{ $Id$ Copyright (c) 1998-2002 by Florian Klaempfl This unit implements the i386 specific class for the register allocator 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 rgcpu; {$i fpcdefs.inc} interface uses cpubase, cpuinfo, aasmbase,aasmtai, cclasses,globtype,cgbase,rgobj; type trgcpu = class(trgobj) procedure add_constraints(reg:Tregister);override; end; tpushedsavedloc = record case byte of 0: (pushed: boolean); 1: (ofs: longint); end; tpushedsavedfpu = array[tsuperregister] of tpushedsavedloc; trgx86fpu = class { The "usableregsxxx" contain all registers of type "xxx" that } { aren't currently allocated to a regvar. The "unusedregsxxx" } { contain all registers of type "xxx" that aren't currently } { allocated } unusedregsfpu,usableregsfpu : Tsuperregisterset; { these counters contain the number of elements in the } { unusedregsxxx/usableregsxxx sets } countunusedregsfpu : byte; { Contains the registers which are really used by the proc itself. It doesn't take care of registers used by called procedures } used_in_proc : tcpuregisterset; {reg_pushes_other : regvarother_longintarray; is_reg_var_other : regvarother_booleanarray; regvar_loaded_other : regvarother_booleanarray;} { tries to hold the amount of times which the current tree is processed } t_times: longint; fpuvaroffset : byte; constructor create; function getregisterfpu(list: taasmoutput) : tregister; procedure ungetregisterfpu(list: taasmoutput; r : tregister); { pushes and restores registers } procedure saveusedfpuregisters(list:Taasmoutput; var saved:Tpushedsavedfpu; const s:Tcpuregisterset); procedure restoreusedfpuregisters(list:Taasmoutput; const saved:Tpushedsavedfpu); { corrects the fpu stack register by ofs } function correct_fpuregister(r : tregister;ofs : byte) : tregister; end; implementation uses systems, verbose; const { This value is used in tsaved. If the array value is equal to this, then this means that this register is not used.} reg_not_saved = $7fffffff; {************************************************************************ trgcpu *************************************************************************} procedure trgcpu.add_constraints(reg:Tregister); var supreg : tsuperregister; begin if getsubreg(reg) in [R_SUBL,R_SUBH] then begin supreg:=getsupreg(reg); {These registers have no 8-bit subregister, so add interferences.} add_edge(supreg,RS_ESI); add_edge(supreg,RS_EDI); add_edge(supreg,RS_EBP); end; end; {****************************************************************************** Trgobj ******************************************************************************} constructor Trgx86fpu.create; var i:Tsuperregister; begin used_in_proc:=[]; t_times := 0; unusedregsfpu:=usableregsfpu; end; function trgx86fpu.getregisterfpu(list: taasmoutput) : tregister; begin { note: don't return R_ST0, see comments above implementation of } { a_loadfpu_* methods in cgcpu (JM) } result:=NR_ST; end; procedure trgx86fpu.ungetregisterfpu(list : taasmoutput; r : tregister); begin { nothing to do, fpu stack management is handled by the load/ } { store operations in cgcpu (JM) } end; function trgx86fpu.correct_fpuregister(r : tregister;ofs : byte) : tregister; begin correct_fpuregister:=r; setsupreg(correct_fpuregister,ofs); end; procedure trgx86fpu.saveusedfpuregisters(list: taasmoutput; var saved : tpushedsavedfpu; const s: tcpuregisterset); var r : tregister; hr : treference; begin used_in_proc:=used_in_proc+s; {$warning TODO firstsavefpureg} (* { don't try to save the fpu registers if not desired (e.g. for } { the 80x86) } if firstsavefpureg <> R_NO then for r.enum:=firstsavefpureg to lastsavefpureg do begin saved[r.enum].ofs:=reg_not_saved; { if the register is used by the calling subroutine and if } { it's not a regvar (those are handled separately) } if not is_reg_var_other[r.enum] and (r.enum in s) and { and is present in use } not(r.enum in unusedregsfpu) then begin { then save it } tg.GetTemp(list,extended_size,tt_persistent,hr); saved[r.enum].ofs:=hr.offset; cg.a_loadfpu_reg_ref(list,OS_FLOAT,r,hr); cg.a_reg_dealloc(list,r); include(unusedregsfpu,r.enum); inc(countunusedregsfpu); end; end; *) end; procedure trgx86fpu.restoreusedfpuregisters(list : taasmoutput; const saved : tpushedsavedfpu); var r,r2 : tregister; hr : treference; begin {$warning TODO firstsavefpureg} (* if firstsavefpureg <> R_NO then for r.enum:=lastsavefpureg downto firstsavefpureg do begin if saved[r.enum].ofs <> reg_not_saved then begin r2.enum:=R_INTREGISTER; r2.number:=NR_FRAME_POINTER_REG; reference_reset_base(hr,r2,saved[r.enum].ofs); cg.a_reg_alloc(list,r); cg.a_loadfpu_ref_reg(list,OS_FLOAT,hr,r); if not (r.enum in unusedregsfpu) then { internalerror(10) in n386cal we always save/restore the reg *state* using save/restoreunusedstate -> the current state may not be real (JM) } else begin dec(countunusedregsfpu); exclude(unusedregsfpu,r.enum); end; tg.UnGetTemp(list,hr); end; end; *) end; (* procedure Trgx86fpu.saveotherregvars(list: taasmoutput; const s: totherregisterset); var r: Tregister; begin if not(cs_regvars in aktglobalswitches) then exit; if firstsavefpureg <> NR_NO then for r.enum := firstsavefpureg to lastsavefpureg do if is_reg_var_other[r.enum] and (r.enum in s) then store_regvar(list,r); end; *) end. { $Log$ Revision 1.40 2003-10-17 15:08:34 peter * commented out more obsolete constants Revision 1.39 2003/10/17 14:38:32 peter * 64k registers supported * fixed some memory leaks Revision 1.38 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 1.37 2003/10/09 21:31:37 daniel * Register allocator splitted, ans abstract now Revision 1.36 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 1.35 2003/09/11 11:55:00 florian * improved arm code generation * move some protected and private field around * the temp. register for register parameters/arguments are now released before the move to the parameter register is done. This improves the code in a lot of cases. Revision 1.34 2003/09/09 20:59:27 daniel * Adding register allocation order Revision 1.33 2003/09/07 22:09:35 peter * preparations for different default calling conventions * various RA fixes Revision 1.32 2003/09/03 15:55:01 peter * NEWRA branch merged Revision 1.31.2.3 2003/08/31 13:50:16 daniel * Remove sorting and use pregenerated indexes * Some work on making things compile Revision 1.31.2.2 2003/08/28 18:35:08 peter * tregister changed to cardinal Revision 1.31.2.1 2003/08/27 19:55:54 peter * first tregister patch Revision 1.31 2003/08/20 09:07:00 daniel * New register coding now mandatory, some more convert_registers calls removed. Revision 1.30 2003/08/17 08:48:02 daniel * Another register allocator bug fixed. * cpu_registers set to 6 for i386 Revision 1.29 2003/06/17 16:51:30 peter * cycle fixes Revision 1.28 2003/06/17 16:34:44 jonas * lots of newra fixes (need getfuncretparaloc implementation for i386)! * renamed all_intregisters to volatile_intregisters and made it processor dependent Revision 1.27 2003/06/13 21:19:31 peter * current_procdef removed, use current_procinfo.procdef instead Revision 1.26 2003/06/12 21:12:20 peter * size para for ungetregisterfpu Revision 1.25 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 1.24 2003/06/03 13:01:59 daniel * Register allocator finished Revision 1.23 2003/06/01 21:38:06 peter * getregisterfpu size parameter added * op_const_reg size parameter added * sparc updates Revision 1.22 2003/05/16 14:33:31 peter * regvar fixes Revision 1.21 2003/04/25 08:25:26 daniel * Ifdefs around a lot of calls to cleartempgen * Fixed registers that are allocated but not freed in several nodes * Tweak to register allocator to cause less spills * 8-bit registers now interfere with esi,edi and ebp Compiler can now compile rtl successfully when using new register allocator Revision 1.20 2003/04/23 14:42:08 daniel * Further register allocator work. Compiler now smaller with new allocator than without. * Somebody forgot to adjust ppu version number Revision 1.19 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 1.18 2003/04/21 19:16:50 peter * count address regs separate Revision 1.17 2003/03/28 19:16:57 peter * generic constructor working for i386 * remove fixed self register * esi added as address register for i386 Revision 1.16 2003/03/17 15:52:57 peter * SUPPORT_MMX define compile fix Revision 1.15 2003/03/08 13:59:17 daniel * Work to handle new register notation in ag386nsm + Added newra version of Ti386moddivnode Revision 1.14 2003/03/08 08:59:07 daniel + $define newra will enable new register allocator + getregisterint will return imaginary registers with $newra + -sr switch added, will skip register allocation so you can see the direct output of the code generator before register allocation Revision 1.13 2003/03/07 21:57:53 daniel * Improved getregisterint Revision 1.12 2003/02/19 22:00:16 daniel * Code generator converted to new register notation - Horribily outdated todo.txt removed Revision 1.11 2003/01/08 18:43:57 daniel * Tregister changed into a record Revision 1.10 2002/10/05 12:43:29 carl * fixes for Delphi 6 compilation (warning : Some features do not work under Delphi) Revision 1.9 2002/08/17 09:23:48 florian * first part of procinfo rewrite Revision 1.8 2002/07/01 18:46:34 peter * internal linker * reorganized aasm layer Revision 1.7 2002/05/16 19:46:52 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 1.6 2002/05/12 16:53:18 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 1.5 2002/04/21 15:43:32 carl * changeregsize -> rg.makeregsize * changeregsize moved from cpubase to here Revision 1.4 2002/04/15 19:44:22 peter * fixed stackcheck that would be called recursively when a stack error was found * generic changeregsize(reg,size) for i386 register resizing * removed some more routines from cga unit * fixed returnvalue handling * fixed default stacksize of linux and go32v2, 8kb was a bit small :-) Revision 1.3 2002/04/04 19:06:13 peter * removed unused units * use tlocation.size in cg.a_*loc*() routines Revision 1.2 2002/04/02 17:11:39 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 1.1 2002/03/31 20:26:40 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 }