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, I<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 I<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 G_VOID,
125 G_SCALAR, or 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 Calls the Perl subroutine in a void context.
132 This flag has 2 effects:
138 It indicates to the subroutine being called that it is executing in
139 a void context (if it executes I<wantarray> the result will be the
144 It ensures that nothing is actually returned from the subroutine.
148 The value returned by the I<call_*> function indicates how many
149 items have been returned by the Perl subroutine--in this case it will
155 Calls the Perl subroutine in a scalar context. This is the default
156 context flag setting for all the I<call_*> functions.
158 This flag has 2 effects:
164 It indicates to the subroutine being called that it is executing in a
165 scalar context (if it executes I<wantarray> the result will be false).
169 It ensures that only a scalar is actually returned from the subroutine.
170 The subroutine can, of course, ignore the I<wantarray> and return a
171 list anyway. If so, then only the last element of the list will be
176 The value returned by the I<call_*> function indicates how many
177 items have been returned by the Perl subroutine - in this case it will
180 If 0, then you have specified the G_DISCARD flag.
182 If 1, then the item actually returned by the Perl subroutine will be
183 stored on the Perl stack - the section I<Returning a Scalar> shows how
184 to access this value on the stack. Remember that regardless of how
185 many items the Perl subroutine returns, only the last one will be
186 accessible from the stack - think of the case where only one value is
187 returned as being a list with only one element. Any other items that
188 were returned will not exist by the time control returns from the
189 I<call_*> function. The section I<Returning a list in a scalar
190 context> shows an example of this behavior.
195 Calls the Perl subroutine in a list context.
197 As with G_SCALAR, this flag has 2 effects:
203 It indicates to the subroutine being called that it is executing in a
204 list context (if it executes I<wantarray> the result will be true).
208 It ensures that all items returned from the subroutine will be
209 accessible when control returns from the I<call_*> function.
213 The value returned by the I<call_*> function indicates how many
214 items have been returned by the Perl subroutine.
216 If 0, then you have specified the G_DISCARD flag.
218 If not 0, then it will be a count of the number of items returned by
219 the subroutine. These items will be stored on the Perl stack. The
220 section I<Returning a list of values> gives an example of using the
221 G_ARRAY flag and the mechanics of accessing the returned items from the
226 By default, the I<call_*> functions place the items returned from
227 by the Perl subroutine on the stack. If you are not interested in
228 these items, then setting this flag will make Perl get rid of them
229 automatically for you. Note that it is still possible to indicate a
230 context to the Perl subroutine by using either G_SCALAR or G_ARRAY.
232 If you do not set this flag then it is I<very> important that you make
233 sure that any temporaries (i.e., parameters passed to the Perl
234 subroutine and values returned from the subroutine) are disposed of
235 yourself. The section I<Returning a Scalar> gives details of how to
236 dispose of these temporaries explicitly and the section I<Using Perl to
237 dispose of temporaries> discusses the specific circumstances where you
238 can ignore the problem and let Perl deal with it for you.
242 Whenever a Perl subroutine is called using one of the I<call_*>
243 functions, it is assumed by default that parameters are to be passed to
244 the subroutine. If you are not passing any parameters to the Perl
245 subroutine, you can save a bit of time by setting this flag. It has
246 the effect of not creating the C<@_> array for the Perl subroutine.
248 Although the functionality provided by this flag may seem
249 straightforward, it should be used only if there is a good reason to do
250 so. The reason for being cautious is that, even if you have specified
251 the G_NOARGS flag, it is still possible for the Perl subroutine that
252 has been called to think that you have passed it parameters.
254 In fact, what can happen is that the Perl subroutine you have called
255 can access the C<@_> array from a previous Perl subroutine. This will
256 occur when the code that is executing the I<call_*> function has
257 itself been called from another Perl subroutine. The code below
272 What has happened is that C<fred> accesses the C<@_> array which
278 It is possible for the Perl subroutine you are calling to terminate
279 abnormally, e.g., by calling I<die> explicitly or by not actually
280 existing. By default, when either of these events occurs, the
281 process will terminate immediately. If you want to trap this
282 type of event, specify the G_EVAL flag. It will put an I<eval { }>
283 around the subroutine call.
285 Whenever control returns from the I<call_*> function you need to
286 check the C<$@> variable as you would in a normal Perl script.
288 The value returned from the I<call_*> function is dependent on
289 what other flags have been specified and whether an error has
290 occurred. Here are all the different cases that can occur:
296 If the I<call_*> function returns normally, then the value
297 returned is as specified in the previous sections.
301 If G_DISCARD is specified, the return value will always be 0.
305 If G_ARRAY is specified I<and> an error has occurred, the return value
310 If G_SCALAR is specified I<and> an error has occurred, the return value
311 will be 1 and the value on the top of the stack will be I<undef>. This
312 means that if you have already detected the error by checking C<$@> and
313 you want the program to continue, you must remember to pop the I<undef>
318 See I<Using G_EVAL> for details on using G_EVAL.
322 Using the G_EVAL flag described above will always set C<$@>: clearing
323 it if there was no error, and setting it to describe the error if there
324 was an error in the called code. This is what you want if your intention
325 is to handle possible errors, but sometimes you just want to trap errors
326 and stop them interfering with the rest of the program.
328 This scenario will mostly be applicable to code that is meant to be called
329 from within destructors, asynchronous callbacks, and signal handlers.
330 In such situations, where the code being called has little relation to the
331 surrounding dynamic context, the main program needs to be insulated from
332 errors in the called code, even if they can't be handled intelligently.
333 It may also be useful to do this with code for C<__DIE__> or C<__WARN__>
334 hooks, and C<tie> functions.
336 The G_KEEPERR flag is meant to be used in conjunction with G_EVAL in
337 I<call_*> functions that are used to implement such code, or with
338 C<eval_sv>. This flag has no effect on the C<call_*> functions when
341 When G_KEEPERR is used, any error in the called code will terminate the
342 call as usual, and the error will not propagate beyond the call (as usual
343 for G_EVAL), but it will not go into C<$@>. Instead the error will be
344 converted into a warning, prefixed with the string "\t(in cleanup)".
345 This can be disabled using C<no warnings 'misc'>. If there is no error,
346 C<$@> will not be cleared.
348 Note that the G_KEEPERR flag does not propagate into inner evals; these
351 The G_KEEPERR flag was introduced in Perl version 5.002.
353 See I<Using G_KEEPERR> for an example of a situation that warrants the
356 =head2 Determining the Context
358 As mentioned above, you can determine the context of the currently
359 executing subroutine in Perl with I<wantarray>. The equivalent test
360 can be made in C by using the C<GIMME_V> macro, which returns
361 C<G_ARRAY> if you have been called in a list context, C<G_SCALAR> if
362 in a scalar context, or C<G_VOID> if in a void context (i.e., the
363 return value will not be used). An older version of this macro is
364 called C<GIMME>; in a void context it returns C<G_SCALAR> instead of
365 C<G_VOID>. An example of using the C<GIMME_V> macro is shown in
366 section I<Using GIMME_V>.
370 Enough of the definition talk! Let's have a few examples.
372 Perl provides many macros to assist in accessing the Perl stack.
373 Wherever possible, these macros should always be used when interfacing
374 to Perl internals. We hope this should make the code less vulnerable
375 to any changes made to Perl in the future.
377 Another point worth noting is that in the first series of examples I
378 have made use of only the I<call_pv> function. This has been done
379 to keep the code simpler and ease you into the topic. Wherever
380 possible, if the choice is between using I<call_pv> and
381 I<call_sv>, you should always try to use I<call_sv>. See
382 I<Using call_sv> for details.
384 =head2 No Parameters, Nothing Returned
386 This first trivial example will call a Perl subroutine, I<PrintUID>, to
387 print out the UID of the process.
394 and here is a C function to call it
402 call_pv("PrintUID", G_DISCARD|G_NOARGS);
407 A few points to note about this example:
413 Ignore C<dSP> and C<PUSHMARK(SP)> for now. They will be discussed in
418 We aren't passing any parameters to I<PrintUID> so G_NOARGS can be
423 We aren't interested in anything returned from I<PrintUID>, so
424 G_DISCARD is specified. Even if I<PrintUID> was changed to
425 return some value(s), having specified G_DISCARD will mean that they
426 will be wiped by the time control returns from I<call_pv>.
430 As I<call_pv> is being used, the Perl subroutine is specified as a
431 C string. In this case the subroutine name has been 'hard-wired' into the
436 Because we specified G_DISCARD, it is not necessary to check the value
437 returned from I<call_pv>. It will always be 0.
441 =head2 Passing Parameters
443 Now let's make a slightly more complex example. This time we want to
444 call a Perl subroutine, C<LeftString>, which will take 2 parameters--a
445 string ($s) and an integer ($n). The subroutine will simply
446 print the first $n characters of the string.
448 So the Perl subroutine would look like this:
453 print substr($s, 0, $n), "\n";
456 The C function required to call I<LeftString> would look like this:
459 call_LeftString(a, b)
469 XPUSHs(sv_2mortal(newSVpv(a, 0)));
470 XPUSHs(sv_2mortal(newSViv(b)));
473 call_pv("LeftString", G_DISCARD);
479 Here are a few notes on the C function I<call_LeftString>.
485 Parameters are passed to the Perl subroutine using the Perl stack.
486 This is the purpose of the code beginning with the line C<dSP> and
487 ending with the line C<PUTBACK>. The C<dSP> declares a local copy
488 of the stack pointer. This local copy should B<always> be accessed
493 If you are going to put something onto the Perl stack, you need to know
494 where to put it. This is the purpose of the macro C<dSP>--it declares
495 and initializes a I<local> copy of the Perl stack pointer.
497 All the other macros which will be used in this example require you to
498 have used this macro.
500 The exception to this rule is if you are calling a Perl subroutine
501 directly from an XSUB function. In this case it is not necessary to
502 use the C<dSP> macro explicitly--it will be declared for you
507 Any parameters to be pushed onto the stack should be bracketed by the
508 C<PUSHMARK> and C<PUTBACK> macros. The purpose of these two macros, in
509 this context, is to count the number of parameters you are
510 pushing automatically. Then whenever Perl is creating the C<@_> array for the
511 subroutine, it knows how big to make it.
513 The C<PUSHMARK> macro tells Perl to make a mental note of the current
514 stack pointer. Even if you aren't passing any parameters (like the
515 example shown in the section I<No Parameters, Nothing Returned>) you
516 must still call the C<PUSHMARK> macro before you can call any of the
517 I<call_*> functions--Perl still needs to know that there are no
520 The C<PUTBACK> macro sets the global copy of the stack pointer to be
521 the same as our local copy. If we didn't do this, I<call_pv>
522 wouldn't know where the two parameters we pushed were--remember that
523 up to now all the stack pointer manipulation we have done is with our
524 local copy, I<not> the global copy.
528 Next, we come to XPUSHs. This is where the parameters actually get
529 pushed onto the stack. In this case we are pushing a string and an
532 See L<perlguts/"XSUBs and the Argument Stack"> for details
533 on how the XPUSH macros work.
537 Because we created temporary values (by means of sv_2mortal() calls)
538 we will have to tidy up the Perl stack and dispose of mortal SVs.
540 This is the purpose of
545 at the start of the function, and
550 at the end. The C<ENTER>/C<SAVETMPS> pair creates a boundary for any
551 temporaries we create. This means that the temporaries we get rid of
552 will be limited to those which were created after these calls.
554 The C<FREETMPS>/C<LEAVE> pair will get rid of any values returned by
555 the Perl subroutine (see next example), plus it will also dump the
556 mortal SVs we have created. Having C<ENTER>/C<SAVETMPS> at the
557 beginning of the code makes sure that no other mortals are destroyed.
559 Think of these macros as working a bit like C<{> and C<}> in Perl
560 to limit the scope of local variables.
562 See the section I<Using Perl to Dispose of Temporaries> for details of
563 an alternative to using these macros.
567 Finally, I<LeftString> can now be called via the I<call_pv> function.
568 The only flag specified this time is G_DISCARD. Because we are passing
569 2 parameters to the Perl subroutine this time, we have not specified
574 =head2 Returning a Scalar
576 Now for an example of dealing with the items returned from a Perl
579 Here is a Perl subroutine, I<Adder>, that takes 2 integer parameters
580 and simply returns their sum.
588 Because we are now concerned with the return value from I<Adder>, the C
589 function required to call it is now a bit more complex.
603 XPUSHs(sv_2mortal(newSViv(a)));
604 XPUSHs(sv_2mortal(newSViv(b)));
607 count = call_pv("Adder", G_SCALAR);
612 croak("Big trouble\n");
614 printf ("The sum of %d and %d is %d\n", a, b, POPi);
621 Points to note this time are
627 The only flag specified this time was G_SCALAR. That means that the C<@_>
628 array will be created and that the value returned by I<Adder> will
629 still exist after the call to I<call_pv>.
633 The purpose of the macro C<SPAGAIN> is to refresh the local copy of the
634 stack pointer. This is necessary because it is possible that the memory
635 allocated to the Perl stack has been reallocated during the
638 If you are making use of the Perl stack pointer in your code you must
639 always refresh the local copy using SPAGAIN whenever you make use
640 of the I<call_*> functions or any other Perl internal function.
644 Although only a single value was expected to be returned from I<Adder>,
645 it is still good practice to check the return code from I<call_pv>
648 Expecting a single value is not quite the same as knowing that there
649 will be one. If someone modified I<Adder> to return a list and we
650 didn't check for that possibility and take appropriate action the Perl
651 stack would end up in an inconsistent state. That is something you
652 I<really> don't want to happen ever.
656 The C<POPi> macro is used here to pop the return value from the stack.
657 In this case we wanted an integer, so C<POPi> was used.
660 Here is the complete list of POP macros available, along with the types
671 The final C<PUTBACK> is used to leave the Perl stack in a consistent
672 state before exiting the function. This is necessary because when we
673 popped the return value from the stack with C<POPi> it updated only our
674 local copy of the stack pointer. Remember, C<PUTBACK> sets the global
675 stack pointer to be the same as our local copy.
680 =head2 Returning a List of Values
682 Now, let's extend the previous example to return both the sum of the
683 parameters and the difference.
685 Here is the Perl subroutine
693 and this is the C function
696 call_AddSubtract(a, b)
707 XPUSHs(sv_2mortal(newSViv(a)));
708 XPUSHs(sv_2mortal(newSViv(b)));
711 count = call_pv("AddSubtract", G_ARRAY);
716 croak("Big trouble\n");
718 printf ("%d - %d = %d\n", a, b, POPi);
719 printf ("%d + %d = %d\n", a, b, POPi);
726 If I<call_AddSubtract> is called like this
728 call_AddSubtract(7, 4);
730 then here is the output
741 We wanted list context, so G_ARRAY was used.
745 Not surprisingly C<POPi> is used twice this time because we were
746 retrieving 2 values from the stack. The important thing to note is that
747 when using the C<POP*> macros they come off the stack in I<reverse>
752 =head2 Returning a List in a Scalar Context
754 Say the Perl subroutine in the previous section was called in a scalar
758 call_AddSubScalar(a, b)
770 XPUSHs(sv_2mortal(newSViv(a)));
771 XPUSHs(sv_2mortal(newSViv(b)));
774 count = call_pv("AddSubtract", G_SCALAR);
778 printf ("Items Returned = %d\n", count);
780 for (i = 1; i <= count; ++i)
781 printf ("Value %d = %d\n", i, POPi);
788 The other modification made is that I<call_AddSubScalar> will print the
789 number of items returned from the Perl subroutine and their value (for
790 simplicity it assumes that they are integer). So if
791 I<call_AddSubScalar> is called
793 call_AddSubScalar(7, 4);
795 then the output will be
800 In this case the main point to note is that only the last item in the
801 list is returned from the subroutine. I<AddSubtract> actually made it back to
802 I<call_AddSubScalar>.
805 =head2 Returning Data from Perl via the Parameter List
807 It is also possible to return values directly via the parameter
808 list--whether it is actually desirable to do it is another matter entirely.
810 The Perl subroutine, I<Inc>, below takes 2 parameters and increments
819 and here is a C function to call it.
834 sva = sv_2mortal(newSViv(a));
835 svb = sv_2mortal(newSViv(b));
842 count = call_pv("Inc", G_DISCARD);
845 croak ("call_Inc: expected 0 values from 'Inc', got %d\n",
848 printf ("%d + 1 = %d\n", a, SvIV(sva));
849 printf ("%d + 1 = %d\n", b, SvIV(svb));
855 To be able to access the two parameters that were pushed onto the stack
856 after they return from I<call_pv> it is necessary to make a note
857 of their addresses--thus the two variables C<sva> and C<svb>.
859 The reason this is necessary is that the area of the Perl stack which
860 held them will very likely have been overwritten by something else by
861 the time control returns from I<call_pv>.
868 Now an example using G_EVAL. Below is a Perl subroutine which computes
869 the difference of its 2 parameters. If this would result in a negative
870 result, the subroutine calls I<die>.
876 die "death can be fatal\n" if $a < $b;
881 and some C to call it
895 XPUSHs(sv_2mortal(newSViv(a)));
896 XPUSHs(sv_2mortal(newSViv(b)));
899 count = call_pv("Subtract", G_EVAL|G_SCALAR);
903 /* Check the eval first */
906 printf ("Uh oh - %s\n", SvPV_nolen(ERRSV));
912 croak("call_Subtract: wanted 1 value from 'Subtract', got %d\n",
915 printf ("%d - %d = %d\n", a, b, POPi);
923 If I<call_Subtract> is called thus
927 the following will be printed
929 Uh oh - death can be fatal
937 We want to be able to catch the I<die> so we have used the G_EVAL
938 flag. Not specifying this flag would mean that the program would
939 terminate immediately at the I<die> statement in the subroutine
948 printf ("Uh oh - %s\n", SvPV_nolen(ERRSV));
952 is the direct equivalent of this bit of Perl
954 print "Uh oh - $@\n" if $@;
956 C<PL_errgv> is a perl global of type C<GV *> that points to the
957 symbol table entry containing the error. C<ERRSV> therefore
958 refers to the C equivalent of C<$@>.
962 Note that the stack is popped using C<POPs> in the block where
963 C<SvTRUE(ERRSV)> is true. This is necessary because whenever a
964 I<call_*> function invoked with G_EVAL|G_SCALAR returns an error,
965 the top of the stack holds the value I<undef>. Because we want the
966 program to continue after detecting this error, it is essential that
967 the stack be tidied up by removing the I<undef>.
972 =head2 Using G_KEEPERR
974 Consider this rather facetious example, where we have used an XS
975 version of the call_Subtract example above inside a destructor:
978 sub new { bless {}, $_[0] }
981 die "death can be fatal" if $a < $b;
984 sub DESTROY { call_Subtract(5, 4); }
985 sub foo { die "foo dies"; }
992 print "Saw: $@" if $@; # should be, but isn't
994 This example will fail to recognize that an error occurred inside the
995 C<eval {}>. Here's why: the call_Subtract code got executed while perl
996 was cleaning up temporaries when exiting the outer braced block, and because
997 call_Subtract is implemented with I<call_pv> using the G_EVAL
998 flag, it promptly reset C<$@>. This results in the failure of the
999 outermost test for C<$@>, and thereby the failure of the error trap.
1001 Appending the G_KEEPERR flag, so that the I<call_pv> call in
1002 call_Subtract reads:
1004 count = call_pv("Subtract", G_EVAL|G_SCALAR|G_KEEPERR);
1006 will preserve the error and restore reliable error handling.
1008 =head2 Using call_sv
1010 In all the previous examples I have 'hard-wired' the name of the Perl
1011 subroutine to be called from C. Most of the time though, it is more
1012 convenient to be able to specify the name of the Perl subroutine from
1013 within the Perl script.
1015 Consider the Perl code below
1019 print "Hello there\n";
1024 Here is a snippet of XSUB which defines I<CallSubPV>.
1031 call_pv(name, G_DISCARD|G_NOARGS);
1033 That is fine as far as it goes. The thing is, the Perl subroutine
1034 can be specified as only a string, however, Perl allows references
1035 to subroutines and anonymous subroutines.
1036 This is where I<call_sv> is useful.
1038 The code below for I<CallSubSV> is identical to I<CallSubPV> except
1039 that the C<name> parameter is now defined as an SV* and we use
1040 I<call_sv> instead of I<call_pv>.
1047 call_sv(name, G_DISCARD|G_NOARGS);
1049 Because we are using an SV to call I<fred> the following can all be used:
1055 CallSubSV( sub { print "Hello there\n" } );
1057 As you can see, I<call_sv> gives you much greater flexibility in
1058 how you can specify the Perl subroutine.
1060 You should note that, if it is necessary to store the SV (C<name> in the
1061 example above) which corresponds to the Perl subroutine so that it can
1062 be used later in the program, it not enough just to store a copy of the
1063 pointer to the SV. Say the code above had been like this:
1065 static SV * rememberSub;
1077 call_sv(rememberSub, G_DISCARD|G_NOARGS);
1079 The reason this is wrong is that, by the time you come to use the
1080 pointer C<rememberSub> in C<CallSavedSub1>, it may or may not still refer
1081 to the Perl subroutine that was recorded in C<SaveSub1>. This is
1082 particularly true for these cases:
1087 SaveSub1( sub { print "Hello there\n" } );
1090 By the time each of the C<SaveSub1> statements above has been executed,
1091 the SV*s which corresponded to the parameters will no longer exist.
1092 Expect an error message from Perl of the form
1094 Can't use an undefined value as a subroutine reference at ...
1096 for each of the C<CallSavedSub1> lines.
1098 Similarly, with this code
1105 you can expect one of these messages (which you actually get is dependent on
1106 the version of Perl you are using)
1108 Not a CODE reference at ...
1109 Undefined subroutine &main::47 called ...
1111 The variable $ref may have referred to the subroutine C<fred>
1112 whenever the call to C<SaveSub1> was made but by the time
1113 C<CallSavedSub1> gets called it now holds the number C<47>. Because we
1114 saved only a pointer to the original SV in C<SaveSub1>, any changes to
1115 $ref will be tracked by the pointer C<rememberSub>. This means that
1116 whenever C<CallSavedSub1> gets called, it will attempt to execute the
1117 code which is referenced by the SV* C<rememberSub>. In this case
1118 though, it now refers to the integer C<47>, so expect Perl to complain
1121 A similar but more subtle problem is illustrated with this code:
1128 This time whenever C<CallSavedSub1> gets called it will execute the Perl
1129 subroutine C<joe> (assuming it exists) rather than C<fred> as was
1130 originally requested in the call to C<SaveSub1>.
1132 To get around these problems it is necessary to take a full copy of the
1133 SV. The code below shows C<SaveSub2> modified to do that.
1135 static SV * keepSub = (SV*)NULL;
1141 /* Take a copy of the callback */
1142 if (keepSub == (SV*)NULL)
1143 /* First time, so create a new SV */
1144 keepSub = newSVsv(name);
1146 /* Been here before, so overwrite */
1147 SvSetSV(keepSub, name);
1153 call_sv(keepSub, G_DISCARD|G_NOARGS);
1155 To avoid creating a new SV every time C<SaveSub2> is called,
1156 the function first checks to see if it has been called before. If not,
1157 then space for a new SV is allocated and the reference to the Perl
1158 subroutine C<name> is copied to the variable C<keepSub> in one
1159 operation using C<newSVsv>. Thereafter, whenever C<SaveSub2> is called,
1160 the existing SV, C<keepSub>, is overwritten with the new value using
1163 =head2 Using call_argv
1165 Here is a Perl subroutine which prints whatever parameters are passed
1172 foreach (@list) { print "$_\n" }
1175 And here is an example of I<call_argv> which will call
1178 static char * words[] = {"alpha", "beta", "gamma", "delta", NULL};
1185 call_argv("PrintList", G_DISCARD, words);
1188 Note that it is not necessary to call C<PUSHMARK> in this instance.
1189 This is because I<call_argv> will do it for you.
1191 =head2 Using call_method
1193 Consider the following Perl code:
1206 my ($self, $index) = @_;
1207 print "$index: $$self[$index]\n";
1213 print "This is Class $class version 1.0\n";
1217 It implements just a very simple class to manage an array. Apart from
1218 the constructor, C<new>, it declares methods, one static and one
1219 virtual. The static method, C<PrintID>, prints out simply the class
1220 name and a version number. The virtual method, C<Display>, prints out a
1221 single element of the array. Here is an all-Perl example of using it.
1223 $a = Mine->new('red', 'green', 'blue');
1230 This is Class Mine version 1.0
1232 Calling a Perl method from C is fairly straightforward. The following
1233 things are required:
1239 A reference to the object for a virtual method or the name of the class
1244 The name of the method
1248 Any other parameters specific to the method
1252 Here is a simple XSUB which illustrates the mechanics of calling both
1253 the C<PrintID> and C<Display> methods from C.
1256 call_Method(ref, method, index)
1263 XPUSHs(sv_2mortal(newSViv(index)));
1266 call_method(method, G_DISCARD);
1269 call_PrintID(class, method)
1274 XPUSHs(sv_2mortal(newSVpv(class, 0)));
1277 call_method(method, G_DISCARD);
1280 So the methods C<PrintID> and C<Display> can be invoked like this:
1282 $a = Mine->new('red', 'green', 'blue');
1283 call_Method($a, 'Display', 1);
1284 call_PrintID('Mine', 'PrintID');
1286 The only thing to note is that, in both the static and virtual methods,
1287 the method name is not passed via the stack--it is used as the first
1288 parameter to I<call_method>.
1290 =head2 Using GIMME_V
1292 Here is a trivial XSUB which prints the context in which it is
1293 currently executing.
1298 I32 gimme = GIMME_V;
1299 if (gimme == G_VOID)
1300 printf ("Context is Void\n");
1301 else if (gimme == G_SCALAR)
1302 printf ("Context is Scalar\n");
1304 printf ("Context is Array\n");
1306 And here is some Perl to test it.
1312 The output from that will be
1318 =head2 Using Perl to Dispose of Temporaries
1320 In the examples given to date, any temporaries created in the callback
1321 (i.e., parameters passed on the stack to the I<call_*> function or
1322 values returned via the stack) have been freed by one of these methods:
1328 Specifying the G_DISCARD flag with I<call_*>
1332 Explicitly using the C<ENTER>/C<SAVETMPS>--C<FREETMPS>/C<LEAVE> pairing
1336 There is another method which can be used, namely letting Perl do it
1337 for you automatically whenever it regains control after the callback
1338 has terminated. This is done by simply not using the
1346 sequence in the callback (and not, of course, specifying the G_DISCARD
1349 If you are going to use this method you have to be aware of a possible
1350 memory leak which can arise under very specific circumstances. To
1351 explain these circumstances you need to know a bit about the flow of
1352 control between Perl and the callback routine.
1354 The examples given at the start of the document (an error handler and
1355 an event driven program) are typical of the two main sorts of flow
1356 control that you are likely to encounter with callbacks. There is a
1357 very important distinction between them, so pay attention.
1359 In the first example, an error handler, the flow of control could be as
1360 follows. You have created an interface to an external library.
1361 Control can reach the external library like this
1363 perl --> XSUB --> external library
1365 Whilst control is in the library, an error condition occurs. You have
1366 previously set up a Perl callback to handle this situation, so it will
1367 get executed. Once the callback has finished, control will drop back to
1368 Perl again. Here is what the flow of control will be like in that
1371 perl --> XSUB --> external library
1375 external library --> call_* --> perl
1377 perl <-- XSUB <-- external library <-- call_* <----+
1379 After processing of the error using I<call_*> is completed,
1380 control reverts back to Perl more or less immediately.
1382 In the diagram, the further right you go the more deeply nested the
1383 scope is. It is only when control is back with perl on the extreme
1384 left of the diagram that you will have dropped back to the enclosing
1385 scope and any temporaries you have left hanging around will be freed.
1387 In the second example, an event driven program, the flow of control
1388 will be more like this
1390 perl --> XSUB --> event handler
1392 event handler --> call_* --> perl
1394 event handler <-- call_* <----+
1396 event handler --> call_* --> perl
1398 event handler <-- call_* <----+
1400 event handler --> call_* --> perl
1402 event handler <-- call_* <----+
1404 In this case the flow of control can consist of only the repeated
1407 event handler --> call_* --> perl
1409 for practically the complete duration of the program. This means that
1410 control may I<never> drop back to the surrounding scope in Perl at the
1413 So what is the big problem? Well, if you are expecting Perl to tidy up
1414 those temporaries for you, you might be in for a long wait. For Perl
1415 to dispose of your temporaries, control must drop back to the
1416 enclosing scope at some stage. In the event driven scenario that may
1417 never happen. This means that, as time goes on, your program will
1418 create more and more temporaries, none of which will ever be freed. As
1419 each of these temporaries consumes some memory your program will
1420 eventually consume all the available memory in your system--kapow!
1422 So here is the bottom line--if you are sure that control will revert
1423 back to the enclosing Perl scope fairly quickly after the end of your
1424 callback, then it isn't absolutely necessary to dispose explicitly of
1425 any temporaries you may have created. Mind you, if you are at all
1426 uncertain about what to do, it doesn't do any harm to tidy up anyway.
1429 =head2 Strategies for Storing Callback Context Information
1432 Potentially one of the trickiest problems to overcome when designing a
1433 callback interface can be figuring out how to store the mapping between
1434 the C callback function and the Perl equivalent.
1436 To help understand why this can be a real problem first consider how a
1437 callback is set up in an all C environment. Typically a C API will
1438 provide a function to register a callback. This will expect a pointer
1439 to a function as one of its parameters. Below is a call to a
1440 hypothetical function C<register_fatal> which registers the C function
1441 to get called when a fatal error occurs.
1443 register_fatal(cb1);
1445 The single parameter C<cb1> is a pointer to a function, so you must
1446 have defined C<cb1> in your code, say something like this
1451 printf ("Fatal Error\n");
1455 Now change that to call a Perl subroutine instead
1457 static SV * callback = (SV*)NULL;
1466 /* Call the Perl sub to process the callback */
1467 call_sv(callback, G_DISCARD);
1475 /* Remember the Perl sub */
1476 if (callback == (SV*)NULL)
1477 callback = newSVsv(fn);
1479 SvSetSV(callback, fn);
1481 /* register the callback with the external library */
1482 register_fatal(cb1);
1484 where the Perl equivalent of C<register_fatal> and the callback it
1485 registers, C<pcb1>, might look like this
1487 # Register the sub pcb1
1488 register_fatal(\&pcb1);
1492 die "I'm dying...\n";
1495 The mapping between the C callback and the Perl equivalent is stored in
1496 the global variable C<callback>.
1498 This will be adequate if you ever need to have only one callback
1499 registered at any time. An example could be an error handler like the
1500 code sketched out above. Remember though, repeated calls to
1501 C<register_fatal> will replace the previously registered callback
1502 function with the new one.
1504 Say for example you want to interface to a library which allows asynchronous
1505 file i/o. In this case you may be able to register a callback whenever
1506 a read operation has completed. To be of any use we want to be able to
1507 call separate Perl subroutines for each file that is opened. As it
1508 stands, the error handler example above would not be adequate as it
1509 allows only a single callback to be defined at any time. What we
1510 require is a means of storing the mapping between the opened file and
1511 the Perl subroutine we want to be called for that file.
1513 Say the i/o library has a function C<asynch_read> which associates a C
1514 function C<ProcessRead> with a file handle C<fh>--this assumes that it
1515 has also provided some routine to open the file and so obtain the file
1518 asynch_read(fh, ProcessRead)
1520 This may expect the C I<ProcessRead> function of this form
1523 ProcessRead(fh, buffer)
1530 To provide a Perl interface to this library we need to be able to map
1531 between the C<fh> parameter and the Perl subroutine we want called. A
1532 hash is a convenient mechanism for storing this mapping. The code
1533 below shows a possible implementation
1535 static HV * Mapping = (HV*)NULL;
1538 asynch_read(fh, callback)
1542 /* If the hash doesn't already exist, create it */
1543 if (Mapping == (HV*)NULL)
1546 /* Save the fh -> callback mapping */
1547 hv_store(Mapping, (char*)&fh, sizeof(fh), newSVsv(callback), 0);
1549 /* Register with the C Library */
1550 asynch_read(fh, asynch_read_if);
1552 and C<asynch_read_if> could look like this
1555 asynch_read_if(fh, buffer)
1562 /* Get the callback associated with fh */
1563 sv = hv_fetch(Mapping, (char*)&fh , sizeof(fh), FALSE);
1564 if (sv == (SV**)NULL)
1565 croak("Internal error...\n");
1568 XPUSHs(sv_2mortal(newSViv(fh)));
1569 XPUSHs(sv_2mortal(newSVpv(buffer, 0)));
1572 /* Call the Perl sub */
1573 call_sv(*sv, G_DISCARD);
1576 For completeness, here is C<asynch_close>. This shows how to remove
1577 the entry from the hash C<Mapping>.
1583 /* Remove the entry from the hash */
1584 (void) hv_delete(Mapping, (char*)&fh, sizeof(fh), G_DISCARD);
1586 /* Now call the real asynch_close */
1589 So the Perl interface would look like this
1593 my($handle, $buffer) = @_;
1596 # Register the Perl callback
1597 asynch_read($fh, \&callback1);
1601 The mapping between the C callback and Perl is stored in the global
1602 hash C<Mapping> this time. Using a hash has the distinct advantage that
1603 it allows an unlimited number of callbacks to be registered.
1605 What if the interface provided by the C callback doesn't contain a
1606 parameter which allows the file handle to Perl subroutine mapping? Say
1607 in the asynchronous i/o package, the callback function gets passed only
1608 the C<buffer> parameter like this
1617 Without the file handle there is no straightforward way to map from the
1618 C callback to the Perl subroutine.
1620 In this case a possible way around this problem is to predefine a
1621 series of C functions to act as the interface to Perl, thus
1624 #define NULL_HANDLE -1
1625 typedef void (*FnMap)();
1637 static struct MapStruct Map [MAX_CB] =
1639 { fn1, NULL, NULL_HANDLE },
1640 { fn2, NULL, NULL_HANDLE },
1641 { fn3, NULL, NULL_HANDLE }
1652 XPUSHs(sv_2mortal(newSVpv(buffer, 0)));
1655 /* Call the Perl sub */
1656 call_sv(Map[index].PerlSub, G_DISCARD);
1681 array_asynch_read(fh, callback)
1686 int null_index = MAX_CB;
1688 /* Find the same handle or an empty entry */
1689 for (index = 0; index < MAX_CB; ++index)
1691 if (Map[index].Handle == fh)
1694 if (Map[index].Handle == NULL_HANDLE)
1698 if (index == MAX_CB && null_index == MAX_CB)
1699 croak ("Too many callback functions registered\n");
1701 if (index == MAX_CB)
1704 /* Save the file handle */
1705 Map[index].Handle = fh;
1707 /* Remember the Perl sub */
1708 if (Map[index].PerlSub == (SV*)NULL)
1709 Map[index].PerlSub = newSVsv(callback);
1711 SvSetSV(Map[index].PerlSub, callback);
1713 asynch_read(fh, Map[index].Function);
1716 array_asynch_close(fh)
1721 /* Find the file handle */
1722 for (index = 0; index < MAX_CB; ++ index)
1723 if (Map[index].Handle == fh)
1726 if (index == MAX_CB)
1727 croak ("could not close fh %d\n", fh);
1729 Map[index].Handle = NULL_HANDLE;
1730 SvREFCNT_dec(Map[index].PerlSub);
1731 Map[index].PerlSub = (SV*)NULL;
1735 In this case the functions C<fn1>, C<fn2>, and C<fn3> are used to
1736 remember the Perl subroutine to be called. Each of the functions holds
1737 a separate hard-wired index which is used in the function C<Pcb> to
1738 access the C<Map> array and actually call the Perl subroutine.
1740 There are some obvious disadvantages with this technique.
1742 Firstly, the code is considerably more complex than with the previous
1745 Secondly, there is a hard-wired limit (in this case 3) to the number of
1746 callbacks that can exist simultaneously. The only way to increase the
1747 limit is by modifying the code to add more functions and then
1748 recompiling. None the less, as long as the number of functions is
1749 chosen with some care, it is still a workable solution and in some
1750 cases is the only one available.
1752 To summarize, here are a number of possible methods for you to consider
1753 for storing the mapping between C and the Perl callback
1757 =item 1. Ignore the problem - Allow only 1 callback
1759 For a lot of situations, like interfacing to an error handler, this may
1760 be a perfectly adequate solution.
1762 =item 2. Create a sequence of callbacks - hard wired limit
1764 If it is impossible to tell from the parameters passed back from the C
1765 callback what the context is, then you may need to create a sequence of C
1766 callback interface functions, and store pointers to each in an array.
1768 =item 3. Use a parameter to map to the Perl callback
1770 A hash is an ideal mechanism to store the mapping between C and Perl.
1775 =head2 Alternate Stack Manipulation
1778 Although I have made use of only the C<POP*> macros to access values
1779 returned from Perl subroutines, it is also possible to bypass these
1780 macros and read the stack using the C<ST> macro (See L<perlxs> for a
1781 full description of the C<ST> macro).
1783 Most of the time the C<POP*> macros should be adequate; the main
1784 problem with them is that they force you to process the returned values
1785 in sequence. This may not be the most suitable way to process the
1786 values in some cases. What we want is to be able to access the stack in
1787 a random order. The C<ST> macro as used when coding an XSUB is ideal
1790 The code below is the example given in the section I<Returning a List
1791 of Values> recoded to use C<ST> instead of C<POP*>.
1794 call_AddSubtract2(a, b)
1806 XPUSHs(sv_2mortal(newSViv(a)));
1807 XPUSHs(sv_2mortal(newSViv(b)));
1810 count = call_pv("AddSubtract", G_ARRAY);
1814 ax = (SP - PL_stack_base) + 1;
1817 croak("Big trouble\n");
1819 printf ("%d + %d = %d\n", a, b, SvIV(ST(0)));
1820 printf ("%d - %d = %d\n", a, b, SvIV(ST(1)));
1833 Notice that it was necessary to define the variable C<ax>. This is
1834 because the C<ST> macro expects it to exist. If we were in an XSUB it
1835 would not be necessary to define C<ax> as it is already defined for
1844 ax = (SP - PL_stack_base) + 1;
1846 sets the stack up so that we can use the C<ST> macro.
1850 Unlike the original coding of this example, the returned
1851 values are not accessed in reverse order. So C<ST(0)> refers to the
1852 first value returned by the Perl subroutine and C<ST(count-1)>
1857 =head2 Creating and Calling an Anonymous Subroutine in C
1859 As we've already shown, C<call_sv> can be used to invoke an
1860 anonymous subroutine. However, our example showed a Perl script
1861 invoking an XSUB to perform this operation. Let's see how it can be
1862 done inside our C code:
1866 SV *cvrv = eval_pv("sub { print 'You will not find me cluttering any namespace!' }", TRUE);
1870 call_sv(cvrv, G_VOID|G_NOARGS);
1872 C<eval_pv> is used to compile the anonymous subroutine, which
1873 will be the return value as well (read more about C<eval_pv> in
1874 L<perlapi/eval_pv>). Once this code reference is in hand, it
1875 can be mixed in with all the previous examples we've shown.
1877 =head1 LIGHTWEIGHT CALLBACKS
1879 Sometimes you need to invoke the same subroutine repeatedly.
1880 This usually happens with a function that acts on a list of
1881 values, such as Perl's built-in sort(). You can pass a
1882 comparison function to sort(), which will then be invoked
1883 for every pair of values that needs to be compared. The first()
1884 and reduce() functions from L<List::Util> follow a similar
1887 In this case it is possible to speed up the routine (often
1888 quite substantially) by using the lightweight callback API.
1889 The idea is that the calling context only needs to be
1890 created and destroyed once, and the sub can be called
1891 arbitrarily many times in between.
1893 It is usual to pass parameters using global variables (typically
1894 $_ for one parameter, or $a and $b for two parameters) rather
1895 than via @_. (It is possible to use the @_ mechanism if you know
1896 what you're doing, though there is as yet no supported API for
1897 it. It's also inherently slower.)
1899 The pattern of macro calls is like this:
1901 dMULTICALL; /* Declare local variables */
1902 I32 gimme = G_SCALAR; /* context of the call: G_SCALAR,
1903 * G_ARRAY, or G_VOID */
1905 PUSH_MULTICALL(cv); /* Set up the context for calling cv,
1906 and set local vars appropriately */
1909 /* set the value(s) af your parameter variables */
1910 MULTICALL; /* Make the actual call */
1913 POP_MULTICALL; /* Tear down the calling context */
1915 For some concrete examples, see the implementation of the
1916 first() and reduce() functions of List::Util 1.18. There you
1917 will also find a header file that emulates the multicall API
1918 on older versions of perl.
1922 L<perlxs>, L<perlguts>, L<perlembed>
1928 Special thanks to the following people who assisted in the creation of
1931 Jeff Okamoto, Tim Bunce, Nick Gianniotis, Steve Kelem, Gurusamy Sarathy
1936 Version 1.3, 14th Apr 1997