fpc/compiler/mips/cgcpu.pas
2024-05-15 22:52:24 +02:00

2049 lines
70 KiB
ObjectPascal

{
Copyright (c) 1998-2012 by Florian Klaempfl and David Zhang
This unit implements the code generator for MIPS
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 cgcpu;
{$i fpcdefs.inc}
interface
uses
globtype, parabase,
cgbase, cgutils, cgobj,
{$ifndef mips64}
cg64f32,
{$endif mips64}
cpupara,
aasmbase, aasmtai, aasmcpu, aasmdata,
cpubase, cpuinfo,
node, symconst, SymType, symdef,
rgcpu;
type
TCGMIPS = class(tcg)
public
procedure init_register_allocators; override;
procedure done_register_allocators; override;
/// { needed by cg64 }
procedure make_simple_ref(list: tasmlist; var ref: treference);
procedure handle_reg_const_reg(list: tasmlist; op: Tasmop; src: tregister; a: tcgint; dst: tregister);
procedure maybeadjustresult(list: TAsmList; op: TOpCg; size: tcgsize; dst: tregister);
procedure overflowcheck_internal(list: TAsmList; arg1, arg2: TRegister);
{ parameter }
procedure a_loadfpu_reg_cgpara(list: tasmlist; size: tcgsize; const r: tregister; const paraloc: TCGPara); override;
procedure a_loadfpu_ref_cgpara(list: tasmlist; size: tcgsize; const ref: treference; const paraloc: TCGPara); override;
procedure a_call_name(list: tasmlist; const s: string; weak : boolean); override;
procedure a_call_reg(list: tasmlist; Reg: TRegister); override;
procedure a_call_sym_pic(list: tasmlist; sym: tasmsymbol);
{ General purpose instructions }
procedure a_op_const_reg(list: tasmlist; Op: TOpCG; size: tcgsize; a: tcgint; reg: TRegister); override;
procedure a_op_reg_reg(list: tasmlist; Op: TOpCG; size: TCGSize; src, dst: TRegister); override;
procedure a_op_const_reg_reg(list: tasmlist; op: TOpCg; size: tcgsize; a: tcgint; src, dst: tregister); override;
procedure a_op_reg_reg_reg(list: tasmlist; op: TOpCg; size: tcgsize; src1, src2, dst: tregister); override;
procedure a_op_const_reg_reg_checkoverflow(list: tasmlist; op: TOpCg; size: tcgsize; a: tcgint; src, dst: tregister; setflags: boolean; var ovloc: tlocation); override;
procedure a_op_reg_reg_reg_checkoverflow(list: tasmlist; op: TOpCg; size: tcgsize; src1, src2, dst: tregister; setflags: boolean; var ovloc: tlocation); override;
{ move instructions }
procedure a_load_const_reg(list: tasmlist; size: tcgsize; a: tcgint; reg: tregister); override;
procedure a_load_const_ref(list: tasmlist; size: tcgsize; a: tcgint; const ref: TReference); override;
procedure a_load_reg_ref(list: tasmlist; FromSize, ToSize: TCgSize; reg: TRegister; const ref: TReference); override;
procedure a_load_ref_reg(list: tasmlist; FromSize, ToSize: TCgSize; const ref: TReference; reg: tregister); override;
procedure a_load_reg_reg(list: tasmlist; FromSize, ToSize: TCgSize; reg1, reg2: tregister); override;
procedure a_loadaddr_ref_reg(list: tasmlist; const ref: TReference; r: tregister); override;
{ fpu move instructions }
procedure a_loadfpu_reg_reg(list: tasmlist; fromsize, tosize: tcgsize; reg1, reg2: tregister); override;
procedure a_loadfpu_ref_reg(list: tasmlist; fromsize, tosize: tcgsize; const ref: TReference; reg: tregister); override;
procedure a_loadfpu_reg_ref(list: tasmlist; fromsize, tosize: tcgsize; reg: tregister; const ref: TReference); override;
{ comparison operations }
procedure a_cmp_const_reg_label(list: tasmlist; size: tcgsize; cmp_op: topcmp; a: tcgint; reg: tregister; l: tasmlabel); override;
procedure a_cmp_reg_reg_label(list: tasmlist; size: tcgsize; cmp_op: topcmp; reg1, reg2: tregister; l: tasmlabel); override;
procedure a_jmp_flags(list: tasmlist; const f: TResFlags; l: tasmlabel); override;
procedure g_flags2reg(list: tasmlist; size: TCgSize; const f: TResFlags; reg: tregister); override;
procedure a_jmp_always(List: tasmlist; l: TAsmLabel); override;
procedure a_jmp_name(list: tasmlist; const s: string); override;
procedure a_mul_reg_reg_pair(list: tasmlist; size: tcgsize; src1,src2,dstlo,dsthi: tregister); override;
procedure g_overflowCheck(List: tasmlist; const Loc: TLocation; def: TDef); override;
procedure g_overflowCheck_loc(List: tasmlist; const Loc: TLocation; def: TDef; ovloc: tlocation); override;
procedure g_proc_entry(list: tasmlist; localsize: longint; nostackframe: boolean); override;
procedure g_proc_exit(list: tasmlist; parasize: longint; nostackframe: boolean); override;
procedure g_concatcopy(list: tasmlist; const Source, dest: treference; len: tcgint); override;
procedure g_concatcopy_unaligned(list: tasmlist; const Source, dest: treference; len: tcgint); override;
procedure g_adjust_self_value(list:TAsmList;procdef: tprocdef;ioffset: tcgint); override;
procedure g_profilecode(list: TAsmList);override;
end;
{$ifndef mips64}
TCg64MPSel = class(tcg64f32)
public
procedure a_load64_reg_ref(list: tasmlist; reg: tregister64; const ref: treference); override;
procedure a_load64_ref_reg(list: tasmlist; const ref: treference; reg: tregister64); override;
procedure a_load64_ref_cgpara(list: tasmlist; const r: treference; const paraloc: tcgpara); override;
procedure a_op64_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; regsrc, regdst: TRegister64); override;
procedure a_op64_const_reg(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regdst: TRegister64); override;
procedure a_op64_const_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regsrc, regdst: tregister64); override;
procedure a_op64_reg_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; regsrc1, regsrc2, regdst: tregister64); override;
procedure a_op64_const_reg_reg_checkoverflow(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regsrc, regdst: tregister64; setflags: boolean; var ovloc: tlocation); override;
procedure a_op64_reg_reg_reg_checkoverflow(list: tasmlist; op: TOpCG; size: tcgsize; regsrc1, regsrc2, regdst: tregister64; setflags: boolean; var ovloc: tlocation); override;
end;
{$endif mips64}
procedure create_codegen;
const
TOpCmp2AsmCond : array[topcmp] of TAsmCond=(C_NONE,
C_EQ,C_GT,C_LT,C_GE,C_LE,C_NE,C_LEU,C_LTU,C_GEU,C_GTU
);
implementation
uses
globals, verbose, systems, cutils,
paramgr, fmodule,
symtable, symsym,
tgobj,
procinfo, cpupi;
const
TOpcg2AsmOp: array[TOpCg] of TAsmOp = (
A_NONE,A_NONE,A_ADDU,A_AND,A_NONE,A_NONE,A_MULT,A_MULTU,A_NONE,A_NONE,
A_OR,A_SRAV,A_SLLV,A_SRLV,A_SUBU,A_XOR,A_NONE,A_NONE
);
procedure TCGMIPS.make_simple_ref(list: tasmlist; var ref: treference);
var
tmpreg, tmpreg1: tregister;
tmpref: treference;
base_replaced: boolean;
begin
{ Enforce some discipline for callers:
- gp is always implicit
- reference is processed only once }
if (ref.base=NR_GP) or (ref.index=NR_GP) then
InternalError(2013022801);
if (ref.refaddr<>addr_no) then
InternalError(2013022802);
{ fixup base/index, if both are present then add them together }
base_replaced:=false;
tmpreg:=ref.base;
if (tmpreg=NR_NO) then
tmpreg:=ref.index
else if (ref.index<>NR_NO) then
begin
tmpreg:=getintregister(list,OS_ADDR);
list.concat(taicpu.op_reg_reg_reg(A_ADDU,tmpreg,ref.base,ref.index));
base_replaced:=true;
end;
ref.base:=tmpreg;
ref.index:=NR_NO;
if (ref.symbol=nil) and
(ref.offset>=simm16lo) and
(ref.offset<=simm16hi-sizeof(pint)) then
exit;
{ Symbol present or offset > 16bits }
if assigned(ref.symbol) then
begin
ref.base:=getintregister(list,OS_ADDR);
reference_reset_symbol(tmpref,ref.symbol,ref.offset,ref.alignment,ref.volatility);
if (cs_create_pic in current_settings.moduleswitches) then
begin
if not (pi_needs_got in current_procinfo.flags) then
InternalError(2013060102);
{ For PIC global symbols offset must be handled separately.
Otherwise (non-PIC or local symbols) offset can be encoded
into relocation even if exceeds 16 bits. }
if (ref.symbol.bind<>AB_LOCAL) then
tmpref.offset:=0;
tmpref.refaddr:=addr_pic;
tmpref.base:=NR_GP;
list.concat(taicpu.op_reg_ref(A_LW,ref.base,tmpref));
end
else
begin
tmpref.refaddr:=addr_high;
list.concat(taicpu.op_reg_ref(A_LUI,ref.base,tmpref));
end;
{ Add original base/index, if any. }
if (tmpreg<>NR_NO) then
list.concat(taicpu.op_reg_reg_reg(A_ADDU,ref.base,tmpreg,ref.base));
if (ref.symbol.bind=AB_LOCAL) or
not (cs_create_pic in current_settings.moduleswitches) then
begin
ref.refaddr:=addr_low;
exit;
end;
{ PIC global symbol }
ref.symbol:=nil;
if (ref.offset>=simm16lo) and
(ref.offset<=simm16hi-sizeof(pint)) then
exit;
{ fallthrough to the case of large offset }
end;
tmpreg1:=getintregister(list,OS_INT);
a_load_const_reg(list,OS_INT,ref.offset,tmpreg1);
if (ref.base=NR_NO) then
ref.base:=tmpreg1 { offset alone, weird but possible }
else
begin
tmpreg:=ref.base;
if (not base_replaced) then
ref.base:=getintregister(list,OS_ADDR);
list.concat(taicpu.op_reg_reg_reg(A_ADDU,ref.base,tmpreg,tmpreg1))
end;
ref.offset:=0;
end;
procedure TCGMIPS.handle_reg_const_reg(list: tasmlist; op: Tasmop; src: tregister; a: tcgint; dst: tregister);
var
tmpreg: tregister;
op2: Tasmop;
negate: boolean;
begin
case op of
A_ADD,A_SUB:
op2:=A_ADDI;
A_ADDU,A_SUBU:
op2:=A_ADDIU;
else
InternalError(2013052001);
end;
negate:=op in [A_SUB,A_SUBU];
{ subtraction is actually addition of negated value, so possible range is
off by one (-32767..32768) }
if (a < simm16lo+ord(negate)) or
(a > simm16hi+ord(negate)) then
begin
tmpreg := GetIntRegister(list, OS_INT);
a_load_const_reg(list, OS_INT, a, tmpreg);
list.concat(taicpu.op_reg_reg_reg(op, dst, src, tmpreg));
end
else
begin
if negate then
a:=-a;
list.concat(taicpu.op_reg_reg_const(op2, dst, src, a));
end;
end;
{****************************************************************************
Assembler code
****************************************************************************}
procedure TCGMIPS.init_register_allocators;
begin
inherited init_register_allocators;
{ Keep RS_R25, i.e. $t9 for PIC call }
if (cs_create_pic in current_settings.moduleswitches) and assigned(current_procinfo) and
(pi_needs_got in current_procinfo.flags) then
begin
current_procinfo.got := NR_GP;
rg[R_INTREGISTER] := Trgintcpu.Create(R_INTREGISTER, R_SUBD,
[RS_R2,RS_R3,RS_R4,RS_R5,RS_R6,RS_R7,RS_R8,RS_R9,
RS_R10,RS_R11,RS_R12,RS_R13,RS_R14,RS_R15,RS_R16,RS_R17,RS_R18,RS_R19,
RS_R20,RS_R21,RS_R22,RS_R23,RS_R24{,RS_R25}],
first_int_imreg, []);
end
else
rg[R_INTREGISTER] := trgintcpu.Create(R_INTREGISTER, R_SUBD,
[RS_R2,RS_R3,RS_R4,RS_R5,RS_R6,RS_R7,RS_R8,RS_R9,
RS_R10,RS_R11,RS_R12,RS_R13,RS_R14,RS_R15,RS_R16,RS_R17,RS_R18,RS_R19,
RS_R20,RS_R21,RS_R22,RS_R23,RS_R24{,RS_R25}],
first_int_imreg, []);
{
rg[R_FPUREGISTER] := trgcpu.Create(R_FPUREGISTER, R_SUBFS,
[RS_F0,RS_F1,RS_F2,RS_F3,RS_F4,RS_F5,RS_F6,RS_F7,
RS_F8,RS_F9,RS_F10,RS_F11,RS_F12,RS_F13,RS_F14,RS_F15,
RS_F16,RS_F17,RS_F18,RS_F19,RS_F20,RS_F21,RS_F22,RS_F23,
RS_F24,RS_F25,RS_F26,RS_F27,RS_F28,RS_F29,RS_F30,RS_F31],
first_fpu_imreg, []);
}
rg[R_FPUREGISTER] := trgcpu.Create(R_FPUREGISTER, R_SUBFS,
[RS_F0,RS_F2,RS_F4,RS_F6, RS_F8,RS_F10,RS_F12,RS_F14,
RS_F16,RS_F18,RS_F20,RS_F22, RS_F24,RS_F26,RS_F28,RS_F30],
first_fpu_imreg, []);
end;
procedure TCGMIPS.done_register_allocators;
begin
rg[R_INTREGISTER].Free;
rg[R_FPUREGISTER].Free;
inherited done_register_allocators;
end;
procedure TCGMIPS.a_loadfpu_ref_cgpara(list: tasmlist; size: tcgsize; const ref: treference; const paraloc: TCGPara);
var
href, href2: treference;
hloc: pcgparalocation;
begin
{ TODO: inherited cannot deal with individual locations for each of OS_32 registers.
Must change parameter management to allocate a single 64-bit register pair,
then this method can be removed. }
href := ref;
hloc := paraloc.location;
while assigned(hloc) do
begin
paramanager.allocparaloc(list,hloc);
case hloc^.loc of
LOC_REGISTER:
a_load_ref_reg(list, hloc^.size, hloc^.size, href, hloc^.Register);
LOC_FPUREGISTER,LOC_CFPUREGISTER :
a_loadfpu_ref_reg(list,hloc^.size,hloc^.size,href,hloc^.register);
LOC_REFERENCE:
begin
paraloc.check_simple_location;
reference_reset_base(href2,paraloc.location^.reference.index,paraloc.location^.reference.offset,ctempposinvalid,paraloc.alignment,[]);
{ concatcopy should choose the best way to copy the data }
g_concatcopy(list,ref,href2,tcgsize2size[size]);
end;
else
internalerror(200408241);
end;
Inc(href.offset, tcgsize2size[hloc^.size]);
hloc := hloc^.Next;
end;
end;
procedure TCGMIPS.a_loadfpu_reg_cgpara(list: tasmlist; size: tcgsize; const r: tregister; const paraloc: TCGPara);
var
href: treference;
begin
if paraloc.Location^.next=nil then
begin
inherited a_loadfpu_reg_cgpara(list,size,r,paraloc);
exit;
end;
tg.GetTemp(list, TCGSize2Size[size], TCGSize2Size[size], tt_normal, href);
a_loadfpu_reg_ref(list, size, size, r, href);
a_loadfpu_ref_cgpara(list, size, href, paraloc);
tg.Ungettemp(list, href);
end;
procedure TCGMIPS.a_call_sym_pic(list: tasmlist; sym: tasmsymbol);
var
href: treference;
begin
reference_reset_symbol(href,sym,0,sizeof(aint),[]);
if (sym.bind=AB_LOCAL) then
href.refaddr:=addr_pic
else
href.refaddr:=addr_pic_call16;
href.base:=NR_GP;
list.concat(taicpu.op_reg_ref(A_LW,NR_PIC_FUNC,href));
if (sym.bind=AB_LOCAL) then
begin
href.refaddr:=addr_low;
href.base:=NR_NO;
list.concat(taicpu.op_reg_ref(A_ADDIU,NR_PIC_FUNC,href));
end;
list.concat(taicpu.op_reg(A_JALR,NR_PIC_FUNC));
{ Delay slot }
list.concat(taicpu.op_none(A_NOP));
{ Restore GP if in PIC mode }
if (cs_create_pic in current_settings.moduleswitches) then
begin
if tcpuprocinfo(current_procinfo).save_gp_ref.offset=0 then
InternalError(2013071001);
list.concat(taicpu.op_reg_ref(A_LW,NR_GP,tcpuprocinfo(current_procinfo).save_gp_ref));
end;
end;
procedure TCGMIPS.a_call_name(list: tasmlist; const s: string; weak: boolean);
var
sym: tasmsymbol;
begin
if assigned(current_procinfo) and
not (pi_do_call in current_procinfo.flags) then
InternalError(2013022101);
if weak then
sym:=current_asmdata.WeakRefAsmSymbol(s,AT_FUNCTION)
else
sym:=current_asmdata.RefAsmSymbol(s,AT_FUNCTION);
if (cs_create_pic in current_settings.moduleswitches) then
a_call_sym_pic(list,sym)
else
begin
list.concat(taicpu.op_sym(A_JAL,sym));
{ Delay slot }
list.concat(taicpu.op_none(A_NOP));
end;
end;
procedure TCGMIPS.a_call_reg(list: tasmlist; Reg: TRegister);
begin
if assigned(current_procinfo) and
not (pi_do_call in current_procinfo.flags) then
InternalError(2013022102);
if (Reg <> NR_PIC_FUNC) then
list.concat(taicpu.op_reg_reg(A_MOVE,NR_PIC_FUNC,reg));
list.concat(taicpu.op_reg(A_JALR,NR_PIC_FUNC));
{ Delay slot }
list.concat(taicpu.op_none(A_NOP));
{ Restore GP if in PIC mode }
if (cs_create_pic in current_settings.moduleswitches) then
begin
if tcpuprocinfo(current_procinfo).save_gp_ref.offset=0 then
InternalError(2013071002);
list.concat(taicpu.op_reg_ref(A_LW,NR_GP,tcpuprocinfo(current_procinfo).save_gp_ref));
end;
end;
{********************** load instructions ********************}
procedure TCGMIPS.a_load_const_reg(list: tasmlist; size: TCGSize; a: tcgint; reg: TRegister);
begin
if (a = 0) then
a_load_reg_reg(list, OS_INT, OS_INT, NR_R0, reg)
else if (a >= simm16lo) and (a <= simm16hi) then
list.concat(taicpu.op_reg_reg_const(A_ADDIU, reg, NR_R0, a))
else if (a>=0) and (a <= 65535) then
list.concat(taicpu.op_reg_reg_const(A_ORI, reg, NR_R0, a))
{$ifdef mips32}
else
{$else}
else if (a>=0) and (a <= high(dword)) then
{$endif}
begin
list.concat(taicpu.op_reg_const(A_LUI, reg, aint(a) shr 16));
if (a and aint($FFFF))<>0 then
list.concat(taicpu.op_reg_reg_const(A_ORI,reg,reg,a and aint($FFFF)));
{$ifdef mips64}
end
else
begin
list.concat(taicpu.op_reg_const(A_LUI, reg, aint(a) shr 48));
if ((a shr 32) and aint($FFFF))<>0 then
list.concat(taicpu.op_reg_reg_const(A_ORI,reg,reg,(a shr 32) and aint($FFFF)));
list.concat(taicpu.op_reg_reg_const(A_SLL, reg, reg, 16));
if ((a shr 16) and aint($FFFF))<>0 then
list.concat(taicpu.op_reg_reg_const(A_ORI,reg,reg,(a shr 16) and aint($FFFF)));
list.concat(taicpu.op_reg_reg_const(A_SLL, reg, reg, 16));
if (a and aint($FFFF))<>0 then
list.concat(taicpu.op_reg_reg_const(A_ORI,reg,reg,a and aint($FFFF)));
{$endif mips64}
end;
end;
procedure TCGMIPS.a_load_const_ref(list: tasmlist; size: tcgsize; a: tcgint; const ref: TReference);
begin
if a = 0 then
a_load_reg_ref(list, size, size, NR_R0, ref)
else
inherited a_load_const_ref(list, size, a, ref);
end;
procedure TCGMIPS.a_load_reg_ref(list: tasmlist; FromSize, ToSize: TCGSize; reg: tregister; const Ref: TReference);
var
op: tasmop;
href: treference;
begin
if (TCGSize2Size[fromsize] < TCGSize2Size[tosize]) then
a_load_reg_reg(list,fromsize,tosize,reg,reg);
if (ref.alignment<>0) and
(ref.alignment<tcgsize2size[tosize]) then
begin
a_load_reg_ref_unaligned(list,FromSize,ToSize,reg,ref);
exit;
end;
case tosize of
OS_8,
OS_S8:
Op := A_SB;
OS_16,
OS_S16:
Op := A_SH;
OS_32,
OS_S32:
Op := A_SW;
OS_64,
OS_S64:
Op := A_SD;
else
InternalError(2002122100);
end;
href:=ref;
make_simple_ref(list,href);
list.concat(taicpu.op_reg_ref(op,reg,href));
end;
procedure TCGMIPS.a_load_ref_reg(list: tasmlist; FromSize, ToSize: TCgSize; const ref: TReference; reg: tregister);
var
op: tasmop;
href: treference;
begin
if (TCGSize2Size[fromsize] >= TCGSize2Size[tosize]) then
fromsize := tosize;
if (ref.alignment<>0) and
(ref.alignment<min(tcgsize2size[fromsize],sizeof(aint))) then
begin
a_load_ref_reg_unaligned(list,FromSize,ToSize,ref,reg);
exit;
end;
case fromsize of
OS_S8:
Op := A_LB;{Load Signed Byte}
OS_8:
Op := A_LBU;{Load Unsigned Byte}
OS_S16:
Op := A_LH;{Load Signed Halfword}
OS_16:
Op := A_LHU;{Load Unsigned Halfword}
OS_S32:
Op := A_LW;{Load Word}
OS_32:
Op := A_LW;//A_LWU;{Load Unsigned Word}
OS_S64,
OS_64:
Op := A_LD;{Load a Long Word}
else
InternalError(2002122101);
end;
href:=ref;
make_simple_ref(list,href);
list.concat(taicpu.op_reg_ref(op,reg,href));
if (fromsize=OS_S8) and (tosize=OS_16) then
a_load_reg_reg(list,fromsize,tosize,reg,reg);
end;
procedure TCGMIPS.a_load_reg_reg(list: tasmlist; fromsize, tosize: tcgsize; reg1, reg2: tregister);
var
instr: taicpu;
done: boolean;
begin
if (tcgsize2size[tosize] < tcgsize2size[fromsize]) or
(
(tcgsize2size[tosize] = tcgsize2size[fromsize]) and (tosize <> fromsize)
) or ((fromsize = OS_S8) and
(tosize = OS_16)) then
begin
done:=true;
case tosize of
OS_8:
list.concat(taicpu.op_reg_reg_const(A_ANDI, reg2, reg1, $ff));
OS_16:
list.concat(taicpu.op_reg_reg_const(A_ANDI, reg2, reg1, $ffff));
{$ifdef cpu64bitalu}
OS_64,
OS_S64,
{$endif cpu64bitalu}
OS_32,
OS_S32:
done:=false;
OS_S8:
begin
if (CPUMIPS_HAS_ISA32R2 in cpu_capabilities[current_settings.cputype]) then
list.concat(taicpu.op_reg_reg(A_SEB,reg2,reg1))
else
begin
list.concat(taicpu.op_reg_reg_const(A_SLL, reg2, reg1, 24));
list.concat(taicpu.op_reg_reg_const(A_SRA, reg2, reg2, 24));
end;
end;
OS_S16:
begin
if (CPUMIPS_HAS_ISA32R2 in cpu_capabilities[current_settings.cputype]) then
list.concat(taicpu.op_reg_reg(A_SEH,reg2,reg1))
else
begin
list.concat(taicpu.op_reg_reg_const(A_SLL, reg2, reg1, 16));
list.concat(taicpu.op_reg_reg_const(A_SRA, reg2, reg2, 16));
end;
end;
else
internalerror(2002090901);
end;
end
else
done:=false;
if (not done) and (reg1 <> reg2) then
begin
{ same size, only a register mov required }
instr := taicpu.op_reg_reg(A_MOVE, reg2, reg1);
list.Concat(instr);
{ Notify the register allocator that we have written a move instruction so
it can try to eliminate it. }
add_move_instruction(instr);
end;
end;
procedure TCGMIPS.a_loadaddr_ref_reg(list: tasmlist; const ref: TReference; r: tregister);
var
href: treference;
hreg: tregister;
begin
{ Enforce some discipline for callers:
- reference must be a "raw" one and not use gp }
if (ref.base=NR_GP) or (ref.index=NR_GP) then
InternalError(2013022803);
if (ref.refaddr<>addr_no) then
InternalError(2013022804);
if (ref.base=NR_NO) and (ref.index<>NR_NO) then
InternalError(200306171);
if (ref.symbol=nil) then
begin
if (ref.base<>NR_NO) then
begin
if (ref.offset<simm16lo) or (ref.offset>simm16hi) then
begin
hreg:=getintregister(list,OS_INT);
a_load_const_reg(list,OS_INT,ref.offset,hreg);
list.concat(taicpu.op_reg_reg_reg(A_ADDU,r,ref.base,hreg));
end
else if (ref.offset<>0) then
list.concat(taicpu.op_reg_reg_const(A_ADDIU,r,ref.base,ref.offset))
else
a_load_reg_reg(list,OS_INT,OS_INT,ref.base,r); { emit optimizable move }
if (ref.index<>NR_NO) then
list.concat(taicpu.op_reg_reg_reg(A_ADDU,r,r,ref.index));
end
else
a_load_const_reg(list,OS_INT,ref.offset,r);
exit;
end;
reference_reset_symbol(href,ref.symbol,ref.offset,ref.alignment,ref.volatility);
if (cs_create_pic in current_settings.moduleswitches) then
begin
if not (pi_needs_got in current_procinfo.flags) then
InternalError(2013060104);
{ For PIC global symbols offset must be handled separately.
Otherwise (non-PIC or local symbols) offset can be encoded
into relocation even if exceeds 16 bits. }
if (href.symbol.bind<>AB_LOCAL) then
href.offset:=0;
href.refaddr:=addr_pic;
href.base:=NR_GP;
list.concat(taicpu.op_reg_ref(A_LW,r,href));
end
else
begin
href.refaddr:=addr_high;
list.concat(taicpu.op_reg_ref(A_LUI,r,href));
end;
{ Add original base/index, if any. }
if (ref.base<>NR_NO) then
list.concat(taicpu.op_reg_reg_reg(A_ADDU,r,r,ref.base));
if (ref.index<>NR_NO) then
list.concat(taicpu.op_reg_reg_reg(A_ADDU,r,r,ref.index));
{ add low part if necessary }
if (ref.symbol.bind=AB_LOCAL) or
not (cs_create_pic in current_settings.moduleswitches) then
begin
href.refaddr:=addr_low;
href.base:=NR_NO;
list.concat(taicpu.op_reg_reg_ref(A_ADDIU,r,r,href));
exit;
end;
if (ref.offset<simm16lo) or (ref.offset>simm16hi) then
begin
hreg:=getintregister(list,OS_INT);
a_load_const_reg(list,OS_INT,ref.offset,hreg);
list.concat(taicpu.op_reg_reg_reg(A_ADDU,r,r,hreg));
end
else if (ref.offset<>0) then
list.concat(taicpu.op_reg_reg_const(A_ADDIU,r,r,ref.offset));
end;
procedure TCGMIPS.a_loadfpu_reg_reg(list: tasmlist; fromsize, tosize: tcgsize; reg1, reg2: tregister);
const
FpuMovInstr: array[OS_F32..OS_F64,OS_F32..OS_F64] of TAsmOp =
((A_MOV_S, A_CVT_D_S),(A_CVT_S_D,A_MOV_D));
var
instr: taicpu;
begin
if (reg1 <> reg2) or (fromsize<>tosize) then
begin
instr := taicpu.op_reg_reg(fpumovinstr[fromsize,tosize], reg2, reg1);
list.Concat(instr);
{ Notify the register allocator that we have written a move instruction so
it can try to eliminate it. }
if (fromsize=tosize) then
add_move_instruction(instr);
end;
end;
procedure TCGMIPS.a_loadfpu_ref_reg(list: tasmlist; fromsize, tosize: tcgsize; const ref: TReference; reg: tregister);
var
href: TReference;
begin
href:=ref;
make_simple_ref(list,href);
case fromsize of
OS_F32:
list.concat(taicpu.op_reg_ref(A_LWC1,reg,href));
OS_F64:
list.concat(taicpu.op_reg_ref(A_LDC1,reg,href));
else
InternalError(2007042701);
end;
if tosize<>fromsize then
a_loadfpu_reg_reg(list,fromsize,tosize,reg,reg);
end;
procedure TCGMIPS.a_loadfpu_reg_ref(list: tasmlist; fromsize, tosize: tcgsize; reg: tregister; const ref: TReference);
var
href: TReference;
begin
if tosize<>fromsize then
a_loadfpu_reg_reg(list,fromsize,tosize,reg,reg);
href:=ref;
make_simple_ref(list,href);
case tosize of
OS_F32:
list.concat(taicpu.op_reg_ref(A_SWC1,reg,href));
OS_F64:
list.concat(taicpu.op_reg_ref(A_SDC1,reg,href));
else
InternalError(2007042702);
end;
end;
procedure TCGMIPS.maybeadjustresult(list: TAsmList; op: TOpCg; size: tcgsize; dst: tregister);
const
overflowops = [OP_MUL,OP_SHL,OP_ADD,OP_SUB,OP_NOT,OP_NEG];
begin
if (op in overflowops) and
(size in [OS_8,OS_S8,OS_16,OS_S16]) then
a_load_reg_reg(list,OS_32,size,dst,dst);
end;
procedure TCGMIPS.overflowcheck_internal(list: tasmlist; arg1, arg2: tregister);
var
carry, hreg: tregister;
begin
if (arg1=arg2) then
InternalError(2013050501);
carry:=GetIntRegister(list,OS_INT);
hreg:=GetIntRegister(list,OS_INT);
list.concat(taicpu.op_reg_reg_reg(A_SLTU,carry,arg1,arg2));
{ if carry<>0, this will cause hardware overflow interrupt }
a_load_const_reg(list,OS_INT,$80000000,hreg);
list.concat(taicpu.op_reg_reg_reg(A_SUB,hreg,hreg,carry));
end;
const
ops_add: array[boolean] of TAsmOp = (A_ADDU, A_ADD);
ops_sub: array[boolean] of TAsmOp = (A_SUBU, A_SUB);
ops_slt: array[boolean] of TAsmOp = (A_SLTU, A_SLT);
ops_slti: array[boolean] of TAsmOp = (A_SLTIU, A_SLTI);
ops_and: array[boolean] of TAsmOp = (A_AND, A_ANDI);
ops_or: array[boolean] of TAsmOp = (A_OR, A_ORI);
ops_xor: array[boolean] of TasmOp = (A_XOR, A_XORI);
procedure TCGMIPS.a_op_const_reg(list: tasmlist; Op: TOpCG; size: tcgsize; a: tcgint; reg: TRegister);
begin
optimize_op_const(size,op,a);
case op of
OP_NONE:
exit;
OP_MOVE:
a_load_const_reg(list,size,a,reg);
OP_NEG,OP_NOT:
internalerror(200306011);
else
a_op_const_reg_reg(list,op,size,a,reg,reg);
end;
end;
procedure TCGMIPS.a_op_reg_reg(list: tasmlist; Op: TOpCG; size: TCGSize; src, dst: TRegister);
begin
case Op of
OP_NEG:
list.concat(taicpu.op_reg_reg_reg(A_SUBU, dst, NR_R0, src));
OP_NOT:
list.concat(taicpu.op_reg_reg_reg(A_NOR, dst, NR_R0, src));
OP_IMUL,OP_MUL:
begin
list.concat(taicpu.op_reg_reg(TOpcg2AsmOp[op], dst, src));
list.concat(taicpu.op_reg(A_MFLO, dst));
end;
else
a_op_reg_reg_reg(list, op, size, src, dst, dst);
exit;
end;
maybeadjustresult(list,op,size,dst);
end;
procedure TCGMIPS.a_op_const_reg_reg(list: tasmlist; op: TOpCg; size: tcgsize; a: tcgint; src, dst: tregister);
var
l: TLocation;
begin
a_op_const_reg_reg_checkoverflow(list, op, size, a, src, dst, false, l);
end;
procedure TCGMIPS.a_op_reg_reg_reg(list: tasmlist; op: TOpCg; size: tcgsize; src1, src2, dst: tregister);
begin
if (TOpcg2AsmOp[op]=A_NONE) then
InternalError(2013070305);
if (op=OP_SAR) then
begin
if (size in [OS_S8,OS_S16]) then
begin
{ Sign-extend before shiting }
list.concat(taicpu.op_reg_reg_const(A_SLL, dst, src2, 32-(tcgsize2size[size]*8)));
list.concat(taicpu.op_reg_reg_const(A_SRA, dst, dst, 32-(tcgsize2size[size]*8)));
src2:=dst;
end
else if not (size in [OS_32,OS_S32]) then
InternalError(2013070306);
end;
list.concat(taicpu.op_reg_reg_reg(TOpCG2AsmOp[op], dst, src2, src1));
maybeadjustresult(list,op,size,dst);
end;
procedure TCGMIPS.a_op_const_reg_reg_checkoverflow(list: tasmlist; op: TOpCg; size: tcgsize; a: tcgint; src, dst: tregister; setflags: boolean; var ovloc: tlocation);
var
signed,immed: boolean;
hreg: TRegister;
asmop: TAsmOp;
begin
a:=aint(a);
ovloc.loc := LOC_VOID;
optimize_op_const(size,op,a);
signed:=(size in [OS_S8,OS_S16,OS_S32]);
if (setflags and (not signed) and (src=dst) and (op in [OP_ADD,OP_SUB])) then
hreg:=GetIntRegister(list,OS_INT)
else
hreg:=dst;
case op of
OP_NONE:
a_load_reg_reg(list,size,size,src,dst);
OP_MOVE:
a_load_const_reg(list,size,a,dst);
OP_ADD:
begin
handle_reg_const_reg(list,ops_add[setflags and signed],src,a,hreg);
if setflags and (not signed) then
overflowcheck_internal(list,hreg,src);
{ does nothing if hreg=dst }
a_load_reg_reg(list,OS_INT,OS_INT,hreg,dst);
end;
OP_SUB:
begin
handle_reg_const_reg(list,ops_sub[setflags and signed],src,a,hreg);
if setflags and (not signed) then
overflowcheck_internal(list,src,hreg);
a_load_reg_reg(list,OS_INT,OS_INT,hreg,dst);
end;
OP_MUL,OP_IMUL:
begin
hreg:=GetIntRegister(list,OS_INT);
a_load_const_reg(list,OS_INT,a,hreg);
a_op_reg_reg_reg_checkoverflow(list,op,size,src,hreg,dst,setflags,ovloc);
exit;
end;
OP_AND,OP_OR,OP_XOR:
begin
{ logical operations zero-extend, not sign-extend, the immediate }
immed:=(a>=0) and (a<=65535);
case op of
OP_AND: asmop:=ops_and[immed];
OP_OR: asmop:=ops_or[immed];
OP_XOR: asmop:=ops_xor[immed];
else
InternalError(2013050401);
end;
if immed then
list.concat(taicpu.op_reg_reg_const(asmop,dst,src,a))
else
begin
hreg:=GetIntRegister(list,OS_INT);
a_load_const_reg(list,OS_INT,a,hreg);
list.concat(taicpu.op_reg_reg_reg(asmop,dst,src,hreg));
end;
end;
OP_SHL:
list.concat(taicpu.op_reg_reg_const(A_SLL,dst,src,a));
OP_SHR:
list.concat(taicpu.op_reg_reg_const(A_SRL,dst,src,a));
OP_SAR:
begin
if (size in [OS_S8,OS_S16]) then
begin
list.concat(taicpu.op_reg_reg_const(A_SLL,dst,src,32-(tcgsize2size[size]*8)));
inc(a,32-tcgsize2size[size]*8);
src:=dst;
end
{$ifdef MIPS64}
else if (size in [OS_64,OS_S64]) then
list.concat(taicpu.op_reg_reg_const(A_DSRA,dst,src,a))
{$endif MIPS64}
else if not (size in [OS_32,OS_S32]) then
InternalError(2013070303);
list.concat(taicpu.op_reg_reg_const(A_SRA,dst,src,a));
end;
else
internalerror(2007012601);
end;
maybeadjustresult(list,op,size,dst);
end;
procedure TCGMIPS.a_op_reg_reg_reg_checkoverflow(list: tasmlist; op: TOpCg; size: tcgsize; src1, src2, dst: tregister; setflags: boolean; var ovloc: tlocation);
var
signed: boolean;
hreg,hreg2: TRegister;
hl: tasmlabel;
begin
ovloc.loc := LOC_VOID;
signed:=(size in [OS_S8,OS_S16,OS_S32]);
if (setflags and (not signed) and (src2=dst) and (op in [OP_ADD,OP_SUB])) then
hreg:=GetIntRegister(list,OS_INT)
else
hreg:=dst;
case op of
OP_ADD:
begin
list.concat(taicpu.op_reg_reg_reg(ops_add[setflags and signed], hreg, src2, src1));
if setflags and (not signed) then
overflowcheck_internal(list, hreg, src2);
a_load_reg_reg(list, OS_INT, OS_INT, hreg, dst);
end;
OP_SUB:
begin
list.concat(taicpu.op_reg_reg_reg(ops_sub[setflags and signed], hreg, src2, src1));
if setflags and (not signed) then
overflowcheck_internal(list, src2, hreg);
a_load_reg_reg(list, OS_INT, OS_INT, hreg, dst);
end;
OP_MUL,OP_IMUL:
begin
if (CPUMIPS_HAS_ISA32R2 in cpu_capabilities[current_settings.cputype]) and
(not setflags) then
{ NOTE: MUL is actually mips32r1 instruction; on older cores it is handled as macro }
list.concat(taicpu.op_reg_reg_reg(A_MUL,dst,src2,src1))
else
begin
list.concat(taicpu.op_reg_reg(TOpCg2AsmOp[op], src2, src1));
list.concat(taicpu.op_reg(A_MFLO, dst));
if setflags then
begin
current_asmdata.getjumplabel(hl);
hreg:=GetIntRegister(list,OS_INT);
list.concat(taicpu.op_reg(A_MFHI,hreg));
if (op=OP_IMUL) then
begin
hreg2:=GetIntRegister(list,OS_INT);
list.concat(taicpu.op_reg_reg_const(A_SRA,hreg2,dst,31));
a_cmp_reg_reg_label(list,OS_INT,OC_EQ,hreg2,hreg,hl);
end
else
a_cmp_reg_reg_label(list,OS_INT,OC_EQ,hreg,NR_R0,hl);
list.concat(taicpu.op_const(A_BREAK,6));
a_label(list,hl);
end;
end;
end;
OP_AND,OP_OR,OP_XOR:
begin
list.concat(taicpu.op_reg_reg_reg(TOpCG2AsmOp[op], dst, src2, src1));
end;
else
internalerror(2007012602);
end;
maybeadjustresult(list,op,size,dst);
end;
{*************** compare instructructions ****************}
procedure TCGMIPS.a_cmp_const_reg_label(list: tasmlist; size: tcgsize; cmp_op: topcmp; a: tcgint; reg: tregister; l: tasmlabel);
var
tmpreg: tregister;
begin
if a = 0 then
a_cmp_reg_reg_label(list,size,cmp_op,NR_R0,reg,l)
else
begin
tmpreg := GetIntRegister(list,OS_INT);
if (a>=simm16lo) and (a<=simm16hi) and
(cmp_op in [OC_LT,OC_B,OC_GTE,OC_AE]) then
begin
list.concat(taicpu.op_reg_reg_const(ops_slti[cmp_op in [OC_LT,OC_GTE]],tmpreg,reg,a));
if cmp_op in [OC_LT,OC_B] then
a_cmp_reg_reg_label(list,size,OC_NE,NR_R0,tmpreg,l)
else
a_cmp_reg_reg_label(list,size,OC_EQ,NR_R0,tmpreg,l);
end
else
begin
a_load_const_reg(list,OS_INT,a,tmpreg);
a_cmp_reg_reg_label(list, size, cmp_op, tmpreg, reg, l);
end;
end;
end;
const
TOpCmp2AsmCond_z : array[OC_GT..OC_LTE] of TAsmCond=(
C_GTZ,C_LTZ,C_GEZ,C_LEZ
);
TOpCmp2AsmCond_eqne: array[topcmp] of TAsmCond = (C_NONE,
{ eq gt lt gte lte ne }
C_NONE, C_NE, C_NE, C_EQ, C_EQ, C_NONE,
{ be b ae a }
C_EQ, C_NE, C_EQ, C_NE
);
procedure TCGMIPS.a_cmp_reg_reg_label(list: tasmlist; size: tcgsize; cmp_op: topcmp; reg1, reg2: tregister; l: tasmlabel);
var
ai : Taicpu;
op: TAsmOp;
hreg: TRegister;
begin
if not (cmp_op in [OC_EQ,OC_NE]) then
begin
if ((reg1=NR_R0) or (reg2=NR_R0)) and (cmp_op in [OC_GT,OC_LT,OC_GTE,OC_LTE]) then
begin
if (reg2=NR_R0) then
begin
ai:=taicpu.op_reg_sym(A_BC,reg1,l);
ai.setcondition(TOpCmp2AsmCond_z[swap_opcmp(cmp_op)]);
end
else
begin
ai:=taicpu.op_reg_sym(A_BC,reg2,l);
ai.setcondition(TOpCmp2AsmCond_z[cmp_op]);
end;
end
else
begin
hreg:=GetIntRegister(list,OS_INT);
op:=ops_slt[cmp_op in [OC_LT,OC_LTE,OC_GT,OC_GTE]];
if (cmp_op in [OC_LTE,OC_GT,OC_BE,OC_A]) then { swap operands }
list.concat(taicpu.op_reg_reg_reg(op,hreg,reg1,reg2))
else
list.concat(taicpu.op_reg_reg_reg(op,hreg,reg2,reg1));
if (TOpCmp2AsmCond_eqne[cmp_op]=C_NONE) then
InternalError(2013051501);
ai:=taicpu.op_reg_reg_sym(A_BC,hreg,NR_R0,l);
ai.SetCondition(TOpCmp2AsmCond_eqne[cmp_op]);
end;
end
else
begin
ai:=taicpu.op_reg_reg_sym(A_BC,reg2,reg1,l);
ai.SetCondition(TOpCmp2AsmCond[cmp_op]);
end;
list.concat(ai);
{ Delay slot }
list.Concat(TAiCpu.Op_none(A_NOP));
end;
procedure TCGMIPS.a_jmp_always(List: tasmlist; l: TAsmLabel);
var
ai : Taicpu;
begin
ai := taicpu.op_sym(A_BA, l);
list.concat(ai);
{ Delay slot }
list.Concat(TAiCpu.Op_none(A_NOP));
end;
procedure TCGMIPS.a_jmp_name(list: tasmlist; const s: string);
begin
List.Concat(TAiCpu.op_sym(A_BA, current_asmdata.RefAsmSymbol(s,AT_FUNCTION)));
{ Delay slot }
list.Concat(TAiCpu.Op_none(A_NOP));
end;
procedure TCGMIPS.a_jmp_flags(list: tasmlist; const f: TResFlags; l: tasmlabel);
var
ai: taicpu;
begin
case f.reg1 of
NR_FCC0..NR_FCC7:
begin
if (f.reg1=NR_FCC0) then
ai:=taicpu.op_sym(A_BC,l)
else
ai:=taicpu.op_reg_sym(A_BC,f.reg1,l);
list.concat(ai);
{ delay slot }
list.concat(taicpu.op_none(A_NOP));
case f.cond of
OC_NE: ai.SetCondition(C_COP1TRUE);
OC_EQ: ai.SetCondition(C_COP1FALSE);
else
InternalError(2014082901);
end;
exit;
end;
else
;
end;
if f.use_const then
a_cmp_const_reg_label(list,OS_INT,f.cond,f.value,f.reg1,l)
else
a_cmp_reg_reg_label(list,OS_INT,f.cond,f.reg2,f.reg1,l);
end;
procedure TCGMIPS.g_flags2reg(list: tasmlist; size: tcgsize; const f: tresflags; reg: tregister);
var
left,right: tregister;
unsigned: boolean;
hl: tasmlabel;
begin
case f.reg1 of
NR_FCC0..NR_FCC7:
begin
if (current_settings.cputype>=cpu_mips4) then
begin
a_load_const_reg(list,size,1,reg);
case f.cond of
OC_NE: list.concat(taicpu.op_reg_reg_reg(A_MOVF,reg,NR_R0,f.reg1));
OC_EQ: list.concat(taicpu.op_reg_reg_reg(A_MOVT,reg,NR_R0,f.reg1));
else
InternalError(2014082902);
end;
end
else
begin
{ TODO: still possible to do branchless by extracting appropriate bit from FCSR? }
current_asmdata.getjumplabel(hl);
a_load_const_reg(list,size,1,reg);
a_jmp_flags(list,f,hl);
a_load_const_reg(list,size,0,reg);
a_label(list,hl);
end;
exit;
end;
else
;
end;
if (f.cond in [OC_EQ,OC_NE]) then
begin
left:=reg;
if f.use_const and (f.value>=0) and (f.value<=65535) then
begin
if (f.value<>0) then
list.concat(taicpu.op_reg_reg_const(A_XORI,reg,f.reg1,f.value))
else
left:=f.reg1;
end
else
begin
if f.use_const then
begin
right:=GetIntRegister(list,OS_INT);
a_load_const_reg(list,OS_INT,f.value,right);
end
else
right:=f.reg2;
list.concat(taicpu.op_reg_reg_reg(A_XOR,reg,f.reg1,right));
end;
if f.cond=OC_EQ then
list.concat(taicpu.op_reg_reg_const(A_SLTIU,reg,left,1))
else
list.concat(taicpu.op_reg_reg_reg(A_SLTU,reg,NR_R0,left));
end
else
begin
{
sle x,a,b --> slt x,b,a; xori x,x,1 immediate not possible (or must be at left)
sgt x,a,b --> slt x,b,a likewise
sge x,a,b --> slt x,a,b; xori x,x,1
slt x,a,b --> unchanged
}
unsigned:=f.cond in [OC_GT,OC_LT,OC_GTE,OC_LTE];
if (f.cond in [OC_GTE,OC_LT,OC_B,OC_AE]) and
f.use_const and
(f.value>=simm16lo) and
(f.value<=simm16hi) then
list.Concat(taicpu.op_reg_reg_const(ops_slti[unsigned],reg,f.reg1,f.value))
else
begin
if f.use_const then
begin
if (f.value=0) then
right:=NR_R0
else
begin
right:=GetIntRegister(list,OS_INT);
a_load_const_reg(list,OS_INT,f.value,right);
end;
end
else
right:=f.reg2;
if (f.cond in [OC_LTE,OC_GT,OC_BE,OC_A]) then
list.Concat(taicpu.op_reg_reg_reg(ops_slt[unsigned],reg,right,f.reg1))
else
list.Concat(taicpu.op_reg_reg_reg(ops_slt[unsigned],reg,f.reg1,right));
end;
if (f.cond in [OC_LTE,OC_GTE,OC_BE,OC_AE]) then
list.Concat(taicpu.op_reg_reg_const(A_XORI,reg,reg,1));
end;
end;
procedure TCGMIPS.a_mul_reg_reg_pair(list: tasmlist; size: tcgsize; src1,src2,dstlo,dsthi: tregister);
var
asmop: tasmop;
begin
case size of
OS_32: asmop:=A_MULTU;
OS_S32: asmop:=A_MULT;
{$ifdef cpu64bitalu}
OS_64: asmop:=A_DMULTU;
OS_S64: asmop:=A_DMULT;
{$endif cpu64bitalu}
else
InternalError(2022020901);
end;
list.concat(taicpu.op_reg_reg(asmop,src1,src2));
if (dstlo<>NR_NO) then
list.concat(taicpu.op_reg(A_MFLO,dstlo));
if (dsthi<>NR_NO) then
list.concat(taicpu.op_reg(A_MFHI,dsthi));
end;
procedure TCGMIPS.g_overflowCheck(List: tasmlist; const Loc: TLocation; def: TDef);
begin
// this is an empty procedure
end;
procedure TCGMIPS.g_overflowCheck_loc(List: tasmlist; const Loc: TLocation; def: TDef; ovloc: tlocation);
begin
// this is an empty procedure
end;
{ *********** entry/exit code and address loading ************ }
procedure FixupOffsets(p:TObject;arg:pointer);
var
sym: tabstractnormalvarsym absolute p;
begin
if (tsym(p).typ=paravarsym) and
(sym.localloc.loc=LOC_REFERENCE) and
(sym.localloc.reference.base=NR_FRAME_POINTER_REG) then
begin
sym.localloc.reference.base:=NR_STACK_POINTER_REG;
Inc(sym.localloc.reference.offset,PLongint(arg)^);
end;
end;
procedure TCGMIPS.g_proc_entry(list: tasmlist; localsize: longint; nostackframe: boolean);
var
lastintoffset,lastfpuoffset,
nextoffset : aint;
i : longint;
ra_save,framesave : taicpu;
fmask,mask : dword;
saveregs : tcpuregisterset;
href: treference;
reg : Tsuperregister;
helplist : TAsmList;
largeoffs : boolean;
begin
list.concat(tai_directive.create(asd_ent,current_procinfo.procdef.mangledname));
if nostackframe then
begin
list.concat(taicpu.op_none(A_P_SET_NOMIPS16));
list.concat(taicpu.op_none(A_P_SET_NOREORDER));
exit;
end;
helplist:=TAsmList.Create;
reference_reset(href,0,[]);
href.base:=NR_STACK_POINTER_REG;
fmask:=0;
nextoffset:=tcpuprocinfo(current_procinfo).floatregstart;
lastfpuoffset:=LocalSize;
for reg := RS_F0 to RS_F31 do { to check: what if F30 is double? }
begin
if reg in (rg[R_FPUREGISTER].used_in_proc-paramanager.get_volatile_registers_fpu(pocall_stdcall)) then
begin
fmask:=fmask or (longword(1) shl ord(reg));
href.offset:=nextoffset;
lastfpuoffset:=nextoffset;
helplist.concat(taicpu.op_reg_ref(A_SWC1,newreg(R_FPUREGISTER,reg,R_SUBFS),href));
inc(nextoffset,4);
{ IEEE Double values are stored in floating point
register pairs f2X/f2X+1,
as the f2X+1 register is not correctly marked as used for now,
we simply assume it is also used if f2X is used
Should be fixed by a proper inclusion of f2X+1 into used_in_proc }
if (ord(reg)-ord(RS_F0)) mod 2 = 0 then
include(rg[R_FPUREGISTER].used_in_proc,succ(reg));
end;
end;
mask:=0;
nextoffset:=tcpuprocinfo(current_procinfo).intregstart;
saveregs:=rg[R_INTREGISTER].used_in_proc-paramanager.get_volatile_registers_int(pocall_stdcall);
if (current_procinfo.flags*[pi_do_call,pi_is_assembler]<>[]) then
include(saveregs,RS_R31);
if (pi_needs_stackframe in current_procinfo.flags) then
include(saveregs,RS_FRAME_POINTER_REG);
lastintoffset:=LocalSize;
framesave:=nil;
ra_save:=nil;
for reg:=RS_R1 to RS_R31 do
begin
if reg in saveregs then
begin
mask:=mask or (longword(1) shl ord(reg));
href.offset:=nextoffset;
lastintoffset:=nextoffset;
if (reg=RS_FRAME_POINTER_REG) then
framesave:=taicpu.op_reg_ref(A_SW,newreg(R_INTREGISTER,reg,R_SUBWHOLE),href)
else if (reg=RS_R31) then
ra_save:=taicpu.op_reg_ref(A_SW,newreg(R_INTREGISTER,reg,R_SUBWHOLE),href)
else
helplist.concat(taicpu.op_reg_ref(A_SW,newreg(R_INTREGISTER,reg,R_SUBWHOLE),href));
inc(nextoffset,4);
end;
end;
//list.concat(Taicpu.Op_reg_reg_const(A_ADDIU,NR_FRAME_POINTER_REG,NR_STACK_POINTER_REG,current_procinfo.para_stack_size));
list.concat(Taicpu.op_none(A_P_SET_NOMIPS16));
list.concat(Taicpu.op_reg_const_reg(A_P_FRAME,current_procinfo.framepointer,LocalSize,NR_R31));
list.concat(Taicpu.op_const_const(A_P_MASK,aint(mask),-(LocalSize-lastintoffset)));
list.concat(Taicpu.op_const_const(A_P_FMASK,aint(Fmask),-(LocalSize-lastfpuoffset)));
list.concat(Taicpu.op_none(A_P_SET_NOREORDER));
if tcpuprocinfo(current_procinfo).setnoat then
list.concat(Taicpu.op_none(A_P_SET_NOAT));
if (cs_create_pic in current_settings.moduleswitches) and
(pi_needs_got in current_procinfo.flags) then
begin
list.concat(Taicpu.op_reg(A_P_CPLOAD,NR_PIC_FUNC));
end;
if (-LocalSize >= simm16lo) and (-LocalSize <= simm16hi) then
begin
list.concat(Taicpu.op_none(A_P_SET_NOMACRO));
list.concat(Taicpu.Op_reg_reg_const(A_ADDIU,NR_STACK_POINTER_REG,NR_STACK_POINTER_REG,-LocalSize));
if assigned(ra_save) then
list.concat(ra_save);
if assigned(framesave) then
begin
list.concat(framesave);
list.concat(Taicpu.op_reg_reg_const(A_ADDIU,NR_FRAME_POINTER_REG,
NR_STACK_POINTER_REG,LocalSize));
end;
end
else
begin
a_load_const_reg(list,OS_32,-LocalSize,NR_R9);
list.concat(Taicpu.Op_reg_reg_reg(A_ADDU,NR_STACK_POINTER_REG,NR_STACK_POINTER_REG,NR_R9));
if assigned(ra_save) then
list.concat(ra_save);
if assigned(framesave) then
begin
list.concat(framesave);
list.concat(Taicpu.op_reg_reg_reg(A_SUBU,NR_FRAME_POINTER_REG,
NR_STACK_POINTER_REG,NR_R9));
end;
{ The instructions before are macros that can extend to multiple instructions,
the settings of R9 to -LocalSize surely does,
but the saving of RA and FP also might, and might
even use AT register, which is why we use R9 instead of AT here for -LocalSize }
list.concat(Taicpu.op_none(A_P_SET_NOMACRO));
end;
if (cs_create_pic in current_settings.moduleswitches) and
(pi_needs_got in current_procinfo.flags) then
begin
largeoffs:=(tcpuprocinfo(current_procinfo).save_gp_ref.offset>simm16hi);
if largeoffs then
list.concat(Taicpu.op_none(A_P_SET_MACRO));
list.concat(Taicpu.op_const(A_P_CPRESTORE,tcpuprocinfo(current_procinfo).save_gp_ref.offset));
if largeoffs then
list.concat(Taicpu.op_none(A_P_SET_NOMACRO));
end;
href.base:=NR_STACK_POINTER_REG;
for i:=0 to MIPS_MAX_REGISTERS_USED_IN_CALL-1 do
if tcpuprocinfo(current_procinfo).register_used[i] then
begin
reg:=parasupregs[i];
href.offset:=i*sizeof(aint)+LocalSize;
list.concat(taicpu.op_reg_ref(A_SW, newreg(R_INTREGISTER,reg,R_SUBWHOLE), href));
end;
list.concatList(helplist);
helplist.Free;
if current_procinfo.has_nestedprocs then
current_procinfo.procdef.parast.SymList.ForEachCall(@FixupOffsets,@LocalSize);
end;
procedure TCGMIPS.g_proc_exit(list: tasmlist; parasize: longint; nostackframe: boolean);
var
href : treference;
stacksize : aint;
saveregs : tcpuregisterset;
nextoffset : aint;
reg : Tsuperregister;
begin
stacksize:=current_procinfo.calc_stackframe_size;
if nostackframe then
begin
list.concat(taicpu.op_reg(A_JR, NR_R31));
list.concat(Taicpu.op_none(A_NOP));
list.concat(Taicpu.op_none(A_P_SET_MACRO));
list.concat(Taicpu.op_none(A_P_SET_REORDER));
end
else
begin
if tcpuprocinfo(current_procinfo).save_gp_ref.offset<>0 then
tg.ungettemp(list,tcpuprocinfo(current_procinfo).save_gp_ref);
reference_reset(href,0,[]);
href.base:=NR_STACK_POINTER_REG;
nextoffset:=tcpuprocinfo(current_procinfo).floatregstart;
for reg := RS_F0 to RS_F31 do
begin
if reg in (rg[R_FPUREGISTER].used_in_proc-paramanager.get_volatile_registers_fpu(pocall_stdcall)) then
begin
href.offset:=nextoffset;
list.concat(taicpu.op_reg_ref(A_LWC1,newreg(R_FPUREGISTER,reg,R_SUBFS),href));
inc(nextoffset,4);
end;
end;
nextoffset:=tcpuprocinfo(current_procinfo).intregstart;
saveregs:=rg[R_INTREGISTER].used_in_proc-paramanager.get_volatile_registers_int(pocall_stdcall);
if (current_procinfo.flags*[pi_do_call,pi_is_assembler]<>[]) then
include(saveregs,RS_R31);
if (pi_needs_stackframe in current_procinfo.flags) then
include(saveregs,RS_FRAME_POINTER_REG);
// GP does not need to be restored on exit
for reg:=RS_R1 to RS_R31 do
begin
if reg in saveregs then
begin
href.offset:=nextoffset;
list.concat(taicpu.op_reg_ref(A_LW,newreg(R_INTREGISTER,reg,R_SUBWHOLE),href));
inc(nextoffset,sizeof(aint));
end;
end;
if (-stacksize >= simm16lo) and (-stacksize <= simm16hi) then
begin
list.concat(taicpu.op_reg(A_JR, NR_R31));
{ correct stack pointer in the delay slot }
list.concat(Taicpu.Op_reg_reg_const(A_ADDIU, NR_STACK_POINTER_REG, NR_STACK_POINTER_REG, stacksize));
end
else
begin
a_load_const_reg(list,OS_32,stacksize,NR_R1);
list.concat(taicpu.op_reg(A_JR, NR_R31));
{ correct stack pointer in the delay slot }
list.concat(taicpu.op_reg_reg_reg(A_ADD,NR_STACK_POINTER_REG,NR_STACK_POINTER_REG,NR_R1));
tcpuprocinfo(current_procinfo).setnoat:=true;
end;
list.concat(Taicpu.op_none(A_P_SET_MACRO));
list.concat(Taicpu.op_none(A_P_SET_REORDER));
end;
list.concat(tai_directive.create(asd_ent_end,current_procinfo.procdef.mangledname));
end;
procedure TCGMIPS.g_concatcopy(list: tasmlist; const Source, dest: treference; len: tcgint);
var
tmpreg1, hreg, countreg: TRegister;
src, dst: TReference;
lab: tasmlabel;
Count, count2: aint;
function reference_is_reusable(const ref: treference): boolean;
begin
result:=(ref.base<>NR_NO) and (ref.index=NR_NO) and
(ref.symbol=nil) and
(ref.offset>=simm16lo) and (ref.offset+len<=simm16hi);
end;
begin
if len > high(longint) then
internalerror(2002072704);
{ A call (to FPC_MOVE) requires the outgoing parameter area to be properly
allocated on stack. This can only be done before tcpuprocinfo.set_first_temp_offset,
i.e. before secondpass. Other internal procedures request correct stack frame
by setting pi_do_call during firstpass, but for this particular one it is impossible.
Therefore, if the current procedure is a leaf one, we have to leave it that way. }
{ anybody wants to determine a good value here :)? }
if (len > 100) and
assigned(current_procinfo) and
(pi_do_call in current_procinfo.flags) then
g_concatcopy_move(list, Source, dest, len)
else if ((source.alignment<>0) and (source.alignment<4)) or
((dest.alignment<>0) and (dest.alignment<4)) then
g_concatcopy_unaligned(list, Source, dest, len)
else
begin
Count := len div 4;
if (count<=4) and reference_is_reusable(source) then
src:=source
else
begin
reference_reset(src,sizeof(aint),source.volatility);
{ load the address of source into src.base }
src.base := GetAddressRegister(list);
a_loadaddr_ref_reg(list, Source, src.base);
end;
if (count<=4) and reference_is_reusable(dest) then
dst:=dest
else
begin
reference_reset(dst,sizeof(aint),dest.volatility);
{ load the address of dest into dst.base }
dst.base := GetAddressRegister(list);
a_loadaddr_ref_reg(list, dest, dst.base);
end;
{ generate a loop }
if Count > 4 then
begin
countreg := GetIntRegister(list, OS_INT);
tmpreg1 := GetIntRegister(list, OS_INT);
a_load_const_reg(list, OS_INT, Count, countreg);
current_asmdata.getjumplabel(lab);
a_label(list, lab);
list.concat(taicpu.op_reg_ref(A_LW, tmpreg1, src));
list.concat(taicpu.op_reg_ref(A_SW, tmpreg1, dst));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, src.base, src.base, 4));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, dst.base, dst.base, 4));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, countreg, countreg, -1));
a_cmp_reg_reg_label(list,OS_INT,OC_GT,NR_R0,countreg,lab);
len := len mod 4;
end;
{ unrolled loop }
Count := len div 4;
if Count > 0 then
begin
tmpreg1 := GetIntRegister(list, OS_INT);
count2:=1;
while count2 <= Count do
begin
list.concat(taicpu.op_reg_ref(A_LW, tmpreg1, src));
list.concat(taicpu.op_reg_ref(A_SW, tmpreg1, dst));
Inc(src.offset, 4);
Inc(dst.offset, 4);
Inc(count2);
end;
len := len mod 4;
end;
if (len and 4) <> 0 then
begin
hreg := GetIntRegister(list, OS_INT);
a_load_ref_reg(list, OS_32, OS_32, src, hreg);
a_load_reg_ref(list, OS_32, OS_32, hreg, dst);
Inc(src.offset, 4);
Inc(dst.offset, 4);
end;
{ copy the leftovers }
if (len and 2) <> 0 then
begin
hreg := GetIntRegister(list, OS_INT);
a_load_ref_reg(list, OS_16, OS_16, src, hreg);
a_load_reg_ref(list, OS_16, OS_16, hreg, dst);
Inc(src.offset, 2);
Inc(dst.offset, 2);
end;
if (len and 1) <> 0 then
begin
hreg := GetIntRegister(list, OS_INT);
a_load_ref_reg(list, OS_8, OS_8, src, hreg);
a_load_reg_ref(list, OS_8, OS_8, hreg, dst);
end;
end;
end;
procedure TCGMIPS.g_concatcopy_unaligned(list: tasmlist; const Source, dest: treference; len: tcgint);
var
src, dst: TReference;
tmpreg1, countreg: TRegister;
i: aint;
lab: tasmlabel;
begin
if (len > 31) and
{ see comment in g_concatcopy }
assigned(current_procinfo) and
(pi_do_call in current_procinfo.flags) then
g_concatcopy_move(list, Source, dest, len)
else
begin
reference_reset(src,sizeof(aint),source.volatility);
reference_reset(dst,sizeof(aint),dest.volatility);
{ load the address of source into src.base }
src.base := GetAddressRegister(list);
a_loadaddr_ref_reg(list, Source, src.base);
{ load the address of dest into dst.base }
dst.base := GetAddressRegister(list);
a_loadaddr_ref_reg(list, dest, dst.base);
{ generate a loop }
if len > 4 then
begin
countreg := GetIntRegister(list, OS_INT);
tmpreg1 := GetIntRegister(list, OS_INT);
a_load_const_reg(list, OS_INT, len, countreg);
current_asmdata.getjumplabel(lab);
a_label(list, lab);
list.concat(taicpu.op_reg_ref(A_LBU, tmpreg1, src));
list.concat(taicpu.op_reg_ref(A_SB, tmpreg1, dst));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, src.base, src.base, 1));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, dst.base, dst.base, 1));
list.concat(taicpu.op_reg_reg_const(A_ADDIU, countreg, countreg, -1));
a_cmp_reg_reg_label(list,OS_INT,OC_GT,NR_R0,countreg,lab);
end
else
begin
{ unrolled loop }
tmpreg1 := GetIntRegister(list, OS_INT);
i := 1;
while i <= len do
begin
list.concat(taicpu.op_reg_ref(A_LBU, tmpreg1, src));
list.concat(taicpu.op_reg_ref(A_SB, tmpreg1, dst));
Inc(src.offset);
Inc(dst.offset);
Inc(i);
end;
end;
end;
end;
procedure TCGMIPS.g_profilecode(list:TAsmList);
var
href: treference;
begin
if not (cs_create_pic in current_settings.moduleswitches) then
begin
reference_reset_symbol(href,current_asmdata.RefAsmSymbol('_gp',AT_DATA),0,sizeof(pint),[]);
a_loadaddr_ref_reg(list,href,NR_GP);
end;
list.concat(taicpu.op_reg_reg(A_MOVE,NR_R1,NR_RA));
list.concat(taicpu.op_reg_reg_const(A_ADDIU,NR_SP,NR_SP,-8));
a_call_sym_pic(list,current_asmdata.RefAsmSymbol('_mcount',AT_FUNCTION));
tcpuprocinfo(current_procinfo).setnoat:=true;
end;
procedure TCGMIPS.g_adjust_self_value(list:TAsmList;procdef: tprocdef;ioffset: tcgint);
begin
{ This method is integrated into g_intf_wrapper and shouldn't be called separately }
InternalError(2013020102);
end;
{$ifndef mips64}
{****************************************************************************
TCG64_MIPSel
****************************************************************************}
procedure TCg64MPSel.a_load64_reg_ref(list: tasmlist; reg: tregister64; const ref: treference);
var
tmpref: treference;
tmpreg: tregister;
incr: shortint;
begin
tmpref := ref;
tcgmips(cg).make_simple_ref(list,tmpref);
if (ref.alignment <4) then
begin
if target_info.endian = endian_big then
begin
inc(tmpref.offset,7);
incr:=-1;
end
else
incr:=1;
list.concat(taicpu.op_reg_ref(A_SB,reg.reglo,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reglo,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reglo,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reglo,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_ref(A_SB,reg.reghi,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reghi,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reghi,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SRL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_SB,reg.reghi,tmpref));
end
else
begin
if target_info.endian = endian_big then
begin
tmpreg := reg.reglo;
reg.reglo := reg.reghi;
reg.reghi := tmpreg;
end;
list.concat(taicpu.op_reg_ref(A_SW,reg.reglo,tmpref));
Inc(tmpref.offset, 4);
list.concat(taicpu.op_reg_ref(A_SW,reg.reghi,tmpref));
end;
end;
procedure TCg64MPSel.a_load64_ref_reg(list: tasmlist; const ref: treference; reg: tregister64);
var
tmpref: treference;
tmpreg: tregister;
incr: shortint;
begin
tmpref := ref;
tcgmips(cg).make_simple_ref(list,tmpref);
if (ref.alignment <4) then
begin
tmpreg:=cg.getintregister(list,OS_INT);
if target_info.endian = endian_little then
begin
inc(tmpref.offset,7);
incr:=-1;
end
else
incr:=1;
list.concat(taicpu.op_reg_ref(A_LBU,reg.reghi,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reghi,reg.reghi,tmpreg));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reghi,reg.reghi,tmpreg));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reghi,reg.reghi,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reghi,reg.reghi,tmpreg));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_ref(A_LBU,reg.reglo,tmpref));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reglo,reg.reglo,tmpreg));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reglo,reg.reglo,tmpreg));
Inc(tmpref.offset,incr);
list.concat(taicpu.op_reg_reg_const(A_SLL,reg.reglo,reg.reglo,8));
list.concat(taicpu.op_reg_ref(A_LBU,tmpreg,tmpref));
list.concat(taicpu.op_reg_reg_reg(A_ADDU,reg.reglo,reg.reglo,tmpreg));
Inc(tmpref.offset,incr);
end
else
begin
if target_info.endian = endian_big then
begin
tmpreg := reg.reglo;
reg.reglo := reg.reghi;
reg.reghi := tmpreg;
end;
list.concat(taicpu.op_reg_ref(A_LW,reg.reglo,tmpref));
Inc(tmpref.offset, 4);
list.concat(taicpu.op_reg_ref(A_LW,reg.reghi,tmpref));
end;
end;
procedure TCg64MPSel.a_load64_ref_cgpara(list: tasmlist; const r: treference; const paraloc: tcgpara);
var
hreg64: tregister64;
begin
{ Override this function to prevent loading the reference twice.
Use here some extra registers, but those are optimized away by the RA }
hreg64.reglo := cg.GetIntRegister(list, OS_S32);
hreg64.reghi := cg.GetIntRegister(list, OS_S32);
a_load64_ref_reg(list, r, hreg64);
a_load64_reg_cgpara(list, hreg64, paraloc);
end;
procedure TCg64MPSel.a_op64_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; regsrc, regdst: TRegister64);
var
tmpreg1: TRegister;
begin
case op of
OP_NEG:
begin
tmpreg1 := cg.GetIntRegister(list, OS_INT);
list.concat(taicpu.op_reg_reg_reg(A_SUBU, regdst.reglo, NR_R0, regsrc.reglo));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, tmpreg1, NR_R0, regdst.reglo));
list.concat(taicpu.op_reg_reg_reg(A_SUBU, regdst.reghi, NR_R0, regsrc.reghi));
list.concat(taicpu.op_reg_reg_reg(A_SUBU, regdst.reghi, regdst.reghi, tmpreg1));
end;
OP_NOT:
begin
list.concat(taicpu.op_reg_reg_reg(A_NOR, regdst.reglo, NR_R0, regsrc.reglo));
list.concat(taicpu.op_reg_reg_reg(A_NOR, regdst.reghi, NR_R0, regsrc.reghi));
end;
else
a_op64_reg_reg_reg(list,op,size,regsrc,regdst,regdst);
end;
end;
procedure TCg64MPSel.a_op64_const_reg(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regdst: TRegister64);
begin
a_op64_const_reg_reg(list, op, size, value, regdst, regdst);
end;
procedure TCg64MPSel.a_op64_const_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regsrc, regdst: tregister64);
var
l: tlocation;
begin
a_op64_const_reg_reg_checkoverflow(list, op, size, Value, regsrc, regdst, False, l);
end;
procedure TCg64MPSel.a_op64_reg_reg_reg(list: tasmlist; op: TOpCG; size: tcgsize; regsrc1, regsrc2, regdst: tregister64);
var
l: tlocation;
begin
a_op64_reg_reg_reg_checkoverflow(list, op, size, regsrc1, regsrc2, regdst, False, l);
end;
procedure TCg64MPSel.a_op64_const_reg_reg_checkoverflow(list: tasmlist; op: TOpCG; size: tcgsize; Value: int64; regsrc, regdst: tregister64; setflags: boolean; var ovloc: tlocation);
var
tmplo,carry: TRegister;
hisize: tcgsize;
begin
carry:=NR_NO;
if (size in [OS_S64]) then
hisize:=OS_S32
else
hisize:=OS_32;
case op of
OP_AND,OP_OR,OP_XOR:
begin
cg.a_op_const_reg_reg(list,op,OS_32,aint(lo(value)),regsrc.reglo,regdst.reglo);
cg.a_op_const_reg_reg(list,op,OS_32,aint(hi(value)),regsrc.reghi,regdst.reghi);
end;
OP_ADD:
begin
if lo(value)<>0 then
begin
tmplo:=cg.GetIntRegister(list,OS_32);
carry:=cg.GetIntRegister(list,OS_32);
tcgmips(cg).handle_reg_const_reg(list,A_ADDU,regsrc.reglo,aint(lo(value)),tmplo);
list.concat(taicpu.op_reg_reg_reg(A_SLTU,carry,tmplo,regsrc.reglo));
cg.a_load_reg_reg(list,OS_32,OS_32,tmplo,regdst.reglo);
end
else
cg.a_load_reg_reg(list,OS_32,OS_32,regsrc.reglo,regdst.reglo);
{ With overflow checking and unsigned args, this generates slighly suboptimal code
($80000000 constant loaded twice). Other cases are fine. Getting it perfect does not
look worth the effort. }
cg.a_op_const_reg_reg_checkoverflow(list,OP_ADD,hisize,aint(hi(value)),regsrc.reghi,regdst.reghi,setflags,ovloc);
if carry<>NR_NO then
cg.a_op_reg_reg_reg_checkoverflow(list,OP_ADD,hisize,carry,regdst.reghi,regdst.reghi,setflags,ovloc);
end;
OP_SUB:
begin
carry:=NR_NO;
if lo(value)<>0 then
begin
tmplo:=cg.GetIntRegister(list,OS_32);
carry:=cg.GetIntRegister(list,OS_32);
tcgmips(cg).handle_reg_const_reg(list,A_SUBU,regsrc.reglo,aint(lo(value)),tmplo);
list.concat(taicpu.op_reg_reg_reg(A_SLTU,carry,regsrc.reglo,tmplo));
cg.a_load_reg_reg(list,OS_32,OS_32,tmplo,regdst.reglo);
end
else
cg.a_load_reg_reg(list,OS_32,OS_32,regsrc.reglo,regdst.reglo);
cg.a_op_const_reg_reg_checkoverflow(list,OP_SUB,hisize,aint(hi(value)),regsrc.reghi,regdst.reghi,setflags,ovloc);
if carry<>NR_NO then
cg.a_op_reg_reg_reg_checkoverflow(list,OP_SUB,hisize,carry,regdst.reghi,regdst.reghi,setflags,ovloc);
end;
else
InternalError(2013050301);
end;
end;
procedure TCg64MPSel.a_op64_reg_reg_reg_checkoverflow(list: tasmlist; op: TOpCG; size: tcgsize; regsrc1, regsrc2, regdst: tregister64; setflags: boolean; var ovloc: tlocation);
var
tmplo,tmphi,carry,hreg: TRegister;
signed: boolean;
begin
case op of
OP_ADD:
begin
signed:=(size in [OS_S64]);
tmplo := cg.GetIntRegister(list,OS_S32);
carry := cg.GetIntRegister(list,OS_S32);
// destreg.reglo could be regsrc1.reglo or regsrc2.reglo
list.concat(taicpu.op_reg_reg_reg(A_ADDU, tmplo, regsrc2.reglo, regsrc1.reglo));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, tmplo, regsrc2.reglo));
cg.a_load_reg_reg(list,OS_INT,OS_INT,tmplo,regdst.reglo);
if signed or (not setflags) then
begin
list.concat(taicpu.op_reg_reg_reg(ops_add[setflags and signed], regdst.reghi, regsrc2.reghi, regsrc1.reghi));
list.concat(taicpu.op_reg_reg_reg(ops_add[setflags and signed], regdst.reghi, regdst.reghi, carry));
end
else
begin
tmphi:=cg.GetIntRegister(list,OS_INT);
hreg:=cg.GetIntRegister(list,OS_INT);
cg.a_load_const_reg(list,OS_INT,$80000000,hreg);
// first add carry to one of the addends
list.concat(taicpu.op_reg_reg_reg(A_ADDU, tmphi, regsrc2.reghi, carry));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, tmphi, regsrc2.reghi));
list.concat(taicpu.op_reg_reg_reg(A_SUB, carry, hreg, carry));
// then add another addend
list.concat(taicpu.op_reg_reg_reg(A_ADDU, regdst.reghi, tmphi, regsrc1.reghi));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, regdst.reghi, tmphi));
list.concat(taicpu.op_reg_reg_reg(A_SUB, carry, hreg, carry));
end;
end;
OP_SUB:
begin
signed:=(size in [OS_S64]);
tmplo := cg.GetIntRegister(list,OS_S32);
carry := cg.GetIntRegister(list,OS_S32);
// destreg.reglo could be regsrc1.reglo or regsrc2.reglo
list.concat(taicpu.op_reg_reg_reg(A_SUBU, tmplo, regsrc2.reglo, regsrc1.reglo));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, regsrc2.reglo,tmplo));
cg.a_load_reg_reg(list,OS_INT,OS_INT,tmplo,regdst.reglo);
if signed or (not setflags) then
begin
list.concat(taicpu.op_reg_reg_reg(ops_sub[setflags and signed], regdst.reghi, regsrc2.reghi, regsrc1.reghi));
list.concat(taicpu.op_reg_reg_reg(ops_sub[setflags and signed], regdst.reghi, regdst.reghi, carry));
end
else
begin
tmphi:=cg.GetIntRegister(list,OS_INT);
hreg:=cg.GetIntRegister(list,OS_INT);
cg.a_load_const_reg(list,OS_INT,$80000000,hreg);
// first subtract the carry...
list.concat(taicpu.op_reg_reg_reg(A_SUBU, tmphi, regsrc2.reghi, carry));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, regsrc2.reghi, tmphi));
list.concat(taicpu.op_reg_reg_reg(A_SUB, carry, hreg, carry));
// ...then the subtrahend
list.concat(taicpu.op_reg_reg_reg(A_SUBU, regdst.reghi, tmphi, regsrc1.reghi));
list.concat(taicpu.op_reg_reg_reg(A_SLTU, carry, tmphi, regdst.reghi));
list.concat(taicpu.op_reg_reg_reg(A_SUB, carry, hreg, carry));
end;
end;
OP_AND,OP_OR,OP_XOR:
begin
cg.a_op_reg_reg_reg(list,op,size,regsrc1.reglo,regsrc2.reglo,regdst.reglo);
cg.a_op_reg_reg_reg(list,op,size,regsrc1.reghi,regsrc2.reghi,regdst.reghi);
end;
else
internalerror(200306017);
end;
end;
{$endif mips64}
procedure create_codegen;
begin
cg:=TCGMIPS.Create;
{$ifdef mips64}
cg128:=tcg128.create;
{$else mips64}
cg64:=TCg64MPSel.Create;
{$endif mips64}
end;
end.