Accessing GNU Modula-2 Built-ins#
This section describes the built-in constants and functions defined in
GNU Modula-2. The following compiler constants can be accessed using
the __ATTRIBUTE__ __BUILTIN__ keywords. These are not
part of the Modula-2 language and they may differ depending upon the
target architecture but they provide a method whereby common
libraries can interface to a different underlying architecture.
The built-in constants are: BITS_PER_UNIT, BITS_PER_WORD,
BITS_PER_CHAR and UNITS_PER_WORD. They are integrated into
GNU Modula-2 by an extension to the ConstFactor rule:
ConstFactor := ConstQualidentOrSet | Number | ConstString |
"(" ConstExpression ")" | "NOT" ConstFactor |
ConstAttribute =:
ConstAttribute := "__ATTRIBUTE__" "__BUILTIN__" "(" "(" Ident ")" ")" =:
Here is an example taken from the ISO library SYSTEM.def :
CONST
BITSPERLOC = __ATTRIBUTE__ __BUILTIN__ ((BITS_PER_UNIT)) ;
LOCSPERWORD = __ATTRIBUTE__ __BUILTIN__ ((UNITS_PER_WORD)) ;
Built-in functions are transparent to the end user. All built-in
functions are declared in DEFINITION MODULE s and are imported
as and when required. Built-in functions are declared in definition
modules by using the __BUILTIN__ keyword. Here is a section of
the ISO library LongMath.def which demonstrates this feature.
PROCEDURE __BUILTIN__ sqrt (x: LONGREAL): LONGREAL;
(* Returns the square root of x *)
This indicates that the function sqrt will be implemented using
the gcc built-in maths library. If gcc cannot utilise the built-in
function (for example if the programmer requested the address of
sqrt) then code is generated to call the alternative function
implemented in the IMPLEMENTATION MODULE.
Sometimes a function exported from the DEFINITION MODULE
will have a different name from the built-in function within gcc. In
such cases the mapping between the GNU Modula-2 function name and the
gcc name is expressed using the keywords __ATTRIBUTE__
__BUILTIN__ ((Ident)). For example the function
sqrt in LongMath.def maps onto the gcc built-in function
sqrtl and this is expressed as:
PROCEDURE __ATTRIBUTE__ __BUILTIN__ ((sqrtl)) sqrt
(x: LONGREAL) : LONGREAL;
(* Returns the positive square root of x *)
The following module Builtins.def enumerates the list of
built-in functions which can be accessed in GNU Modula-2. It also
serves to define the parameter and return value for each function:
DEFINITION MODULE Builtins ;
FROM SYSTEM IMPORT ADDRESS ;
(* floating point intrinsic procedure functions *)
isfinitef
PROCEDURE __BUILTIN__ isfinitef (x: SHORTREAL) : BOOLEAN ;
isfinite
PROCEDURE __BUILTIN__ isfinite (x: REAL) : BOOLEAN ;
isfinitel
PROCEDURE __BUILTIN__ isfinitel (x: LONGREAL) : BOOLEAN ;
sinf
PROCEDURE __BUILTIN__ sinf (x: SHORTREAL) : SHORTREAL ;
sin
PROCEDURE __BUILTIN__ sin (x: REAL) : REAL ;
sinl
PROCEDURE __BUILTIN__ sinl (x: LONGREAL) : LONGREAL ;
cosf
PROCEDURE __BUILTIN__ cosf (x: SHORTREAL) : SHORTREAL ;
cos
PROCEDURE __BUILTIN__ cos (x: REAL) : REAL ;
cosl
PROCEDURE __BUILTIN__ cosl (x: LONGREAL) : LONGREAL ;
sqrtf
PROCEDURE __BUILTIN__ sqrtf (x: SHORTREAL) : SHORTREAL ;
sqrt
PROCEDURE __BUILTIN__ sqrt (x: REAL) : REAL ;
sqrtl
PROCEDURE __BUILTIN__ sqrtl (x: LONGREAL) : LONGREAL ;
atan2f
PROCEDURE __BUILTIN__ atan2f (x, y: SHORTREAL) : SHORTREAL ;
atan2
PROCEDURE __BUILTIN__ atan2 (x, y: REAL) : REAL ;
atan2l
PROCEDURE __BUILTIN__ atan2l (x, y: LONGREAL) : LONGREAL ;
fabsf
PROCEDURE __BUILTIN__ fabsf (x: SHORTREAL) : SHORTREAL ;
fabs
PROCEDURE __BUILTIN__ fabs (x: REAL) : REAL ;
fabsl
PROCEDURE __BUILTIN__ fabsl (x: LONGREAL) : LONGREAL ;
logf
PROCEDURE __BUILTIN__ logf (x: SHORTREAL) : SHORTREAL ;
log
PROCEDURE __BUILTIN__ log (x: REAL) : REAL ;
logl
PROCEDURE __BUILTIN__ logl (x: LONGREAL) : LONGREAL ;
expf
PROCEDURE __BUILTIN__ expf (x: SHORTREAL) : SHORTREAL ;
exp
PROCEDURE __BUILTIN__ exp (x: REAL) : REAL ;
expl
PROCEDURE __BUILTIN__ expl (x: LONGREAL) : LONGREAL ;
log10f
PROCEDURE __BUILTIN__ log10f (x: SHORTREAL) : SHORTREAL ;
log10
PROCEDURE __BUILTIN__ log10 (x: REAL) : REAL ;
log10l
PROCEDURE __BUILTIN__ log10l (x: LONGREAL) : LONGREAL ;
exp10f
PROCEDURE __BUILTIN__ exp10f (x: SHORTREAL) : SHORTREAL ;
exp10
PROCEDURE __BUILTIN__ exp10 (x: REAL) : REAL ;
exp10l
PROCEDURE __BUILTIN__ exp10l (x: LONGREAL) : LONGREAL ;
ilogbf
PROCEDURE __BUILTIN__ ilogbf (x: SHORTREAL) : INTEGER ;
ilogb
PROCEDURE __BUILTIN__ ilogb (x: REAL) : INTEGER ;
ilogbl
PROCEDURE __BUILTIN__ ilogbl (x: LONGREAL) : INTEGER ;
huge_val
PROCEDURE __BUILTIN__ huge_val () : REAL ;
huge_valf
PROCEDURE __BUILTIN__ huge_valf () : SHORTREAL ;
huge_vall
PROCEDURE __BUILTIN__ huge_vall () : LONGREAL ;
significand
PROCEDURE __BUILTIN__ significand (r: REAL) : REAL ;
significandf
PROCEDURE __BUILTIN__ significandf (s: SHORTREAL) : SHORTREAL ;
significandl
PROCEDURE __BUILTIN__ significandl (l: LONGREAL) : LONGREAL ;
modf
PROCEDURE __BUILTIN__ modf (x: REAL; VAR y: REAL) : REAL ;
modff
PROCEDURE __BUILTIN__ modff (x: SHORTREAL;
VAR y: SHORTREAL) : SHORTREAL ;
modfl
PROCEDURE __BUILTIN__ modfl (x: LONGREAL; VAR y: LONGREAL) : LONGREAL ;
signbit
PROCEDURE __BUILTIN__ signbit (r: REAL) : INTEGER ;
signbitf
PROCEDURE __BUILTIN__ signbitf (s: SHORTREAL) : INTEGER ;
signbitl
PROCEDURE __BUILTIN__ signbitl (l: LONGREAL) : INTEGER ;
nextafter
PROCEDURE __BUILTIN__ nextafter (x, y: REAL) : REAL ;
nextafterf
PROCEDURE __BUILTIN__ nextafterf (x, y: SHORTREAL) : SHORTREAL ;
nextafterl
PROCEDURE __BUILTIN__ nextafterl (x, y: LONGREAL) : LONGREAL ;
nexttoward
PROCEDURE __BUILTIN__ nexttoward (x, y: REAL) : LONGREAL ;
nexttowardf
PROCEDURE __BUILTIN__ nexttowardf (x, y: SHORTREAL) : LONGREAL ;
nexttowardl
PROCEDURE __BUILTIN__ nexttowardl (x, y: LONGREAL) : LONGREAL ;
scalb
PROCEDURE __BUILTIN__ scalb (x, n: REAL) : REAL ;
scalbf
PROCEDURE __BUILTIN__ scalbf (x, n: SHORTREAL) : SHORTREAL ;
scalbl
PROCEDURE __BUILTIN__ scalbl (x, n: LONGREAL) : LONGREAL ;
scalbln
PROCEDURE __BUILTIN__ scalbln (x: REAL; n: LONGINT) : REAL ;
scalblnf
PROCEDURE __BUILTIN__ scalblnf (x: SHORTREAL; n: LONGINT) : SHORTREAL ;
scalblnl
PROCEDURE __BUILTIN__ scalblnl (x: LONGREAL; n: LONGINT) : LONGREAL ;
scalbn
PROCEDURE __BUILTIN__ scalbn (x: REAL; n: INTEGER) : REAL ;
scalbnf
PROCEDURE __BUILTIN__ scalbnf (x: SHORTREAL; n: INTEGER) : SHORTREAL ;
scalbnl
PROCEDURE __BUILTIN__ scalbnl (x: LONGREAL; n: INTEGER) : LONGREAL ;
(* complex arithmetic intrincic procedure functions *)
cabsf
PROCEDURE __BUILTIN__ cabsf (z: SHORTCOMPLEX) : SHORTREAL ;
cabs
PROCEDURE __BUILTIN__ cabs (z: COMPLEX) : REAL ;
cabsl
PROCEDURE __BUILTIN__ cabsl (z: LONGCOMPLEX) : LONGREAL ;
cargf
PROCEDURE __BUILTIN__ cargf (z: SHORTCOMPLEX) : SHORTREAL ;
carg
PROCEDURE __BUILTIN__ carg (z: COMPLEX) : REAL ;
cargl
PROCEDURE __BUILTIN__ cargl (z: LONGCOMPLEX) : LONGREAL ;
conjf
PROCEDURE __BUILTIN__ conjf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
conj
PROCEDURE __BUILTIN__ conj (z: COMPLEX) : COMPLEX ;
conjl
PROCEDURE __BUILTIN__ conjl (z: LONGCOMPLEX) : LONGCOMPLEX ;
cpowerf
PROCEDURE __BUILTIN__ cpowerf (base: SHORTCOMPLEX;
exp: SHORTREAL) : SHORTCOMPLEX ;
cpower
PROCEDURE __BUILTIN__ cpower (base: COMPLEX; exp: REAL) : COMPLEX ;
cpowerl
PROCEDURE __BUILTIN__ cpowerl (base: LONGCOMPLEX;
exp: LONGREAL) : LONGCOMPLEX ;
csqrtf
PROCEDURE __BUILTIN__ csqrtf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
csqrt
PROCEDURE __BUILTIN__ csqrt (z: COMPLEX) : COMPLEX ;
csqrtl
PROCEDURE __BUILTIN__ csqrtl (z: LONGCOMPLEX) : LONGCOMPLEX ;
cexpf
PROCEDURE __BUILTIN__ cexpf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
cexp
PROCEDURE __BUILTIN__ cexp (z: COMPLEX) : COMPLEX ;
cexpl
PROCEDURE __BUILTIN__ cexpl (z: LONGCOMPLEX) : LONGCOMPLEX ;
clnf
PROCEDURE __BUILTIN__ clnf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
cln
PROCEDURE __BUILTIN__ cln (z: COMPLEX) : COMPLEX ;
clnl
PROCEDURE __BUILTIN__ clnl (z: LONGCOMPLEX) : LONGCOMPLEX ;
csinf
PROCEDURE __BUILTIN__ csinf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
csin
PROCEDURE __BUILTIN__ csin (z: COMPLEX) : COMPLEX ;
csinl
PROCEDURE __BUILTIN__ csinl (z: LONGCOMPLEX) : LONGCOMPLEX ;
ccosf
PROCEDURE __BUILTIN__ ccosf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
ccos
PROCEDURE __BUILTIN__ ccos (z: COMPLEX) : COMPLEX ;
ccosl
PROCEDURE __BUILTIN__ ccosl (z: LONGCOMPLEX) : LONGCOMPLEX ;
ctanf
PROCEDURE __BUILTIN__ ctanf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
ctan
PROCEDURE __BUILTIN__ ctan (z: COMPLEX) : COMPLEX ;
ctanl
PROCEDURE __BUILTIN__ ctanl (z: LONGCOMPLEX) : LONGCOMPLEX ;
carcsinf
PROCEDURE __BUILTIN__ carcsinf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
carcsin
PROCEDURE __BUILTIN__ carcsin (z: COMPLEX) : COMPLEX ;
carcsinl
PROCEDURE __BUILTIN__ carcsinl (z: LONGCOMPLEX) : LONGCOMPLEX ;
carccosf
PROCEDURE __BUILTIN__ carccosf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
carccos
PROCEDURE __BUILTIN__ carccos (z: COMPLEX) : COMPLEX ;
carccosl
PROCEDURE __BUILTIN__ carccosl (z: LONGCOMPLEX) : LONGCOMPLEX ;
carctanf
PROCEDURE __BUILTIN__ carctanf (z: SHORTCOMPLEX) : SHORTCOMPLEX ;
carctan
PROCEDURE __BUILTIN__ carctan (z: COMPLEX) : COMPLEX ;
carctanl
PROCEDURE __BUILTIN__ carctanl (z: LONGCOMPLEX) : LONGCOMPLEX ;
(* memory and string intrincic procedure functions *)
alloca
PROCEDURE __BUILTIN__ alloca (i: CARDINAL) : ADDRESS ;
memcpy
PROCEDURE __BUILTIN__ memcpy (dest, src: ADDRESS;
nbytes: CARDINAL) : ADDRESS ;
index
PROCEDURE __BUILTIN__ index (s: ADDRESS; c: INTEGER) : ADDRESS ;
rindex
PROCEDURE __BUILTIN__ rindex (s: ADDRESS; c: INTEGER) : ADDRESS ;
memcmp
PROCEDURE __BUILTIN__ memcmp (s1, s2: ADDRESS;
nbytes: CARDINAL) : INTEGER ;
memset
PROCEDURE __BUILTIN__ memset (s: ADDRESS; c: INTEGER;
nbytes: CARDINAL) : ADDRESS ;
memmove
PROCEDURE __BUILTIN__ memmove (s1, s2: ADDRESS;
nbytes: CARDINAL) : ADDRESS ;
strcat
PROCEDURE __BUILTIN__ strcat (dest, src: ADDRESS) : ADDRESS ;
strncat
PROCEDURE __BUILTIN__ strncat (dest, src: ADDRESS;
nbytes: CARDINAL) : ADDRESS ;
strcpy
PROCEDURE __BUILTIN__ strcpy (dest, src: ADDRESS) : ADDRESS ;
strncpy
PROCEDURE __BUILTIN__ strncpy (dest, src: ADDRESS;
nbytes: CARDINAL) : ADDRESS ;
strcmp
PROCEDURE __BUILTIN__ strcmp (s1, s2: ADDRESS) : INTEGER ;
strncmp
PROCEDURE __BUILTIN__ strncmp (s1, s2: ADDRESS;
nbytes: CARDINAL) : INTEGER ;
strlen
PROCEDURE __BUILTIN__ strlen (s: ADDRESS) : INTEGER ;
strstr
PROCEDURE __BUILTIN__ strstr (haystack, needle: ADDRESS) : ADDRESS ;
strpbrk
PROCEDURE __BUILTIN__ strpbrk (s, accept: ADDRESS) : ADDRESS ;
strspn
PROCEDURE __BUILTIN__ strspn (s, accept: ADDRESS) : CARDINAL ;
strcspn
PROCEDURE __BUILTIN__ strcspn (s, accept: ADDRESS) : CARDINAL ;
strchr
PROCEDURE __BUILTIN__ strchr (s: ADDRESS; c: INTEGER) : ADDRESS ;
strrchr
PROCEDURE __BUILTIN__ strrchr (s: ADDRESS; c: INTEGER) : ADDRESS ;
(*
longjmp - this GCC builtin restricts the val to always 1.
*)
(* do not use these two builtins, as gcc, only really
anticipates that the Ada front end should use them
and it only uses them in its runtime exception handling.
We leave them here in the hope that someday they will
behave more like their libc counterparts. *)
longjmp
PROCEDURE __BUILTIN__ longjmp (env: ADDRESS; val: INTEGER) ;
setjmp
PROCEDURE __BUILTIN__ setjmp (env: ADDRESS) : INTEGER ;
(*
frame_address - returns the address of the frame.
The current frame is obtained if level is 0,
the next level up if level is 1 etc.
*)
frame_address
PROCEDURE __BUILTIN__ frame_address (level: CARDINAL) : ADDRESS ;
(*
return_address - returns the return address of function.
The current function return address is
obtained if level is 0,
the next level up if level is 1 etc.
*)
return_address
PROCEDURE __BUILTIN__ return_address (level: CARDINAL) : ADDRESS ;
(*
alloca_trace - this is a no-op which is used for internal debugging.
*)
alloca_trace
PROCEDURE alloca_trace (returned: ADDRESS; nBytes: CARDINAL) : ADDRESS ;
END Builtins.
Although this module exists and will result in the generation of
in-line code if optimization flags are passed to GNU Modula-2, users
are advised to utilize the same functions from more generic libraries.
The built-in mechanism will be applied to these generic
libraries where appropriate. Note for the mathematical routines to
be in-lined you need to specify the -ffast-math -O options.