libmp -- traditional BSD multiple precision integer arithmetic library
#include <mp.h>
Function prototypes are given in the main body of the text.
Applications using this interface must be linked with -lmp (this library)
and -lcrypto (crypto(3)).
This interface is obsolete in favor of the crypto(3) BIGNUM library.
libmp is the traditional BSD multiple precision integer arithmetic
library. It has a number of problems, and is unsuitable for use in any
programs where reliability is a concern. It is provided here for compatibility
only.
These routines perform arithmetic on integers of arbitrary precision
stored using the defined type MINT. Pointers to MINT are initialized
using itom() or xtom(), and must be recycled with mfree() when they are
no longer needed. Routines which store a result in one of their arguments
expect that the latter has also been initialized prior to being
passed to it. The following routines are defined and implemented:
MINT *itom(short n);
MINT *xtom(const char *s);
char *mtox(const MINT *mp);
void mfree(MINT *mp);
itom() returns an MINT with the value of n. xtom() returns an MINT
with the value of s, which is treated to be in hexadecimal. The
return values from itom() and xtom() must be released with mfree()
when they are no longer needed. mtox() returns a null-terminated
hexadecimal string having the value of mp; its return value must be
released with free() (free(3)) when it is no longer needed.
void madd(const MINT *mp1, const MINT *mp2, MINT *rmp);
void msub(const MINT *mp1, const MINT *mp2, MINT *rmp);
void mult(const MINT *mp1, const MINT *mp2, MINT *rmp);
madd(), msub(), and mult() store the sum, difference, or product,
respectively, of mp1 and mp2 in rmp.
void mdiv(const MINT *nmp, const MINT *dmp, MINT *qmp, MINT *rmp);
void sdiv(const MINT *nmp, short d, MINT *qmp, short *ro);
mdiv() computes the quotient and remainder of nmp and dmp and
stores the result in qmp and rmp, respectively. sdiv() is similar
to mdiv() except the divisor (dmp or d) and remainder (rmp or ro)
are ordinary integers.
void pow(const MINT *bmp, const MINT *emp, const MINT *mmp, MINT *rmp);
void rpow(const MINT *bmp, short e, MINT *rmp);
rpow() computes the result of bmp raised to the empth power and
reduced modulo mmp; the result is stored in rmp. pow() computes
the result of bmp raised to the eth power and stores the result in
rmp.
void min(MINT *mp);
void mout(const MINT *mp);
min() reads a line from standard input, tries to interpret it as a
decimal number, and if successful, stores the result in mp. mout()
prints the value, in decimal, of mp to standard output (without a
trailing newline).
void gcd(const MINT *mp1, const MINT *mp2, MINT *rmp);
gcd() computes the greatest common divisor of mp1 and mp2 and
stores the result in rmp.
int mcmp(const MINT *mp1, const MINT *mp2);
mcmp compares the values of mp1 and mp2 and returns 0 if the two
values are equal, a value greater than 0 if mp1 is greater than
mp2, and a value less than 0 if mp2 is greater than mp1.
void move(const MINT *smp, MINT *tmp);
move() copies the value of smp to tmp (both values must be initialized).
void msqrt(const MINT *nmp, MINT *xmp, MINT *rmp);
msqrt() computes the square root and remainder of nmp and stores
them in xmp and rmp, respectively.
This version of libmp is implemented in terms of the crypto(3) BIGNUM
library.
Running out of memory or illegal operations result in error messages on
standard error and a call to abort(3).
abort(3), bn(3), crypto(3), free(3), malloc(3), math(3)
A libmp library appeared in 4.3BSD. FreeBSD 2.2 shipped with a libmp
implemented in terms of libgmp. This implementation appeared in
FreeBSD 5.0.
The pow() routine exists in both libmp and libm with incompatible semantics.
Errors are reported via output to standard error and abnormal program
termination instead of via return values. The application cannot control
this behavior.
It is not clear whether the string returned by mtox() may be written to
by the caller. This implementation allows it, but others may not. Ideally,
mtox() would take a pointer to a buffer to fill in.
It is not clear whether using the same variable as both source and destination
in a single invocation is permitted. Some of the calls in this
implementation allow this, while others do not.
FreeBSD 5.2.1 September 7, 1989 FreeBSD 5.2.1 [ Back ] |