$@>. This results in the failure of the
1051 outermost test for C<$@>, and thereby the failure of the error trap.
1053 Appending the G_KEEPERR flag, so that the I<call_pv> call in
1054 call_Subtract reads:
1056 count = call_pv("Subtract", G_EVAL|G_SCALAR|G_KEEPERR);
1058 will preserve the error and restore reliable error handling.
1060 =head2 Using call_sv
1062 In all the previous examples I have 'hard-wired' the name of the Perl
1063 subroutine to be called from C. Most of the time though, it is more
1064 convenient to be able to specify the name of the Perl subroutine from
1065 within the Perl script, and you'll want to use
1066 L<call_sv|perlapi/call_sv>.
1068 Consider the Perl code below
1072 print "Hello there\n";
1077 Here is a snippet of XSUB which defines I<CallSubPV>.
1084 call_pv(name, G_DISCARD|G_NOARGS);
1086 That is fine as far as it goes. The thing is, the Perl subroutine
1087 can be specified as only a string, however, Perl allows references
1088 to subroutines and anonymous subroutines.
1089 This is where I<call_sv> is useful.
1091 The code below for I<CallSubSV> is identical to I<CallSubPV> except
1092 that the C<name> parameter is now defined as an SV* and we use
1093 I<call_sv> instead of I<call_pv>.
1100 call_sv(name, G_DISCARD|G_NOARGS);
1102 Because we are using an SV to call I<fred> the following can all be used:
1108 CallSubSV( sub { print "Hello there\n" } );
1110 As you can see, I<call_sv> gives you much greater flexibility in
1111 how you can specify the Perl subroutine.
1113 You should note that, if it is necessary to store the SV (C<name> in the
1114 example above) which corresponds to the Perl subroutine so that it can
1115 be used later in the program, it not enough just to store a copy of the
1116 pointer to the SV. Say the code above had been like this:
1118 static SV * rememberSub;
1130 call_sv(rememberSub, G_DISCARD|G_NOARGS);
1132 The reason this is wrong is that, by the time you come to use the
1133 pointer C<rememberSub> in C<CallSavedSub1>, it may or may not still refer
1134 to the Perl subroutine that was recorded in C<SaveSub1>. This is
1135 particularly true for these cases:
1140 SaveSub1( sub { print "Hello there\n" } );
1143 By the time each of the C<SaveSub1> statements above has been executed,
1144 the SV*s which corresponded to the parameters will no longer exist.
1145 Expect an error message from Perl of the form
1147 Can't use an undefined value as a subroutine reference at ...
1149 for each of the C<CallSavedSub1> lines.
1151 Similarly, with this code
1158 you can expect one of these messages (which you actually get is dependent on
1159 the version of Perl you are using)
1161 Not a CODE reference at ...
1162 Undefined subroutine &main::47 called ...
1164 The variable $ref may have referred to the subroutine C<fred>
1165 whenever the call to C<SaveSub1> was made but by the time
1166 C<CallSavedSub1> gets called it now holds the number C<47>. Because we
1167 saved only a pointer to the original SV in C<SaveSub1>, any changes to
1168 $ref will be tracked by the pointer C<rememberSub>. This means that
1169 whenever C<CallSavedSub1> gets called, it will attempt to execute the
1170 code which is referenced by the SV* C<rememberSub>. In this case
1171 though, it now refers to the integer C<47>, so expect Perl to complain
1174 A similar but more subtle problem is illustrated with this code:
1181 This time whenever C<CallSavedSub1> gets called it will execute the Perl
1182 subroutine C<joe> (assuming it exists) rather than C<fred> as was
1183 originally requested in the call to C<SaveSub1>.
1185 To get around these problems it is necessary to take a full copy of the
1186 SV. The code below shows C<SaveSub2> modified to do that.
1188 /* this isn't thread-safe */
1189 static SV * keepSub = (SV*)NULL;
1195 /* Take a copy of the callback */
1196 if (keepSub == (SV*)NULL)
1197 /* First time, so create a new SV */
1198 keepSub = newSVsv(name);
1200 /* Been here before, so overwrite */
1201 SvSetSV(keepSub, name);
1207 call_sv(keepSub, G_DISCARD|G_NOARGS);
1209 To avoid creating a new SV every time C<SaveSub2> is called,
1210 the function first checks to see if it has been called before. If not,
1211 then space for a new SV is allocated and the reference to the Perl
1212 subroutine C<name> is copied to the variable C<keepSub> in one
1213 operation using C<newSVsv>. Thereafter, whenever C<SaveSub2> is called,
1214 the existing SV, C<keepSub>, is overwritten with the new value using
1217 Note: using a static or global variable to store the SV isn't
1218 thread-safe. You can either use the C<MY_CXT> mechanism documented in
1219 L<perlxs/Safely Storing Static Data in XS> which is fast, or store the
1220 values in perl global variables, using get_sv(), which is much slower.
1222 =head2 Using call_argv
1224 Here is a Perl subroutine which prints whatever parameters are passed
1231 foreach (@list) { print "$_\n" }
1234 And here is an example of I<call_argv> which will call
1237 static char * words[] = {"alpha", "beta", "gamma", "delta", NULL};
1242 call_argv("PrintList", G_DISCARD, words);
1245 Note that it is not necessary to call C<PUSHMARK> in this instance.
1246 This is because I<call_argv> will do it for you.
1248 =head2 Using call_method
1250 Consider the following Perl code:
1263 my ($self, $index) = @_;
1264 print "$index: $$self[$index]\n";
1270 print "This is Class $class version 1.0\n";
1274 It implements just a very simple class to manage an array. Apart from
1275 the constructor, C<new>, it declares methods, one static and one
1276 virtual. The static method, C<PrintID>, prints out simply the class
1277 name and a version number. The virtual method, C<Display>, prints out a
1278 single element of the array. Here is an all-Perl example of using it.
1280 $a = Mine->new('red', 'green', 'blue');
1287 This is Class Mine version 1.0
1289 Calling a Perl method from C is fairly straightforward. The following
1290 things are required:
1296 A reference to the object for a virtual method or the name of the class
1301 The name of the method
1305 Any other parameters specific to the method
1309 Here is a simple XSUB which illustrates the mechanics of calling both
1310 the C<PrintID> and C<Display> methods from C.
1313 call_Method(ref, method, index)
1321 PUSHs(sv_2mortal(newSViv(index)));
1324 call_method(method, G_DISCARD);
1327 call_PrintID(class, method)
1332 XPUSHs(sv_2mortal(newSVpv(class, 0)));
1335 call_method(method, G_DISCARD);
1338 So the methods C<PrintID> and C<Display> can be invoked like this:
1340 $a = Mine->new('red', 'green', 'blue');
1341 call_Method($a, 'Display', 1);
1342 call_PrintID('Mine', 'PrintID');
1344 The only thing to note is that, in both the static and virtual methods,
1345 the method name is not passed via the stack--it is used as the first
1346 parameter to I<call_method>.
1348 =head2 Using GIMME_V
1350 Here is a trivial XSUB which prints the context in which it is
1351 currently executing.
1357 if (gimme == G_VOID)
1358 printf ("Context is Void\n");
1359 else if (gimme == G_SCALAR)
1360 printf ("Context is Scalar\n");
1362 printf ("Context is Array\n");
1364 And here is some Perl to test it.
1370 The output from that will be
1376 =head2 Using Perl to Dispose of Temporaries
1378 In the examples given to date, any temporaries created in the callback
1379 (i.e., parameters passed on the stack to the I<call_*> function or
1380 values returned via the stack) have been freed by one of these methods:
1386 Specifying the G_DISCARD flag with I<call_*>
1390 Explicitly using the C<ENTER>/C<SAVETMPS>--C<FREETMPS>/C<LEAVE> pairing
1394 There is another method which can be used, namely letting Perl do it
1395 for you automatically whenever it regains control after the callback
1396 has terminated. This is done by simply not using the
1404 sequence in the callback (and not, of course, specifying the G_DISCARD
1407 If you are going to use this method you have to be aware of a possible
1408 memory leak which can arise under very specific circumstances. To
1409 explain these circumstances you need to know a bit about the flow of
1410 control between Perl and the callback routine.
1412 The examples given at the start of the document (an error handler and
1413 an event driven program) are typical of the two main sorts of flow
1414 control that you are likely to encounter with callbacks. There is a
1415 very important distinction between them, so pay attention.
1417 In the first example, an error handler, the flow of control could be as
1418 follows. You have created an interface to an external library.
1419 Control can reach the external library like this
1421 perl --> XSUB --> external library
1423 Whilst control is in the library, an error condition occurs. You have
1424 previously set up a Perl callback to handle this situation, so it will
1425 get executed. Once the callback has finished, control will drop back to
1426 Perl again. Here is what the flow of control will be like in that
1429 perl --> XSUB --> external library
1433 external library --> call_* --> perl
1435 perl <-- XSUB <-- external library <-- call_* <----+
1437 After processing of the error using I<call_*> is completed,
1438 control reverts back to Perl more or less immediately.
1440 In the diagram, the further right you go the more deeply nested the
1441 scope is. It is only when control is back with perl on the extreme
1442 left of the diagram that you will have dropped back to the enclosing
1443 scope and any temporaries you have left hanging around will be freed.
1445 In the second example, an event driven program, the flow of control
1446 will be more like this
1448 perl --> XSUB --> event handler
1450 event handler --> call_* --> perl
1452 event handler <-- call_* <----+
1454 event handler --> call_* --> perl
1456 event handler <-- call_* <----+
1458 event handler --> call_* --> perl
1460 event handler <-- call_* <----+
1462 In this case the flow of control can consist of only the repeated
1465 event handler --> call_* --> perl
1467 for practically the complete duration of the program. This means that
1468 control may I<never> drop back to the surrounding scope in Perl at the
1471 So what is the big problem? Well, if you are expecting Perl to tidy up
1472 those temporaries for you, you might be in for a long wait. For Perl
1473 to dispose of your temporaries, control must drop back to the
1474 enclosing scope at some stage. In the event driven scenario that may
1475 never happen. This means that, as time goes on, your program will
1476 create more and more temporaries, none of which will ever be freed. As
1477 each of these temporaries consumes some memory your program will
1478 eventually consume all the available memory in your system--kapow!
1480 So here is the bottom line--if you are sure that control will revert
1481 back to the enclosing Perl scope fairly quickly after the end of your
1482 callback, then it isn't absolutely necessary to dispose explicitly of
1483 any temporaries you may have created. Mind you, if you are at all
1484 uncertain about what to do, it doesn't do any harm to tidy up anyway.
1487 =head2 Strategies for Storing Callback Context Information
1490 Potentially one of the trickiest problems to overcome when designing a
1491 callback interface can be figuring out how to store the mapping between
1492 the C callback function and the Perl equivalent.
1494 To help understand why this can be a real problem first consider how a
1495 callback is set up in an all C environment. Typically a C API will
1496 provide a function to register a callback. This will expect a pointer
1497 to a function as one of its parameters. Below is a call to a
1498 hypothetical function C<register_fatal> which registers the C function
1499 to get called when a fatal error occurs.
1501 register_fatal(cb1);
1503 The single parameter C<cb1> is a pointer to a function, so you must
1504 have defined C<cb1> in your code, say something like this
1509 printf ("Fatal Error\n");
1513 Now change that to call a Perl subroutine instead
1515 static SV * callback = (SV*)NULL;
1524 /* Call the Perl sub to process the callback */
1525 call_sv(callback, G_DISCARD);
1533 /* Remember the Perl sub */
1534 if (callback == (SV*)NULL)
1535 callback = newSVsv(fn);
1537 SvSetSV(callback, fn);
1539 /* register the callback with the external library */
1540 register_fatal(cb1);
1542 where the Perl equivalent of C<register_fatal> and the callback it
1543 registers, C<pcb1>, might look like this
1545 # Register the sub pcb1
1546 register_fatal(\&pcb1);
1550 die "I'm dying...\n";
1553 The mapping between the C callback and the Perl equivalent is stored in
1554 the global variable C<callback>.
1556 This will be adequate if you ever need to have only one callback
1557 registered at any time. An example could be an error handler like the
1558 code sketched out above. Remember though, repeated calls to
1559 C<register_fatal> will replace the previously registered callback
1560 function with the new one.
1562 Say for example you want to interface to a library which allows asynchronous
1563 file i/o. In this case you may be able to register a callback whenever
1564 a read operation has completed. To be of any use we want to be able to
1565 call separate Perl subroutines for each file that is opened. As it
1566 stands, the error handler example above would not be adequate as it
1567 allows only a single callback to be defined at any time. What we
1568 require is a means of storing the mapping between the opened file and
1569 the Perl subroutine we want to be called for that file.
1571 Say the i/o library has a function C<asynch_read> which associates a C
1572 function C<ProcessRead> with a file handle C<fh>--this assumes that it
1573 has also provided some routine to open the file and so obtain the file
1576 asynch_read(fh, ProcessRead)
1578 This may expect the C I<ProcessRead> function of this form
1581 ProcessRead(fh, buffer)
1588 To provide a Perl interface to this library we need to be able to map
1589 between the C<fh> parameter and the Perl subroutine we want called. A
1590 hash is a convenient mechanism for storing this mapping. The code
1591 below shows a possible implementation
1593 static HV * Mapping = (HV*)NULL;
1596 asynch_read(fh, callback)
1600 /* If the hash doesn't already exist, create it */
1601 if (Mapping == (HV*)NULL)
1604 /* Save the fh -> callback mapping */
1605 hv_store(Mapping, (char*)&fh, sizeof(fh), newSVsv(callback), 0);
1607 /* Register with the C Library */
1608 asynch_read(fh, asynch_read_if);
1610 and C<asynch_read_if> could look like this
1613 asynch_read_if(fh, buffer)
1620 /* Get the callback associated with fh */
1621 sv = hv_fetch(Mapping, (char*)&fh , sizeof(fh), FALSE);
1622 if (sv == (SV**)NULL)
1623 croak("Internal error...\n");
1627 PUSHs(sv_2mortal(newSViv(fh)));
1628 PUSHs(sv_2mortal(newSVpv(buffer, 0)));
1631 /* Call the Perl sub */
1632 call_sv(*sv, G_DISCARD);
1635 For completeness, here is C<asynch_close>. This shows how to remove
1636 the entry from the hash C<Mapping>.
1642 /* Remove the entry from the hash */
1643 (void) hv_delete(Mapping, (char*)&fh, sizeof(fh), G_DISCARD);
1645 /* Now call the real asynch_close */
1648 So the Perl interface would look like this
1652 my($handle, $buffer) = @_;
1655 # Register the Perl callback
1656 asynch_read($fh, \&callback1);
1660 The mapping between the C callback and Perl is stored in the global
1661 hash C<Mapping> this time. Using a hash has the distinct advantage that
1662 it allows an unlimited number of callbacks to be registered.
1664 What if the interface provided by the C callback doesn't contain a
1665 parameter which allows the file handle to Perl subroutine mapping? Say
1666 in the asynchronous i/o package, the callback function gets passed only
1667 the C<buffer> parameter like this
1676 Without the file handle there is no straightforward way to map from the
1677 C callback to the Perl subroutine.
1679 In this case a possible way around this problem is to predefine a
1680 series of C functions to act as the interface to Perl, thus
1683 #define NULL_HANDLE -1
1684 typedef void (*FnMap)();
1696 static struct MapStruct Map [MAX_CB] =
1698 { fn1, NULL, NULL_HANDLE },
1699 { fn2, NULL, NULL_HANDLE },
1700 { fn3, NULL, NULL_HANDLE }
1711 XPUSHs(sv_2mortal(newSVpv(buffer, 0)));
1714 /* Call the Perl sub */
1715 call_sv(Map[index].PerlSub, G_DISCARD);
1740 array_asynch_read(fh, callback)
1745 int null_index = MAX_CB;
1747 /* Find the same handle or an empty entry */
1748 for (index = 0; index < MAX_CB; ++index)
1750 if (Map[index].Handle == fh)
1753 if (Map[index].Handle == NULL_HANDLE)
1757 if (index == MAX_CB && null_index == MAX_CB)
1758 croak ("Too many callback functions registered\n");
1760 if (index == MAX_CB)
1763 /* Save the file handle */
1764 Map[index].Handle = fh;
1766 /* Remember the Perl sub */
1767 if (Map[index].PerlSub == (SV*)NULL)
1768 Map[index].PerlSub = newSVsv(callback);
1770 SvSetSV(Map[index].PerlSub, callback);
1772 asynch_read(fh, Map[index].Function);
1775 array_asynch_close(fh)
1780 /* Find the file handle */
1781 for (index = 0; index < MAX_CB; ++ index)
1782 if (Map[index].Handle == fh)
1785 if (index == MAX_CB)
1786 croak ("could not close fh %d\n", fh);
1788 Map[index].Handle = NULL_HANDLE;
1789 SvREFCNT_dec(Map[index].PerlSub);
1790 Map[index].PerlSub = (SV*)NULL;
1794 In this case the functions C<fn1>, C<fn2>, and C<fn3> are used to
1795 remember the Perl subroutine to be called. Each of the functions holds
1796 a separate hard-wired index which is used in the function C<Pcb> to
1797 access the C<Map> array and actually call the Perl subroutine.
1799 There are some obvious disadvantages with this technique.
1801 Firstly, the code is considerably more complex than with the previous
1804 Secondly, there is a hard-wired limit (in this case 3) to the number of
1805 callbacks that can exist simultaneously. The only way to increase the
1806 limit is by modifying the code to add more functions and then
1807 recompiling. None the less, as long as the number of functions is
1808 chosen with some care, it is still a workable solution and in some
1809 cases is the only one available.
1811 To summarize, here are a number of possible methods for you to consider
1812 for storing the mapping between C and the Perl callback
1816 =item 1. Ignore the problem - Allow only 1 callback
1818 For a lot of situations, like interfacing to an error handler, this may
1819 be a perfectly adequate solution.
1821 =item 2. Create a sequence of callbacks - hard wired limit
1823 If it is impossible to tell from the parameters passed back from the C
1824 callback what the context is, then you may need to create a sequence of C
1825 callback interface functions, and store pointers to each in an array.
1827 =item 3. Use a parameter to map to the Perl callback
1829 A hash is an ideal mechanism to store the mapping between C and Perl.
1834 =head2 Alternate Stack Manipulation
1837 Although I have made use of only the C<POP*> macros to access values
1838 returned from Perl subroutines, it is also possible to bypass these
1839 macros and read the stack using the C<ST> macro (See L<perlxs> for a
1840 full description of the C<ST> macro).
1842 Most of the time the C<POP*> macros should be adequate; the main
1843 problem with them is that they force you to process the returned values
1844 in sequence. This may not be the most suitable way to process the
1845 values in some cases. What we want is to be able to access the stack in
1846 a random order. The C<ST> macro as used when coding an XSUB is ideal
1849 The code below is the example given in the section L</Returning a List
1850 of Values> recoded to use C<ST> instead of C<POP*>.
1853 call_AddSubtract2(a, b)
1866 PUSHs(sv_2mortal(newSViv(a)));
1867 PUSHs(sv_2mortal(newSViv(b)));
1870 count = call_pv("AddSubtract", G_ARRAY);
1874 ax = (SP - PL_stack_base) + 1;
1877 croak("Big trouble\n");
1879 printf ("%d + %d = %d\n", a, b, SvIV(ST(0)));
1880 printf ("%d - %d = %d\n", a, b, SvIV(ST(1)));
1893 Notice that it was necessary to define the variable C<ax>. This is
1894 because the C<ST> macro expects it to exist. If we were in an XSUB it
1895 would not be necessary to define C<ax> as it is already defined for
1904 ax = (SP - PL_stack_base) + 1;
1906 sets the stack up so that we can use the C<ST> macro.
1910 Unlike the original coding of this example, the returned
1911 values are not accessed in reverse order. So C<ST(0)> refers to the
1912 first value returned by the Perl subroutine and C<ST(count-1)>
1917 =head2 Creating and Calling an Anonymous Subroutine in C
1919 As we've already shown, C<call_sv> can be used to invoke an
1920 anonymous subroutine. However, our example showed a Perl script
1921 invoking an XSUB to perform this operation. Let's see how it can be
1922 done inside our C code:
1928 print 'You will not find me cluttering any namespace!'
1933 call_sv(cvrv, G_VOID|G_NOARGS);
1935 C<eval_pv> is used to compile the anonymous subroutine, which
1936 will be the return value as well (read more about C<eval_pv> in
1937 L<perlapi/eval_pv>). Once this code reference is in hand, it
1938 can be mixed in with all the previous examples we've shown.
1940 =head1 LIGHTWEIGHT CALLBACKS
1942 Sometimes you need to invoke the same subroutine repeatedly.
1943 This usually happens with a function that acts on a list of
1944 values, such as Perl's built-in sort(). You can pass a
1945 comparison function to sort(), which will then be invoked
1946 for every pair of values that needs to be compared. The first()
1947 and reduce() functions from L<List::Util> follow a similar
1950 In this case it is possible to speed up the routine (often
1951 quite substantially) by using the lightweight callback API.
1952 The idea is that the calling context only needs to be
1953 created and destroyed once, and the sub can be called
1954 arbitrarily many times in between.
1956 It is usual to pass parameters using global variables (typically
1957 $_ for one parameter, or $a and $b for two parameters) rather
1958 than via @_. (It is possible to use the @_ mechanism if you know
1959 what you're doing, though there is as yet no supported API for
1960 it. It's also inherently slower.)
1962 The pattern of macro calls is like this:
1964 dMULTICALL; /* Declare local variables */
1965 U8 gimme = G_SCALAR; /* context of the call: G_SCALAR,
1966 * G_ARRAY, or G_VOID */
1968 PUSH_MULTICALL(cv); /* Set up the context for calling cv,
1969 and set local vars appropriately */
1972 /* set the value(s) af your parameter variables */
1973 MULTICALL; /* Make the actual call */
1976 POP_MULTICALL; /* Tear down the calling context */
1978 For some concrete examples, see the implementation of the
1979 first() and reduce() functions of List::Util 1.18. There you
1980 will also find a header file that emulates the multicall API
1981 on older versions of perl.
1985 L<perlxs>, L<perlguts>, L<perlembed>
1991 Special thanks to the following people who assisted in the creation of
1994 Jeff Okamoto, Tim Bunce, Nick Gianniotis, Steve Kelem, Gurusamy Sarathy
1999 Last updated for perl 5.23.1.