source: libcfa/src/stdlib.hfa @ a9e10e2

//
// Cforall Version 1.0.0 Copyright (C) 2016 University of Waterloo
//
// The contents of this file are covered under the licence agreement in the
// file "LICENCE" distributed with Cforall.
//
// stdlib --
//
// Author           : Peter A. Buhr
// Created On       : Thu Jan 28 17:12:35 2016
// Last Modified By : Peter A. Buhr
// Last Modified On : Thu Feb  2 11:30:04 2023
// Update Count     : 766
//

#pragma once

#include "bits/defs.hfa"                                                                // OPTIONAL_THREAD
#include "bits/align.hfa"                                                               // libAlign
#include "bits/random.hfa"                                                              // prng

#include <stdlib.h>                                                                             // *alloc, strto*, ato*
#include <heap.hfa>


// Reduce includes by explicitly defining these routines.
extern "C" {
        void * memalign( size_t alignment, size_t size );       // malloc.h
        void * pvalloc( size_t size );                                          // malloc.h
        void * memset( void * dest, int fill, size_t size ); // string.h
        void * memcpy( void * dest, const void * src, size_t size ); // string.h
} // extern "C"

//---------------------------------------

#ifndef EXIT_FAILURE
#define EXIT_FAILURE    1                                                               // failing exit status
#define EXIT_SUCCESS    0                                                               // successful exit status
#endif // ! EXIT_FAILURE

//---------------------------------------

#include "common.hfa"

//---------------------------------------

static inline forall( T & | sized(T) ) {
        // CFA safe equivalents, i.e., implicit size specification

        T * malloc( void ) {
                if ( _Alignof(T) <= libAlign() ) return (T *)malloc( sizeof(T) ); // C allocation
                else return (T *)memalign( _Alignof(T), sizeof(T) );
        } // malloc

        T * aalloc( size_t dim ) {
                if ( _Alignof(T) <= libAlign() ) return (T *)aalloc( dim, sizeof(T) ); // C allocation
                else return (T *)amemalign( _Alignof(T), dim, sizeof(T) );
        } // aalloc

        T * calloc( size_t dim ) {
                if ( _Alignof(T) <= libAlign() ) return (T *)calloc( dim, sizeof(T) ); // C allocation
                else return (T *)cmemalign( _Alignof(T), dim, sizeof(T) );
        } // calloc

        T * resize( T * ptr, size_t size ) {                            // CFA resize, eliminate return-type cast
                if ( _Alignof(T) <= libAlign() ) return (T *)resize( (void *)ptr, size ); // CFA resize
                else return (T *)resize( (void *)ptr, _Alignof(T), size ); // CFA resize
        } // resize

        T * realloc( T * ptr, size_t size ) {                           // CFA realloc, eliminate return-type cast
                if ( _Alignof(T) <= libAlign() ) return (T *)realloc( (void *)ptr, size ); // C realloc
                else return (T *)realloc( (void *)ptr, _Alignof(T), size ); // CFA realloc
        } // realloc

        T * memalign( size_t align ) {
                return (T *)memalign( align, sizeof(T) );               // C memalign
        } // memalign

        T * amemalign( size_t align, size_t dim ) {
                return (T *)amemalign( align, dim, sizeof(T) ); // CFA amemalign
        } // amemalign

        T * cmemalign( size_t align, size_t dim  ) {
                return (T *)cmemalign( align, dim, sizeof(T) ); // CFA cmemalign
        } // cmemalign

        T * aligned_alloc( size_t align ) {
                return (T *)aligned_alloc( align, sizeof(T) );  // C aligned_alloc
        } // aligned_alloc

        int posix_memalign( T ** ptr, size_t align ) {
                return posix_memalign( (void **)ptr, align, sizeof(T) ); // C posix_memalign
        } // posix_memalign

        T * valloc( void ) {
                return (T *)valloc( sizeof(T) );                                // C valloc
        } // valloc

        T * pvalloc( void ) {
                return (T *)pvalloc( sizeof(T) );                               // C pvalloc
        } // pvalloc
} // distribution

/*
        FIX ME : fix alloc interface after Ticker Number 214 is resolved, define and add union to S_fill. Then, modify postfix-fill functions to support T * with nmemb, char, and T object of any size. Finally, change alloc_internal.
        Or, just follow the instructions below for that.

        1. Replace the current forall-block that contains defintions of S_fill and S_realloc with following:
                forall( T & | sized(T) ) {
                        union  U_fill           { char c; T * a; T t; };
                        struct S_fill           { char tag; U_fill(T) fill; };
                        struct S_realloc        { inline T *; };
                }

        2. Replace all current postfix-fill functions with following for updated S_fill:
                S_fill(T) ?`fill( char a )                                      { S_fill(T) ret = {'c'}; ret.fill.c = a; return ret; }
                S_fill(T) ?`fill( T    a )                                      { S_fill(T) ret = {'t'}; memcpy(&ret.fill.t, &a, sizeof(T)); return ret; }
                S_fill(T) ?`fill( T    a[], size_t nmemb )      { S_fill(T) ret = {'a', nmemb}; ret.fill.a = a; return ret; }

        3. Replace the alloc_internal$ function which is outside ttype forall-block with following function:
                T * alloc_internal$( void * Resize, T * Realloc, size_t Align, size_t Dim, S_fill(T) Fill) {
                        T * ptr = NULL;
                        size_t size = sizeof(T);
                        size_t copy_end = 0;

                        if(Resize) {
                                ptr = (T*) (void *) resize( (int *)Resize, Align, Dim * size );
                        } else if (Realloc) {
                                if (Fill.tag != '0') copy_end = min(malloc_size( Realloc ), Dim * size);
                                ptr = (T*) (void *) realloc( (int *)Realloc, Align, Dim * size );
                        } else {
                                ptr = (T*) (void *) memalign( Align, Dim * size );
                        }

                        if(Fill.tag == 'c') {
                                memset( (char *)ptr + copy_end, (int)Fill.fill.c, Dim * size - copy_end );
                        } else if(Fill.tag == 't') {
                                for ( int i = copy_end; i <= Dim * size - size ; i += size ) {
                                        memcpy( (char *)ptr + i, &Fill.fill.t, size );
                                }
                        } else if(Fill.tag == 'a') {
                                memcpy( (char *)ptr + copy_end, Fill.fill.a, min(Dim * size - copy_end, size * Fill.nmemb) );
                        }

                        return ptr;
                } // alloc_internal$
*/

typedef struct S_align                  { inline size_t;  } T_align;
typedef struct S_resize                 { inline void *;  }     T_resize;

forall( T & ) {
        struct S_fill           { char tag; char c; size_t size; T * at; char t[50]; };
        struct S_realloc        { inline T *; };
}

static inline T_align   ?`align   ( size_t a )  { return (T_align){a}; }
static inline T_resize  ?`resize  ( void * a )  { return (T_resize){a}; }

static inline forall( T & | sized(T) ) {
        S_fill(T) ?`fill ( T t ) {
                S_fill(T) ret = { 't' };
                size_t size = sizeof(T);
                if ( size > sizeof(ret.t) ) {
                        abort( "ERROR: const object of size greater than 50 bytes given for dynamic memory fill\n" );
                } // if
                memcpy( &ret.t, &t, size );
                return ret;
        }
        S_fill(T)               ?`fill ( zero_t ) = void; // FIX ME: remove this once ticket 214 is resolved
        S_fill(T)               ?`fill ( T * a )                                { return (S_fill(T)){ 'T', '0', 0, a }; } // FIX ME: remove this once ticket 214 is resolved
        S_fill(T)               ?`fill ( char c )                               { return (S_fill(T)){ 'c', c }; }
        S_fill(T)               ?`fill ( T a[], size_t nmemb )  { return (S_fill(T)){ 'a', '0', nmemb * sizeof(T), a }; }

        S_realloc(T)    ?`realloc ( T * a )                             { return (S_realloc(T)){a}; }

        T * alloc_internal$( void * Resize, T * Realloc, size_t Align, size_t Dim, S_fill(T) Fill ) {
                T * ptr = NULL;
                size_t size = sizeof(T);
                size_t copy_end = 0;

                if ( Resize ) {
                        ptr = (T*) (void *) resize( (void *)Resize, Align, Dim * size );
                } else if ( Realloc ) {
                        if ( Fill.tag != '0' ) copy_end = min(malloc_size( Realloc ), Dim * size );
                        ptr = (T *) (void *) realloc( (void *)Realloc, Align, Dim * size );
                } else {
                        ptr = (T *) (void *) memalign( Align, Dim * size );
                }

                if ( Fill.tag == 'c' ) {
                        memset( (char *)ptr + copy_end, (int)Fill.c, Dim * size - copy_end );
                } else if ( Fill.tag == 't' ) {
                        for ( i; copy_end ~ Dim * size ~ size ) {
                                #pragma GCC diagnostic push
                                #pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
                                assert( size <= sizeof(Fill.t) );
                                memcpy( (char *)ptr + i, &Fill.t, size );
                                #pragma GCC diagnostic pop
                        }
                } else if ( Fill.tag == 'a' ) {
                        memcpy( (char *)ptr + copy_end, Fill.at, min(Dim * size - copy_end, Fill.size) );
                } else if ( Fill.tag == 'T' ) {
                        memcpy( (char *)ptr + copy_end, Fill.at, Dim * size );
                }

                return ptr;
        } // alloc_internal$

        forall( TT... | { T * alloc_internal$( void *, T *, size_t, size_t, S_fill(T), TT ); } ) {
                T * alloc_internal$( void *, T *, size_t Align, size_t Dim, S_fill(T) Fill, T_resize Resize, TT rest ) {
                return alloc_internal$( Resize, (T*)0p, Align, Dim, Fill, rest);
                }

                T * alloc_internal$( void *, T *, size_t Align, size_t Dim, S_fill(T) Fill, S_realloc(T) Realloc, TT rest ) {
                return alloc_internal$( (void*)0p, Realloc, Align, Dim, Fill, rest);
                }

                T * alloc_internal$( void * Resize, T * Realloc, size_t, size_t Dim, S_fill(T) Fill, T_align Align, TT rest ) {
                return alloc_internal$( Resize, Realloc, Align, Dim, Fill, rest);
                }

                T * alloc_internal$( void * Resize, T * Realloc, size_t Align, size_t Dim, S_fill(T), S_fill(T) Fill, TT rest ) {
                return alloc_internal$( Resize, Realloc, Align, Dim, Fill, rest );
                }

            T * alloc( TT all ) {
                return alloc_internal$( (void*)0p, (T*)0p, (_Alignof(T) > libAlign() ? _Alignof(T) : libAlign()), (size_t)1, (S_fill(T)){'0'}, all );
            }

            T * alloc( size_t dim, TT all ) {
                return alloc_internal$( (void*)0p, (T*)0p, (_Alignof(T) > libAlign() ? _Alignof(T) : libAlign()), dim, (S_fill(T)){'0'}, all );
            }
        } // distribution TT
} // distribution T

static inline forall( T & | sized(T) ) {
        // CFA safe initialization/copy, i.e., implicit size specification, non-array types
        T * memset( T * dest, char fill ) {
                return (T *)memset( dest, fill, sizeof(T) );
        } // memset

        T * memcpy( T * dest, const T * src ) {
                return (T *)memcpy( dest, src, sizeof(T) );
        } // memcpy

        // CFA safe initialization/copy, i.e., implicit size specification, array types
        T * amemset( T dest[], char fill, size_t dim ) {
                return (T *)(void *)memset( dest, fill, dim * sizeof(T) ); // C memset
        } // amemset

        T * amemcpy( T dest[], const T src[], size_t dim ) {
                return (T *)(void *)memcpy( dest, src, dim * sizeof(T) ); // C memcpy
        } // amemcpy
} // distribution

// CFA deallocation for multiple objects
static inline forall( T & )                                                     // FIX ME, problems with 0p in list
void free( T * ptr ) {
        free( (void *)ptr );                                                            // C free
} // free
static inline forall( T &, TT... | { void free( TT ); } )
void free( T * ptr, TT rest ) {
        free( ptr );
        free( rest );
} // free

// CFA allocation/deallocation and constructor/destructor, non-array types
static inline forall( T & | sized(T), TT... | { void ?{}( T &, TT ); } )
T * new( TT p ) {
        return &(*(T *)malloc()){ p };                                          // run constructor
} // new

static inline forall( T & | { void ^?{}( T & ); } )
void delete( T * ptr ) {
        // special case for 0-sized object => always call destructor
        if ( ptr || sizeof(ptr) == 0 ) {                                        // ignore null but not 0-sized objects
                ^(*ptr){};                                                                              // run destructor
        } // if
        free( ptr );                                                                            // always call free
} // delete
static inline forall( T &, TT... | { void ^?{}( T & ); void delete( TT ); } )
void delete( T * ptr, TT rest ) {
        delete( ptr );
        delete( rest );
} // delete

// CFA allocation/deallocation and constructor/destructor, array types
forall( T & | sized(T), TT... | { void ?{}( T &, TT ); } ) T * anew( size_t dim, TT p );
forall( T & | sized(T) | { void ^?{}( T & ); } ) void adelete( T arr[] );
forall( T & | sized(T) | { void ^?{}( T & ); }, TT... | { void adelete( TT ); } ) void adelete( T arr[], TT rest );

//---------------------------------------

static inline {
        int strto( const char sptr[], char ** eptr, int base ) { return (int)strtol( sptr, eptr, base ); }
        unsigned int strto( const char sptr[], char ** eptr, int base ) { return (unsigned int)strtoul( sptr, eptr, base ); }
        long int strto( const char sptr[], char ** eptr, int base ) { return strtol( sptr, eptr, base ); }
        unsigned long int strto( const char sptr[], char ** eptr, int base ) { return strtoul( sptr, eptr, base ); }
        long long int strto( const char sptr[], char ** eptr, int base ) { return strtoll( sptr, eptr, base ); }
        unsigned long long int strto( const char sptr[], char ** eptr, int base ) { return strtoull( sptr, eptr, base ); }

        float strto( const char sptr[], char ** eptr ) { return strtof( sptr, eptr ); }
        double strto( const char sptr[], char ** eptr ) { return strtod( sptr, eptr ); }
        long double strto( const char sptr[], char ** eptr ) { return strtold( sptr, eptr ); }
} // distribution

float _Complex strto( const char sptr[], char ** eptr );
double _Complex strto( const char sptr[], char ** eptr );
long double _Complex strto( const char sptr[], char ** eptr );

static inline {
        int ato( const char sptr[] ) { return (int)strtol( sptr, 0p, 10 ); }
        unsigned int ato( const char sptr[] ) { return (unsigned int)strtoul( sptr, 0p, 10 ); }
        long int ato( const char sptr[] ) { return strtol( sptr, 0p, 10 ); }
        unsigned long int ato( const char sptr[] ) { return strtoul( sptr, 0p, 10 ); }
        long long int ato( const char sptr[] ) { return strtoll( sptr, 0p, 10 ); }
        unsigned long long int ato( const char sptr[] ) { return strtoull( sptr, 0p, 10 ); }

        float ato( const char sptr[] ) { return strtof( sptr, 0p ); }
        double ato( const char sptr[] ) { return strtod( sptr, 0p ); }
        long double ato( const char sptr[] ) { return strtold( sptr, 0p ); }

        float _Complex ato( const char sptr[] ) { return strto( sptr, 0p ); }
        double _Complex ato( const char sptr[] ) { return strto( sptr, 0p ); }
        long double _Complex ato( const char sptr[] ) { return strto( sptr, 0p ); }
} // distribution

//---------------------------------------

forall( E | { int ?<?( E, E ); } ) {
        E * bsearch( E key, const E * vals, size_t dim );
        size_t bsearch( E key, const E * vals, size_t dim );
        E * bsearchl( E key, const E * vals, size_t dim );
        size_t bsearchl( E key, const E * vals, size_t dim );
        E * bsearchu( E key, const E * vals, size_t dim );
        size_t bsearchu( E key, const E * vals, size_t dim );
} // distribution

forall( K, E | { int ?<?( K, K ); K getKey( const E & ); } ) {
        E * bsearch( K key, const E * vals, size_t dim );
        size_t bsearch( K key, const E * vals, size_t dim );
        E * bsearchl( K key, const E * vals, size_t dim );
        size_t bsearchl( K key, const E * vals, size_t dim );
        E * bsearchu( K key, const E * vals, size_t dim );
        size_t bsearchu( K key, const E * vals, size_t dim );
} // distribution

forall( E | { int ?<?( E, E ); } ) {
        void qsort( E * vals, size_t dim );
} // distribution

//---------------------------------------

extern "C" {                                                                                    // override C version
        void srandom( unsigned int seed );
        long int random( void );                                                        // GENERATES POSITIVE AND NEGATIVE VALUES
        // For positive values, use unsigned int, e.g., unsigned int r = random() % 100U;
} // extern "C"

static inline {
        long int random( long int l, long int u ) { if ( u < l ) [u, l] = [l, u]; return lrand48() % (u - l + 1) + l; } // [l,u]
        long int random( long int u ) { return random( 0, u - 1 ); } // [0,u)
        unsigned long int random( void ) { return lrand48(); }
        unsigned long int random( unsigned long int u ) { return lrand48() % u; } // [0,u)
        unsigned long int random( unsigned long int l, unsigned long int u ) { if ( u < l ) [u, l] = [l, u]; return lrand48() % (u - l + 1) + l; } // [l,u]

        char random( void ) { return (unsigned long int)random(); }
        char random( char u ) { return (unsigned long int)random( (unsigned long int)u ); } // [0,u)
        char random( char l, char u ) { return random( (unsigned long int)l, (unsigned long int)u ); } // [l,u)
        int random( void ) { return (long int)random(); }
        int random( int u ) { return (long int)random( (long int)u ); } // [0,u]
        int random( int l, int u ) { return random( (long int)l, (long int)u ); } // [l,u)
        unsigned int random( void ) { return (unsigned long int)random(); }
        unsigned int random( unsigned int u ) { return (unsigned long int)random( (unsigned long int)u ); } // [0,u]
        unsigned int random( unsigned int l, unsigned int u ) { return random( (unsigned long int)l, (unsigned long int)u ); } // [l,u)
} // distribution

float random( void );                                                                   // [0.0, 1.0)
double random( void );                                                                  // [0.0, 1.0)
float _Complex random( void );                                                  // [0.0, 1.0)+[0.0, 1.0)i
double _Complex random( void );                                                 // [0.0, 1.0)+[0.0, 1.0)i
long double _Complex random( void );                                    // [0.0, 1.0)+[0.0, 1.0)i

//---------------------------------------

// Sequential Pseudo Random-Number Generator : generate repeatable sequence of values that appear random.
//
// Declaration :
//   PRNG sprng = { 1009 } - set starting seed versus random seed
//
// Interface :
//   set_seed( sprng, 1009 ) - set starting seed for ALL kernel threads versus random seed
//   get_seed( sprng ) - read seed
//   prng( sprng ) - generate random value in range [0,UINT_MAX]
//   prng( sprng, u ) - generate random value in range [0,u)
//   prng( sprng, l, u ) - generate random value in range [l,u]
//   calls( sprng ) - number of generated random value so far
//
// Examples : generate random number between 5-21
//   prng( sprng ) % 17 + 5;    values 0-16 + 5 = 5-21
//   prng( sprng, 16 + 1 ) + 5;
//   prng( sprng, 5, 21 );
//   calls( sprng );

forall( PRNG &, R )
trait basic_prng {
        void set_seed( PRNG & prng, R seed );                           // set seed
        R get_seed( PRNG & prng );                                                      // get seed
        R prng( PRNG & prng );
        void ?{}( PRNG & prng );                                                        // random seed
        void ?{}( PRNG & prng, R seed );                                        // fixed seed
}; // basic_prng

static inline forall( PRNG &, R | basic_prng( PRNG, R ) | { R ?%?( R, R ); } ) {
        R prng( PRNG & prng, R u ) { return prng( prng ) % u; } // [0,u)
}
static inline forall( PRNG &, R | basic_prng( PRNG, R ) | { R ?+?( R, R ); R ?-?( R, R ); R ?%?( R, R ); void ?{}( R &, one_t ); } ) {
        R prng( PRNG & prng, R l, R u ) { return prng( prng, u - l + (R){1} ) + l; } // [l,u]
}

struct PRNG32 {
        uint32_t callcnt;                                                                       // call count
        uint32_t seed;                                                                          // current seed
        PRNG_STATE_32_T state;                                                          // random state
}; // PRNG

static inline {
        void set_seed( PRNG32 & prng, uint32_t seed_ ) with( prng ) { seed = seed_; PRNG_SET_SEED_32( state, seed ); }
        uint32_t get_seed( PRNG32 & prng ) __attribute__(( warn_unused_result )) with( prng ) { return seed; }
        uint32_t prng( PRNG32 & prng ) __attribute__(( warn_unused_result )) with( prng ) { callcnt += 1; return PRNG_NAME_32( state ); } // [0,UINT_MAX]
        uint32_t prng( PRNG32 & prng, uint32_t u ) __attribute__(( warn_unused_result )) { return prng( prng ) % u; } // [0,u)
        uint32_t prng( PRNG32 & prng, uint32_t l, uint32_t u ) __attribute__(( warn_unused_result )) { return prng( prng, u - l + 1 ) + l; } // [l,u]
        uint32_t calls( PRNG32 & prng ) __attribute__(( warn_unused_result )) with( prng ) { return callcnt; }
        void ?{}( PRNG32 & prng ) with( prng ) { callcnt = 0; set_seed( prng, rdtscl() ); } // random seed
        void ?{}( PRNG32 & prng, uint32_t seed ) with( prng ) { callcnt = 0; set_seed( prng, seed ); } // fixed seed
} // distribution

struct PRNG64 {
        uint64_t callcnt;                                                                       // call count
        uint64_t seed;                                                                          // current seed
        PRNG_STATE_64_T state;                                                          // random state
}; // PRNG

static inline {
        void set_seed( PRNG64 & prng, uint64_t seed_ ) with( prng ) { seed = seed_; PRNG_SET_SEED_64( state, seed ); }
        uint64_t get_seed( PRNG64 & prng ) __attribute__(( warn_unused_result )) with( prng ) { return seed; }
        uint64_t prng( PRNG64 & prng ) __attribute__(( warn_unused_result )) with( prng ) { callcnt += 1; return PRNG_NAME_64( state ); } // [0,UINT_MAX]
        uint64_t prng( PRNG64 & prng, uint64_t u ) __attribute__(( warn_unused_result )) { return prng( prng ) % u; } // [0,u)
        uint64_t prng( PRNG64 & prng, uint64_t l, uint64_t u ) __attribute__(( warn_unused_result )) { return prng( prng, u - l + 1 ) + l; } // [l,u]
        uint64_t calls( PRNG64 & prng ) __attribute__(( warn_unused_result )) with( prng ) { return callcnt; }
        void ?{}( PRNG64 & prng ) with( prng ) { callcnt = 0; set_seed( prng, rdtscl() ); } // random seed
        void ?{}( PRNG64 & prng, uint64_t seed ) with( prng ) { callcnt = 0; set_seed( prng, seed ); } // fixed seed
} // distribution

// Concurrent Pseudo Random-Number Generator : generate repeatable sequence of values that appear random.
//
// Interface :
//   set_seed( 1009 ) - fixed seed for all kernel threads versus random seed
//   get_seed() - read seed
//   prng() - generate random value in range [0,UINT_MAX]
//   prng( u ) - generate random value in range [0,u)
//   prng( l, u ) - generate random value in range [l,u]
//
// Examples : generate random number between 5-21
//   prng() % 17 + 5;   values 0-16 + 5 = 5-21
//   prng( 16 + 1 ) + 5;
//   prng( 5, 21 );

// Harmonize with concurrency/thread.hfa.
void set_seed( size_t seed_ ) OPTIONAL_THREAD;                  // set global seed
size_t get_seed() __attribute__(( warn_unused_result )); // get global seed
size_t prng( void ) __attribute__(( warn_unused_result )) OPTIONAL_THREAD; // [0,UINT_MAX]
static inline {
        size_t prng( size_t u ) __attribute__(( warn_unused_result )) { return prng() % u; } // [0,u)
        size_t prng( size_t l, size_t u ) __attribute__(( warn_unused_result )) { return prng( u - l + 1 ) + l; } // [l,u]
} // distribution

//---------------------------------------

extern bool threading_enabled( void ) OPTIONAL_THREAD;

// Local Variables: //
// mode: c //
// tab-width: 4 //
// End: //