freepascal.compiler/compiler/i386/n386mat.pas

{
    $Id$
    Copyright (c) 1998-2002 by Florian Klaempfl

    Generate i386 assembler for math nodes

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    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.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

 ****************************************************************************
}
unit n386mat;

{$i fpcdefs.inc}

interface

    uses
      node,nmat,ncgmat,nx86mat;

    type
      ti386moddivnode = class(tmoddivnode)
         procedure pass_2;override;
      end;

      ti386shlshrnode = class(tshlshrnode)
         procedure pass_2;override;
         { everything will be handled in pass_2 }
         function first_shlshr64bitint: tnode; override;
      end;

      ti386unaryminusnode = class(tx86unaryminusnode)
      end;

      ti386notnode = class(tx86notnode)
      end;


implementation

    uses
      globtype,systems,
      cutils,verbose,globals,
      symconst,symdef,aasmbase,aasmtai,defutil,
      cgbase,pass_1,pass_2,
      ncon,
      cpubase,cpuinfo,
      cga,ncgutil,cgobj;

{*****************************************************************************
                             TI386MODDIVNODE
*****************************************************************************}

    procedure ti386moddivnode.pass_2;

    var  hreg1,hreg2:Tregister;
         power:longint;
         hl:Tasmlabel;
         op:Tasmop;

    begin
      secondpass(left);
      if codegenerror then
        exit;
      secondpass(right);
      if codegenerror then
        exit;

      if is_64bitint(resulttype.def) then
        { should be handled in pass_1 (JM) }
        internalerror(200109052);
      { put numerator in register }
      location_reset(location,LOC_REGISTER,OS_INT);
      location_force_reg(exprasmlist,left.location,OS_INT,false);
      hreg1:=left.location.register;

      if (nodetype=divn) and (right.nodetype=ordconstn) and
         ispowerof2(tordconstnode(right).value,power) then
        begin
          { for signed numbers, the numerator must be adjusted before the
            shift instruction, but not wih unsigned numbers! Otherwise,
            "Cardinal($ffffffff) div 16" overflows! (JM) }
          if is_signed(left.resulttype.def) Then
            begin
              if (aktOptProcessor <> class386) and
                 not(cs_littlesize in aktglobalswitches) then
                { use a sequence without jumps, saw this in
                  comp.compilers (JM) }
                begin
                  { no jumps, but more operations }
                  hreg2:=cg.getintregister(exprasmlist,OS_INT);
                  emit_reg_reg(A_MOV,S_L,hreg1,hreg2);
                  {If the left value is signed, hreg2=$ffffffff, otherwise 0.}
                  emit_const_reg(A_SAR,S_L,31,hreg2);
                  {If signed, hreg2=right value-1, otherwise 0.}
                  emit_const_reg(A_AND,S_L,tordconstnode(right).value-1,hreg2);
                  { add to the left value }
                  emit_reg_reg(A_ADD,S_L,hreg2,hreg1);
                  { release EDX if we used it }
                  cg.ungetregister(exprasmlist,hreg2);
                  { do the shift }
                  emit_const_reg(A_SAR,S_L,power,hreg1);
                end
              else
                begin
                  { a jump, but less operations }
                  emit_reg_reg(A_TEST,S_L,hreg1,hreg1);
                  objectlibrary.getlabel(hl);
                  cg.a_jmp_flags(exprasmlist,F_NS,hl);
                  if power=1 then
                    emit_reg(A_INC,S_L,hreg1)
                  else
                    emit_const_reg(A_ADD,S_L,tordconstnode(right).value-1,hreg1);
                  cg.a_label(exprasmlist,hl);
                  emit_const_reg(A_SAR,S_L,power,hreg1);
                end
            end
          else
            emit_const_reg(A_SHR,S_L,power,hreg1);
          location.register:=hreg1;
        end
      else
        begin
          {Bring denominator to a register.}
          cg.ungetregister(exprasmlist,hreg1);
          cg.getexplicitregister(exprasmlist,NR_EAX);
          emit_reg_reg(A_MOV,S_L,hreg1,NR_EAX);
          cg.getexplicitregister(exprasmlist,NR_EDX);
          {Sign extension depends on the left type.}
          if torddef(left.resulttype.def).typ=u32bit then
            emit_reg_reg(A_XOR,S_L,NR_EDX,NR_EDX)
          else
            emit_none(A_CDQ,S_NO);

          {Division depends on the right type.}
          if Torddef(right.resulttype.def).typ=u32bit then
            op:=A_DIV
          else
            op:=A_IDIV;

          if right.location.loc in [LOC_REFERENCE,LOC_CREFERENCE] then
            emit_ref(op,S_L,right.location.reference)
          else if right.location.loc in [LOC_REGISTER,LOC_CREGISTER] then
            emit_reg(op,S_L,right.location.register)
          else
            begin
              hreg1:=cg.getintregister(exprasmlist,right.location.size);
              cg.a_load_loc_reg(exprasmlist,OS_32,right.location,hreg1);
              cg.ungetregister(exprasmlist,hreg1);
              emit_reg(op,S_L,hreg1);
            end;
          location_release(exprasmlist,right.location);

          {Copy the result into a new register. Release EAX & EDX.}
          if nodetype=divn then
            begin
              cg.ungetregister(exprasmlist,NR_EDX);
              cg.ungetregister(exprasmlist,NR_EAX);
              location.register:=cg.getintregister(exprasmlist,OS_INT);
              emit_reg_reg(A_MOV,S_L,NR_EAX,location.register);
            end
          else
            begin
              cg.ungetregister(exprasmlist,NR_EAX);
              cg.ungetregister(exprasmlist,NR_EDX);
              location.register:=cg.getintregister(exprasmlist,OS_INT);
              emit_reg_reg(A_MOV,S_L,NR_EDX,location.register);
            end;
        end;
    end;


{*****************************************************************************
                             TI386SHLRSHRNODE
*****************************************************************************}


    function ti386shlshrnode.first_shlshr64bitint: tnode;

    begin
      result := nil;
    end;

    procedure ti386shlshrnode.pass_2;

    var hregisterhigh,hregisterlow:Tregister;
        op:Tasmop;
        l1,l2,l3:Tasmlabel;

    begin
      secondpass(left);
      secondpass(right);

      { determine operator }
      if nodetype=shln then
        op:=A_SHL
      else
        op:=A_SHR;

      if is_64bitint(left.resulttype.def) then
        begin
          location_reset(location,LOC_REGISTER,OS_64);

          { load left operator in a register }
          location_force_reg(exprasmlist,left.location,OS_64,false);
          hregisterhigh:=left.location.registerhigh;
          hregisterlow:=left.location.registerlow;

          { shifting by a constant directly coded: }
          if (right.nodetype=ordconstn) then
            begin
              { shrd/shl works only for values <=31 !! }
              if Tordconstnode(right).value>63 then
                begin
                  cg.a_load_const_reg(exprasmlist,OS_32,0,hregisterhigh);
                  cg.a_load_const_reg(exprasmlist,OS_32,0,hregisterlow);
                  location.registerlow:=hregisterlow;
                  location.registerhigh:=hregisterhigh;
                end
              else if Tordconstnode(right).value>31 then
                begin
                  if nodetype=shln then
                    begin
                      emit_reg_reg(A_XOR,S_L,hregisterhigh,hregisterhigh);
                      if ((tordconstnode(right).value and 31) <> 0) then
                        emit_const_reg(A_SHL,S_L,tordconstnode(right).value and 31,
                                       hregisterlow);
                    end
                  else
                    begin
                      emit_reg_reg(A_XOR,S_L,hregisterlow,hregisterlow);
                      if ((tordconstnode(right).value and 31) <> 0) then
                        emit_const_reg(A_SHR,S_L,tordconstnode(right).value and 31,
                                       hregisterhigh);
                    end;
                  location.registerhigh:=hregisterlow;
                  location.registerlow:=hregisterhigh;
                end
              else
                begin
                  if nodetype=shln then
                    begin
                      emit_const_reg_reg(A_SHLD,S_L,tordconstnode(right).value and 31,
                                         hregisterlow,hregisterhigh);
                      emit_const_reg(A_SHL,S_L,tordconstnode(right).value and 31,
                                     hregisterlow);
                    end
                  else
                    begin
                      emit_const_reg_reg(A_SHRD,S_L,tordconstnode(right).value and 31,
                                         hregisterhigh,hregisterlow);
                      emit_const_reg(A_SHR,S_L,tordconstnode(right).value and 31,
                                     hregisterhigh);
                    end;
                  location.registerlow:=hregisterlow;
                  location.registerhigh:=hregisterhigh;
                end;
            end
          else
            begin
              { load right operators in a register }
              cg.getexplicitregister(exprasmlist,NR_ECX);
              cg.a_load_loc_reg(exprasmlist,OS_32,right.location,NR_ECX);
              if right.location.loc<>LOC_CREGISTER then
                location_release(exprasmlist,right.location);

              { left operator is already in a register }
              { hence are both in a register }
              { is it in the case ECX ? }

              { the damned shift instructions work only til a count of 32 }
              { so we've to do some tricks here                           }
              objectlibrary.getlabel(l1);
              objectlibrary.getlabel(l2);
              objectlibrary.getlabel(l3);
              emit_const_reg(A_CMP,S_L,64,NR_ECX);
              cg.a_jmp_flags(exprasmlist,F_L,l1);
              emit_reg_reg(A_XOR,S_L,hregisterlow,hregisterlow);
              emit_reg_reg(A_XOR,S_L,hregisterhigh,hregisterhigh);
              cg.a_jmp_always(exprasmlist,l3);
              cg.a_label(exprasmlist,l1);
              emit_const_reg(A_CMP,S_L,32,NR_ECX);
              cg.a_jmp_flags(exprasmlist,F_L,l2);
              emit_const_reg(A_SUB,S_L,32,NR_ECX);
              if nodetype=shln then
                begin
                  emit_reg_reg(A_SHL,S_L,NR_CL,hregisterlow);
                  emit_reg_reg(A_MOV,S_L,hregisterlow,hregisterhigh);
                  emit_reg_reg(A_XOR,S_L,hregisterlow,hregisterlow);
                  cg.a_jmp_always(exprasmlist,l3);
                  cg.a_label(exprasmlist,l2);
                  emit_reg_reg_reg(A_SHLD,S_L,NR_CL,hregisterlow,hregisterhigh);
                  emit_reg_reg(A_SHL,S_L,NR_CL,hregisterlow);
                end
              else
                begin
                  emit_reg_reg(A_SHR,S_L,NR_CL,hregisterhigh);
                  emit_reg_reg(A_MOV,S_L,hregisterhigh,hregisterlow);
                  emit_reg_reg(A_XOR,S_L,hregisterhigh,hregisterhigh);
                  cg.a_jmp_always(exprasmlist,l3);
                  cg.a_label(exprasmlist,l2);
                  emit_reg_reg_reg(A_SHRD,S_L,NR_CL,hregisterhigh,hregisterlow);
                  emit_reg_reg(A_SHR,S_L,NR_CL,hregisterhigh);
                end;
              cg.a_label(exprasmlist,l3);

              cg.ungetregister(exprasmlist,NR_ECX);
              location.registerlow:=hregisterlow;
              location.registerhigh:=hregisterhigh;
            end;
        end
      else
        begin
          { load left operators in a register }
          location_copy(location,left.location);
          location_force_reg(exprasmlist,location,OS_INT,false);

          { shifting by a constant directly coded: }
          if (right.nodetype=ordconstn) then
            { l shl 32 should 0 imho, but neither TP nor Delphi do it in this way (FK)}
            emit_const_reg(op,S_L,tordconstnode(right).value and 31,location.register)
          else
            begin
              { load right operators in a ECX }
              if right.location.loc<>LOC_CREGISTER then
                location_release(exprasmlist,right.location);
              cg.getexplicitregister(exprasmlist,NR_ECX);
              cg.a_load_loc_reg(exprasmlist,OS_32,right.location,NR_ECX);

              { right operand is in ECX }
              cg.ungetregister(exprasmlist,NR_ECX);
              emit_reg_reg(op,S_L,NR_CL,location.register);
            end;
        end;
    end;


begin
   cunaryminusnode:=ti386unaryminusnode;
   cmoddivnode:=ti386moddivnode;
   cshlshrnode:=ti386shlshrnode;
   cnotnode:=ti386notnode;
end.
{
  $Log$
  Revision 1.69  2004-01-20 12:59:37  florian
    * common addnode code for x86-64 and i386

  Revision 1.68  2003/12/26 13:19:16  florian
    * rtl and compiler compile with -Cfsse2

  Revision 1.67  2003/12/25 01:07:09  florian
    + $fputype directive support
    + single data type operations with sse unit
    * fixed more x86-64 stuff

  Revision 1.66  2003/12/10 17:28:41  peter
    * int64 shl/shr > 63 returns 0

  Revision 1.65  2003/10/10 17:48:14  peter
    * old trgobj moved to x86/rgcpu and renamed to trgx86fpu
    * tregisteralloctor renamed to trgobj
    * removed rgobj from a lot of units
    * moved location_* and reference_* to cgobj
    * first things for mmx register allocation

  Revision 1.64  2003/10/09 21:31:37  daniel
    * Register allocator splitted, ans abstract now

  Revision 1.63  2003/10/01 20:34:49  peter
    * procinfo unit contains tprocinfo
    * cginfo renamed to cgbase
    * moved cgmessage to verbose
    * fixed ppc and sparc compiles

  Revision 1.62  2003/09/29 20:58:56  peter
    * optimized releasing of registers

  Revision 1.61  2003/09/28 21:48:20  peter
    * fix register leaks

  Revision 1.60  2003/09/03 15:55:01  peter
    * NEWRA branch merged

  Revision 1.59.2.2  2003/08/31 13:50:16  daniel
    * Remove sorting and use pregenerated indexes
    * Some work on making things compile

  Revision 1.59.2.1  2003/08/29 17:29:00  peter
    * next batch of updates

  Revision 1.59  2003/07/02 22:18:04  peter
    * paraloc splitted in callerparaloc,calleeparaloc
    * sparc calling convention updates

  Revision 1.58  2003/06/13 21:19:31  peter
    * current_procdef removed, use current_procinfo.procdef instead

  Revision 1.57  2003/06/03 21:11:09  peter
    * cg.a_load_* get a from and to size specifier
    * makeregsize only accepts newregister
    * i386 uses generic tcgnotnode,tcgunaryminus

  Revision 1.56  2003/06/03 13:01:59  daniel
    * Register allocator finished

  Revision 1.55  2003/05/31 15:04:31  peter
    * load_loc_reg update

  Revision 1.54  2003/05/22 21:32:29  peter
    * removed some unit dependencies

  Revision 1.53  2003/04/22 23:50:23  peter
    * firstpass uses expectloc
    * checks if there are differences between the expectloc and
      location.loc from secondpass in EXTDEBUG

  Revision 1.52  2003/04/22 14:33:38  peter
    * removed some notes/hints

  Revision 1.51  2003/04/22 10:09:35  daniel
    + Implemented the actual register allocator
    + Scratch registers unavailable when new register allocator used
    + maybe_save/maybe_restore unavailable when new register allocator used

  Revision 1.50  2003/04/21 19:15:26  peter
    * when ecx is not available allocated another register

  Revision 1.49  2003/04/17 10:02:48  daniel
    * Tweaked register allocate/deallocate positition to less interferences
      are generated.

  Revision 1.48  2003/03/28 19:16:57  peter
    * generic constructor working for i386
    * remove fixed self register
    * esi added as address register for i386

  Revision 1.47  2003/03/08 20:36:41  daniel
    + Added newra version of Ti386shlshrnode
    + Added interference graph construction code

  Revision 1.46  2003/03/08 13:59:17  daniel
    * Work to handle new register notation in ag386nsm
    + Added newra version of Ti386moddivnode

  Revision 1.45  2003/02/19 22:00:15  daniel
    * Code generator converted to new register notation
    - Horribily outdated todo.txt removed

  Revision 1.44  2003/01/13 18:37:44  daniel
    * Work on register conversion

  Revision 1.43  2003/01/13 14:54:34  daniel
    * Further work to convert codegenerator register convention;
      internalerror bug fixed.

  Revision 1.42  2003/01/08 18:43:57  daniel
   * Tregister changed into a record

  Revision 1.41  2002/11/25 17:43:26  peter
    * splitted defbase in defutil,symutil,defcmp
    * merged isconvertable and is_equal into compare_defs(_ext)
    * made operator search faster by walking the list only once

  Revision 1.40  2002/09/07 15:25:10  peter
    * old logs removed and tabs fixed

  Revision 1.39  2002/08/15 15:15:55  carl
    * jmpbuf size allocation for exceptions is now cpu specific (as it should)
    * more generic nodes for maths
    * several fixes for better m68k support

  Revision 1.38  2002/08/14 19:18:16  carl
    * bugfix of unaryminus node with left LOC_CREGISTER

  Revision 1.37  2002/08/12 15:08:42  carl
    + stab register indexes for powerpc (moved from gdb to cpubase)
    + tprocessor enumeration moved to cpuinfo
    + linker in target_info is now a class
    * many many updates for m68k (will soon start to compile)
    - removed some ifdef or correct them for correct cpu

  Revision 1.36  2002/08/11 14:32:30  peter
    * renamed current_library to objectlibrary

  Revision 1.35  2002/08/11 13:24:17  peter
    * saving of asmsymbols in ppu supported
    * asmsymbollist global is removed and moved into a new class
      tasmlibrarydata that will hold the info of a .a file which
      corresponds with a single module. Added librarydata to tmodule
      to keep the library info stored for the module. In the future the
      objectfiles will also be stored to the tasmlibrarydata class
    * all getlabel/newasmsymbol and friends are moved to the new class

  Revision 1.34  2002/08/02 07:44:31  jonas
    * made assigned() handling generic
    * add nodes now can also evaluate constant expressions at compile time
      that contain nil nodes

  Revision 1.33  2002/07/20 11:58:02  florian
    * types.pas renamed to defbase.pas because D6 contains a types
      unit so this would conflicts if D6 programms are compiled
    + Willamette/SSE2 instructions to assembler added

  Revision 1.32  2002/07/01 18:46:33  peter
    * internal linker
    * reorganized aasm layer

  Revision 1.31  2002/05/18 13:34:25  peter
    * readded missing revisions

  Revision 1.30  2002/05/16 19:46:51  carl
  + defines.inc -> fpcdefs.inc to avoid conflicts if compiling by hand
  + try to fix temp allocation (still in ifdef)
  + generic constructor calls
  + start of tassembler / tmodulebase class cleanup

  Revision 1.28  2002/05/13 19:54:38  peter
    * removed n386ld and n386util units
    * maybe_save/maybe_restore added instead of the old maybe_push

  Revision 1.27  2002/05/12 16:53:17  peter
    * moved entry and exitcode to ncgutil and cgobj
    * foreach gets extra argument for passing local data to the
      iterator function
    * -CR checks also class typecasts at runtime by changing them
      into as
    * fixed compiler to cycle with the -CR option
    * fixed stabs with elf writer, finally the global variables can
      be watched
    * removed a lot of routines from cga unit and replaced them by
      calls to cgobj
    * u32bit-s32bit updates for and,or,xor nodes. When one element is
      u32bit then the other is typecasted also to u32bit without giving
      a rangecheck warning/error.
    * fixed pascal calling method with reversing also the high tree in
      the parast, detected by tcalcst3 test

  Revision 1.26  2002/04/04 19:06:12  peter
    * removed unused units
    * use tlocation.size in cg.a_*loc*() routines

  Revision 1.25  2002/04/02 17:11:36  peter
    * tlocation,treference update
    * LOC_CONSTANT added for better constant handling
    * secondadd splitted in multiple routines
    * location_force_reg added for loading a location to a register
      of a specified size
    * secondassignment parses now first the right and then the left node
      (this is compatible with Kylix). This saves a lot of push/pop especially
      with string operations
    * adapted some routines to use the new cg methods

  Revision 1.24  2002/03/31 20:26:39  jonas
    + a_loadfpu_* and a_loadmm_* methods in tcg
    * register allocation is now handled by a class and is mostly processor
      independent (+rgobj.pas and i386/rgcpu.pas)
    * temp allocation is now handled by a class (+tgobj.pas, -i386\tgcpu.pas)
    * some small improvements and fixes to the optimizer
    * some register allocation fixes
    * some fpuvaroffset fixes in the unary minus node
    * push/popusedregisters is now called rg.save/restoreusedregisters and
      (for i386) uses temps instead of push/pop's when using -Op3 (that code is
      also better optimizable)
    * fixed and optimized register saving/restoring for new/dispose nodes
    * LOC_FPU locations now also require their "register" field to be set to
      R_ST, not R_ST0 (the latter is used for LOC_CFPUREGISTER locations only)
    - list field removed of the tnode class because it's not used currently
      and can cause hard-to-find bugs

  Revision 1.23  2002/03/04 19:10:14  peter
    * removed compiler warnings

}