1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
|
{
Copyright (c) 1998-2014 by the Free Pascal development team
This unit calls the optimization procedures to optimize the assembler
code for m68k
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 aoptcpu;
{$i fpcdefs.inc}
{$define DEBUG_AOPTCPU}
Interface
uses
cpubase, aoptobj, aoptcpub, aopt, aasmtai,
cgbase;
Type
TCpuAsmOptimizer = class(TAsmOptimizer)
function RegLoadedWithNewValue(reg: tregister; hp: tai): boolean; override;
function PeepHoleOptPass1Cpu(var p: tai): boolean; override;
function TryToOptimizeMove(var p: tai): boolean;
{ outputs a debug message into the assembler file }
procedure DebugMsg(const s: string; p: tai);
End;
Implementation
uses
cutils, aasmcpu, cgutils, globals, verbose, cpuinfo, itcpugas;
{ Range check must be disabled explicitly as conversions between signed and unsigned
32-bit values are done without explicit typecasts }
{$R-}
function opname(var p: tai): string;
begin
result:=upcase(gas_op2str[taicpu(p).opcode]);
end;
function RefsEqual(const r1, r2: treference): boolean;
begin
RefsEqual :=
(r1.offset = r2.offset) and
(r1.base = r2.base) and
(r1.index = r2.index) and (r1.scalefactor = r2.scalefactor) and
(r1.symbol=r2.symbol) and (r1.refaddr = r2.refaddr) and
(r1.relsymbol = r2.relsymbol);
end;
function MatchOperand(const oper1: TOper; const oper2: TOper): boolean;
begin
result := oper1.typ = oper2.typ;
if result then
case oper1.typ of
top_const:
Result:=oper1.val = oper2.val;
top_reg:
Result:=oper1.reg = oper2.reg;
top_ref:
Result:=RefsEqual(oper1.ref^, oper2.ref^);
else
internalerror(2016112401);
end
end;
function TCpuAsmOptimizer.RegLoadedWithNewValue(reg: tregister; hp: tai): boolean;
var
p: taicpu;
begin
if not assigned(hp) or
(hp.typ <> ait_instruction) then
begin
Result := false;
exit;
end;
p := taicpu(hp);
Result :=
((p.opcode in [A_MOVE,A_MOVEA,A_MVS,A_MVZ,A_MOVEQ,A_LEA]) and
(p.oper[1]^.typ = top_reg) and
(SuperRegistersEqual(p.oper[1]^.reg,reg)) and
((p.oper[0]^.typ = top_const) or
((p.oper[0]^.typ = top_reg) and
not(SuperRegistersEqual(p.oper[0]^.reg,reg))) or
((p.oper[0]^.typ = top_ref) and
not RegInRef(reg,p.oper[0]^.ref^)))) or
((p.opcode = A_FMOVE) and
(p.oper[1]^.typ = top_reg) and
(SuperRegistersEqual(p.oper[1]^.reg,reg)) and
((p.oper[0]^.typ = top_realconst) or
((p.oper[0]^.typ = top_reg) and
not(SuperRegistersEqual(p.oper[0]^.reg,reg))))) or
((p.opcode = A_MOVEM) and
(p.oper[1]^.typ = top_regset) and
((getsupreg(reg) in p.oper[1]^.dataregset) or
(getsupreg(reg) in p.oper[1]^.addrregset))) or
((p.opcode = A_FMOVEM) and
(p.oper[1]^.typ = top_regset) and
(getsupreg(reg) in p.oper[1]^.fpuregset));
end;
{$ifdef DEBUG_AOPTCPU}
procedure TCpuAsmOptimizer.DebugMsg(const s: string; p : tai);
begin
asml.insertbefore(tai_comment.Create(strpnew(s)), p);
end;
{$else DEBUG_AOPTCPU}
procedure TCpuAsmOptimizer.DebugMsg(const s: string; p : tai);inline;
begin
end;
{$endif DEBUG_AOPTCPU}
function TCpuAsmOptimizer.TryToOptimizeMove(var p: tai): boolean;
var
next, next2: tai;
opstr: string[15];
begin
result:=false;
if GetNextInstruction(p,next) and
(next.typ = ait_instruction) and
(taicpu(next).opcode = taicpu(p).opcode) and
(taicpu(p).opsize = taicpu(next).opsize) and
(taicpu(p).oper[1]^.typ = top_reg) and
MatchOperand(taicpu(p).oper[1]^,taicpu(next).oper[0]^) then
begin
if not(RegInOp(taicpu(p).oper[1]^.reg,taicpu(next).oper[1]^)) and
RegEndOfLife(taicpu(next).oper[0]^.reg, taicpu(next)) then
begin
opstr:=opname(p);
case taicpu(p).oper[0]^.typ of
top_reg:
begin
{ move %reg0, %tmpreg; move %tmpreg, <ea> -> move %reg0, <ea> }
taicpu(p).loadOper(1,taicpu(next).oper[1]^);
asml.remove(next);
next.free;
result:=true;
{ also remove leftover move %reg0, %reg0, which can occur as the result
of the previous optimization, if %reg0 and %tmpreg was different types
(addr vs. data), so these moves were left in by the cg }
if MatchOperand(taicpu(p).oper[0]^,taicpu(p).oper[1]^) then
begin
DebugMsg('Optimizer: '+opstr+' + '+opstr+' removed',p);
GetNextInstruction(p,next);
asml.remove(p);
p.free;
p:=next;
end
else
DebugMsg('Optimizer: '+opstr+' + '+opstr+' to '+opstr+' #1',p)
end;
top_const:
begin
// DebugMsg('Optimizer: '+opstr+' + '+opstr+' to '+opstr+' #2',p);
end;
top_ref:
begin
{ move ref, %tmpreg; move %tmpreg, <ea> -> move ref, <ea> }
{ we only want to do this when <ea> is a reg or a simple reference }
with taicpu(next).oper[1]^ do
if (taicpu(next).opcode <> A_FMOVE) and
((typ = top_reg) or
((typ = top_ref) and
((ref^.index = NR_NO) or
(ref^.base = NR_NO)) and
(ref^.symbol = nil) and
(ref^.offset = 0))) then
begin
DebugMsg('Optimizer: '+opstr+' + '+opstr+' to '+opstr+' #3',p);
taicpu(p).loadOper(1,taicpu(next).oper[1]^);
asml.remove(next);
next.free;
result:=true;
end;
end;
end;
end;
exit;
end;
if GetNextInstruction(p,next) and
(next.typ = ait_instruction) and
GetNextInstruction(next,next2) and
(next2.typ = ait_instruction) and
(taicpu(next).opcode <> taicpu(p).opcode) and
(taicpu(next2).opcode = taicpu(p).opcode) and
(taicpu(p).oper[0]^.typ = top_reg) and
(taicpu(p).oper[1]^.typ = top_reg) and
(getregtype(taicpu(p).oper[0]^.reg) = getregtype(taicpu(p).oper[1]^.reg)) and
MatchOperand(taicpu(p).oper[1]^,taicpu(next2).oper[0]^) and
MatchOperand(taicpu(next2).oper[1]^,taicpu(p).oper[0]^) and
(taicpu(p).opsize = taicpu(next2).opsize) and
((taicpu(p).opcode = A_FMOVE) or
(taicpu(p).opsize = taicpu(next).opsize)) then
begin
opstr:=opname(p);
if not(RegInOp(taicpu(p).oper[1]^.reg,taicpu(next2).oper[1]^)) and
not(RegInOp(taicpu(p).oper[1]^.reg,taicpu(next).oper[0]^)) and
RegEndOfLife(taicpu(p).oper[0]^.reg, taicpu(next2)) then
begin
{ move %reg0, %tmpreg
op ???, %tmpreg
move %tmpreg, %reg0
to:
op ???, %reg0 }
if MatchOperand(taicpu(p).oper[1]^,taicpu(next).oper[taicpu(next).ops-1]^) then
begin
{
Disabled, because it breaks some tests... :( (KB)
DebugMsg('Optimizer: '+opstr+' + OP + '+opstr+' to OP #1',next);
taicpu(next).loadOper(taicpu(next).ops-1,taicpu(p).oper[0]^);
asml.remove(p);
asml.remove(next2);
p.free;
next2.free;
result:=true;
}
end;
end;
end;
end;
function TCpuAsmOptimizer.PeepHoleOptPass1Cpu(var p: tai): boolean;
var
next: tai;
tmpref: treference;
tmpsingle: single;
begin
result:=false;
case p.typ of
ait_instruction:
begin
//asml.insertbefore(tai_comment.Create(strpnew('pass1 called for instr')), p);
case taicpu(p).opcode of
A_MOVE:
result:=TryToOptimizeMove(p);
{ LEA (Ax),Ax is a NOP if src and dest reg is equal, so remove it. }
A_LEA:
if not assigned(taicpu(p).oper[0]^.ref^.symbol) and
(((taicpu(p).oper[0]^.ref^.base = taicpu(p).oper[1]^.reg) and
(taicpu(p).oper[0]^.ref^.index = NR_NO)) or
((taicpu(p).oper[0]^.ref^.index = taicpu(p).oper[1]^.reg) and
(taicpu(p).oper[0]^.ref^.base = NR_NO))) and
(taicpu(p).oper[0]^.ref^.offset = 0) then
begin
DebugMsg('Optimizer: LEA 0(Ax),Ax removed',p);
GetNextInstruction(p,next);
asml.remove(p);
p.free;
p:=next;
result:=true;
end;
{ Address register sub/add can be replaced with ADDQ/SUBQ or LEA if the value is in the
SmallInt range, which is shorter to encode and faster to execute on most 68k }
A_SUB,A_SUBA,A_ADD,A_ADDA:
if (taicpu(p).oper[1]^.typ = top_reg) and isaddressregister(taicpu(p).oper[1]^.reg) and
(taicpu(p).oper[0]^.typ = top_const) then
begin
if isvalueforaddqsubq(taicpu(p).oper[0]^.val) then
begin
DebugMsg('Optimizer: SUB/ADD #val,Ax to SUBQ/ADDQ',p);
taicpu(p).opsize:=S_L; // this is safe, because we're targetting an address reg
if taicpu(p).opcode in [A_ADD,A_ADDA] then
taicpu(p).opcode:=A_ADDQ
else
taicpu(p).opcode:=A_SUBQ;
result:=true;
end
else
if isvalue16bit(abs(taicpu(p).oper[0]^.val)) then
begin
DebugMsg('Optimizer: SUB/ADD #val,Ax to LEA val(Ax),Ax',p);
if taicpu(p).opcode in [A_SUB,A_SUBA] then
reference_reset_base(tmpref,taicpu(p).oper[1]^.reg,-taicpu(p).oper[0]^.val,ctempposinvalid,0,[])
else
reference_reset_base(tmpref,taicpu(p).oper[1]^.reg,taicpu(p).oper[0]^.val,ctempposinvalid,0,[]);
taicpu(p).opcode:=A_LEA;
taicpu(p).loadref(0,tmpref);
result:=true;
end;
end;
{ MOVEA #0,Ax to SUBA Ax,Ax, because it's shorter }
A_MOVEA:
if (taicpu(p).oper[0]^.typ = top_const) and
(taicpu(p).oper[0]^.val = 0) then
begin
DebugMsg('Optimizer: MOVEA #0,Ax to SUBA Ax,Ax',p);
taicpu(p).opcode:=A_SUBA;
taicpu(p).opsize:=S_L; { otherwise it will be .W -> BOOM }
taicpu(p).loadoper(0,taicpu(p).oper[1]^);
result:=true;
end;
{ CLR.L Dx on a 68000 is slower than MOVEQ #0,Dx }
A_CLR:
if (current_settings.cputype in [cpu_mc68000]) and
(taicpu(p).oper[0]^.typ = top_reg) and
(taicpu(p).opsize = S_L) and
isintregister(taicpu(p).oper[0]^.reg) then
begin
//DebugMsg('Optimizer: CLR.L Dx to MOVEQ #0,Dx',p);
taicpu(p).opcode:=A_MOVEQ;
taicpu(p).loadoper(1,taicpu(p).oper[0]^);
taicpu(p).loadconst(0,0);
taicpu(p).ops:=2;
result:=true;
end;
{ CMP #0,<ea> equals to TST <ea>, just shorter and TST is more flexible anyway }
A_CMP,A_CMPI:
if (taicpu(p).oper[0]^.typ = top_const) and
(taicpu(p).oper[0]^.val = 0) then
begin
DebugMsg('Optimizer: CMP #0 to TST',p);
taicpu(p).opcode:=A_TST;
taicpu(p).loadoper(0,taicpu(p).oper[1]^);
taicpu(p).clearop(1);
taicpu(p).ops:=1;
result:=true;
end;
A_FCMP:
if (taicpu(p).oper[0]^.typ = top_realconst) then
begin
if (taicpu(p).oper[0]^.val_real = 0.0) then
begin
DebugMsg('Optimizer: FCMP #0.0 to FTST',p);
taicpu(p).opcode:=A_FTST;
taicpu(p).opsize:=S_FX;
taicpu(p).loadoper(0,taicpu(p).oper[1]^);
taicpu(p).clearop(1);
taicpu(p).ops:=1;
result:=true;
end
else
begin
tmpsingle:=taicpu(p).oper[0]^.val_real;
if (taicpu(p).opsize = S_FD) and
((taicpu(p).oper[0]^.val_real - tmpsingle) = 0.0) then
begin
DebugMsg('Optimizer: FCMP const to lesser precision',p);
taicpu(p).opsize:=S_FS;
result:=true;
end;
end;
end;
A_FMOVE,A_FMUL,A_FADD,A_FSUB,A_FDIV:
begin
if (taicpu(p).opcode = A_FMOVE) and TryToOptimizeMove(p) then
begin
result:=true;
exit;
end;
if (taicpu(p).oper[0]^.typ = top_realconst) then
begin
tmpsingle:=taicpu(p).oper[0]^.val_real;
if (taicpu(p).opsize = S_FD) and
((taicpu(p).oper[0]^.val_real - tmpsingle) = 0.0) then
begin
DebugMsg('Optimizer: FMOVE/FMUL/FADD/FSUB/FDIV const to lesser precision',p);
taicpu(p).opsize:=S_FS;
result:=true;
end;
end;
end;
end;
end;
end;
end;
begin
casmoptimizer:=TCpuAsmOptimizer;
end.
|
