{ This file is part of the Free Pascal run time library. Copyright (c) 2000 by Jonas Maebe and other members of the Free Pascal development team Implementation of mathematical Routines (only for real) See the file COPYING.FPC, included in this distribution, for details about the copyright. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. **********************************************************************} {**************************************************************************** EXTENDED data type routines ****************************************************************************} {$define FPC_SYSTEM_HAS_ABS} function fpc_abs_real(d : valreal) : valreal;compilerproc; begin { Function is handled internal in the compiler } runerror(207); result:=0; end; {$define FPC_SYSTEM_HAS_SQR} function fpc_sqr_real(d : valreal) : valreal;compilerproc; begin { Function is handled internal in the compiler } runerror(207); result:=0; end; {$define FPC_SYSTEM_HAS_TRUNC} function fpc_trunc_real(d : valreal) : int64;compilerproc; assembler; nostackframe; { input: d in fr1 } { output: result in r3 } asm fctidz f1, f1 stfd f1, -8(r1) ld r3, -8(r1) end; {$ifndef FPC_SYSTEM_HAS_ROUND} {$define FPC_SYSTEM_HAS_ROUND} function fpc_round_real(d : valreal) : int64; compilerproc;assembler; nostackframe; { exactly the same as trunc, except that one fctiwz has become fctiw } { input: d in fr1 } { output: result in r3 } asm fctid f1, f1 stfd f1, -8(r1) ld r3, -8(r1) end; {$endif not FPC_SYSTEM_HAS_ROUND} {**************************************************************************** Int to real helpers ****************************************************************************} {$define FPC_SYSTEM_HAS_INT64_TO_DOUBLE} function fpc_int64_to_double(i: int64): double; compilerproc;assembler; { input: i in r3 } { output: double(i) in f0 } {from "PowerPC Microprocessor Family: Programming Environments Manual for 64 and 32-Bit Microprocessors", v2.0, pg. 698 } var temp : int64; asm std r3,temp // store dword lfd f0,temp // load float fcfid f0,f0 // convert to fpu int end; { we wouls have to generate the .localfunc directive for ELFv2, and moreover it must appear at the start right after setting up the TOC pointer, but the local variables will cause the code generator to already insert the stack allocation before that... -> disable this routine for ELFv2 } {$if not defined(aix) and (not defined(linux) or (defined(_ELF_CALL) and (_ELF_CALL = 1))) } {$define FPC_SYSTEM_HAS_QWORD_TO_DOUBLE} function fpc_qword_to_double(q: qword): double; compilerproc;assembler; const longint_to_real_helper: qword = $80000000; {from "PowerPC Microprocessor Family: Programming Environments Manual for 64 and 32-Bit Microprocessors", v2.0, pg. 698, *exact version* } { input: q in r3 } { output: double(q) in f0 } var temp1, temp2: qword; asm {$ifndef darwin} // load 2^32 into f4 lis r4, longint_to_real_helper@highesta ori r4, r4, longint_to_real_helper@highera sldi r4, r4, 32 oris r4, r4, longint_to_real_helper@ha lfd f4, longint_to_real_helper@l(r4) {$else not darwin} {$ifdef FPC_PIC} mflr r0 bcl 20,31,.Lpiclab .Lpiclab: mflr r5 mtlr r0 addis r4,r5,(longint_to_real_helper-.Lpiclab)@ha lfd f2,(longint_to_real_helper-.Lpiclab)@l(r4) {$else FPC_PIC} lis r4, longint_to_real_helper@ha lfd f4, longint_to_real_helper@l(r4) {$endif FPC_PIC} {$endif not darwin} rldicl r4,r3,32,32 // isolate high half rldicl r0,r3,0,32 // isolate low half std r4,temp1 // store dword both std r0,temp2 lfd f2,temp1 // load float both lfd f0,temp2 // load float both fcfid f2,f2 // convert each half to fcfid f0,f0 // fpu int (no rnd) fmadd f0,f4,f2,f0 // (2**32)*high+low (only add can rnd) end; {$endif ndef aix}