-
Notifications
You must be signed in to change notification settings - Fork 1
/
codec.py
executable file
·753 lines (692 loc) · 33.2 KB
/
codec.py
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
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
#!/usr/bin/python
# **********************************************************
# Copyright (c) 2021 Google, Inc. All rights reserved.
# Copyright (c) 2016 - 2022 ARM Limited. All rights reserved.
# **********************************************************
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
#
# * Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
#
# * Neither the name of ARM Limited nor the names of its contributors may be
# used to endorse or promote products derived from this software without
# specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL ARM LIMITED OR CONTRIBUTORS BE LIABLE
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
# OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
# DAMAGE.
# This script reads opnd_defs.txt, codec_<version>.txt files and generates:
# opnd_decode_funcs.h
# opnd_encode_funcs.h
# encode_gen_<version>.h
# decode_gen_<version>.h
# opcode_names.h
# opcode_api.h
# opcode_opnd_pairs.h
# It is automatically run by cmake when opnd_defs.txt or codec_<version>.txt
# change.
import os
import re
import sys
N = 32 # bits in an instruction word
ONES = (1 << N) - 1
# Stores instances of FallthroughDecode objects for resolution of overlapping
# encodings.
FALLTHROUGH = dict()
opnd_header = '/* This file was generated by codec.py from opnd_defs.txt and the codec_<version>.txt files. */\n\n'
opcode_header = '/* This file was generated by codec.py from codec_<version>.txt files. */\n\n'
class Opcode:
def __init__(self, name, nzcv_rw, feat):
self.name = name
self.nzcv_rw = nzcv_rw
self.feat = feat
class Opnd:
def __init__(self, gen, used, must_be_set):
self.gen = gen
self.used = used
self.non_zero = must_be_set
class Opndset:
def __init__(self, fixed, dsts, srcs, enc_order):
for (ds, i, ot) in enc_order:
if not ((ds == 'dst' or ds == 'src') and
(dsts if ds == 'dst' else srcs)[i] == ot):
raise Exception
self.fixed = fixed
self.dsts = dsts
self.srcs = srcs
self.enc_order = enc_order
class FallthroughDecode:
def __init__(self, opcode, opndset='', decode_clause='', decode_function=''):
self.flag_name = opcode + '_fallthrough_flag'
self.opndset = opndset
self.decode_clause = decode_clause
self.decode_function = decode_function
class Pattern:
def __init__(self, pattern, opcode_bits, opnd_bits, high_soft_bits, opcode, opndset, enum, feat):
self.pattern = pattern
self.opcode_bits = opcode_bits
self.opnd_bits = opnd_bits
# High soft bits are bits that are allowed to vary by the spec but
# default to 1.They are represented by ^ in codec.txt
self.high_soft_bits = high_soft_bits
self.opcode = opcode
self.opndset = opndset
self.enum = enum
self.feat = feat
def __iter__(self):
for field in (self.opcode_bits, self.opnd_bits, self.opcode, self.opndset):
yield field
def all_opnds(self):
if isinstance(self.opndset, tuple):
return self.opndset[0] + self.opndset[1]
if isinstance(self.opndset, str):
return [self.opndset]
return self.opndset
def ignored_bit_mask(self):
return ~(self.opnd_bits | self.high_soft_bits)
def set_bits(self):
return self.opcode_bits | self.high_soft_bits
def codec_header(isa_version):
return '/* This file was generated by codec.py from codec_%s.txt. */\n\n' % isa_version
def fallthrough_instr_id(opcode, opcode_bits, opnd_bits):
return '%s_%08x_%08x' % (opcode, opcode_bits, opnd_bits)
def generate_opndset_decoders(opndsettab, opndtab):
c = []
c += ['bool {} = false;'.format(opcode.flag_name) for opcode in
FALLTHROUGH.values()]
c += ['\n']
for name in sorted(opndsettab):
opnd_set = opndsettab[name]
(dsts, srcs) = (opnd_set.dsts, opnd_set.srcs)
c += ['/* %s <- %s */' % (opnd_set.dsts, opnd_set.srcs)]
c += ['static bool',
'decode_opnds%s(uint enc, byte *pc, '
'instr_t *instr, int opcode)' % name, '{']
if dsts + srcs != []:
vars = (['dst%d' % i for i in range(len(dsts))] +
['src%d' % i for i in range(len(srcs))])
tests = (['!decode_opnd_%s(enc & 0x%08x, opcode, pc, &dst%d)' %
(dsts[i], opndtab[dsts[i]].gen | opndtab[dsts[i]].used, i)
for i in range(len(dsts))]
+
['!decode_opnd_%s(enc & 0x%08x, opcode, pc, &src%d)' %
(srcs[i], opndtab[srcs[i]].gen | opndtab[srcs[i]].used, i)
for i in range(len(srcs))])
c += [' opnd_t ' + ', '.join(vars) + ';']
c += [' if (' + ' ||\n '.join(tests) + ')']
c += [' return false;']
c.append(' instr_set_opcode(instr, opcode);')
c.append(' instr_set_num_opnds(instr, %d, %d);' %
(len(dsts), len(srcs)))
for i in range(len(dsts)):
c.append(' instr_set_dst(instr, %d, dst%d);' % (i, i))
for i in range(len(srcs)):
c.append(' instr_set_src(instr, %d, src%d);' % (i, i))
c.append(' return true;')
c.append('}')
c.append('')
return '\n'.join(c) + '\n'
def generate_decoder(patterns, opndsettab, opndtab, opc_props, curr_isa, next_isa):
# Recursive function to generate nested conditionals in main decoder.
def gen(c, pats, depth):
def reorder_key(t):
f, v, m, t = t
return (m, t, f, v)
def indent_append(text):
c.append('{}{}'.format(' ' * depth, text))
if len(pats) < 4:
else_str = ''
for pattern in sorted(pats, key=reorder_key):
not_zero_mask = 0
try:
opnd_set = opndsettab[pattern.opndset]
for mask in (opndtab[o].non_zero for o in opnd_set.dsts + opnd_set.srcs):
not_zero_mask |= mask
except KeyError:
pass
indent_append('%sif ((enc & 0x%08x) == 0x%08x%s)' % (
else_str,
((1 << N) - 1) & pattern.ignored_bit_mask(),
pattern.opcode_bits,
' && (enc & 0x%08x) != 0' % not_zero_mask if not_zero_mask else ''))
if not else_str:
else_str = 'else '
if opc_props[pattern.opcode].nzcv_rw != 'n':
c[-1] = c[-1] + ' {'
# Uncomment this for debug output in generated code:
# indent_append(' // %s->%s' % (m, opc_props[m].nzcv_rw))
if opc_props[pattern.opcode].nzcv_rw == 'r':
indent_append(' instr->eflags |= EFLAGS_READ_NZCV;')
elif opc_props[pattern.opcode].nzcv_rw == 'w':
indent_append(' instr->eflags |= EFLAGS_WRITE_NZCV;')
elif opc_props[pattern.opcode].nzcv_rw in ['rw', 'wr']:
indent_append(' instr->eflags |= (EFLAGS_READ_NZCV | '
'EFLAGS_WRITE_NZCV);')
elif opc_props[pattern.opcode].nzcv_rw in ['er', 'ew']:
indent_append(
' // instr->eflags handling for %s is '
'manually handled in codec.c\'s decode_common().' % pattern.opcode)
else:
indent_append(' ASSERT(0);')
enc_key = fallthrough_instr_id(
pattern.opcode, pattern.opcode_bits, pattern.opnd_bits)
if enc_key in FALLTHROUGH and patten.opndset == FALLTHROUGH[enc_key].opndset:
indent_append(' %s = true;' % FALLTHROUGH[enc_key].flag_name)
FALLTHROUGH[enc_key].decode_clause = \
'if ((enc & 0x%08x) == 0x%08x && %s == true)' % \
(((1 << N) - 1) & pattern.ignored_bit_mask(), pattern.opcode_bits, \
FALLTHROUGH[enc_key].flag_name)
FALLTHROUGH[enc_key].decode_function = \
'return decode_opnds%s(enc, pc, instr, OP_%s);' % \
(pattern.opndset, pattern.opcode)
else:
indent_append(' return decode_opnds%s(enc, pc, '
'instr, OP_%s);' % (pattern.opndset, pattern.opcode))
if opc_props[pattern.opcode].nzcv_rw != 'n':
indent_append('}')
return
# Look for best bit to test. We aim to reduce the number of patterns
# remaining.
best_switch_bit = -1
least_patterns_selected = len(pats)
for switch_bit in range(N):
bit_not_set_or_variable = 0
bit_set_or_variable = 0
for p in pats:
# In how many patterns is this bit not set or included
# in the variable bits.
if (1 << switch_bit) & (~p.opcode_bits | p.opnd_bits | p.high_soft_bits):
bit_not_set_or_variable += 1
# How many patterns have this b set and or in the variable
# bits.
if (1 << switch_bit) & (p.opcode_bits | p.opnd_bits | p.high_soft_bits):
bit_set_or_variable += 1
patterns_selected = max(bit_not_set_or_variable, bit_set_or_variable)
if patterns_selected < least_patterns_selected:
best_switch_bit = switch_bit
least_patterns_selected = patterns_selected
indent_append('if ((enc >> %d & 1) == 0) {' % (best_switch_bit,))
pats0 = []
pats1 = []
# Split the decode tree on this bit, if the bit
# lies in the operand bits, then it goes in both trees.
for p in pats:
if (1 << best_switch_bit) & (~p.opcode_bits | p.opnd_bits | p.high_soft_bits):
pats0.append(p)
if (1 << best_switch_bit) & (p.opcode_bits | p.opnd_bits | p.high_soft_bits):
pats1.append(p)
gen(c, pats0, depth + 1)
indent_append('} else {')
gen(c, pats1, depth + 1)
indent_append('}')
c = ['static bool',
'decoder_' + curr_isa + '(uint enc, byte *pc, instr_t *instr)',
'{']
gen(c, patterns, 1)
for opcode in FALLTHROUGH.values():
c += [' %s' % opcode.decode_clause]
c += [' %s' % opcode.decode_function]
# Call the next version of the decoder if defined.
if next_isa != '':
c.append(' return decoder_' + next_isa + '(enc, pc, instr);')
else:
c.append(' return false;')
c.append('}')
return '\n'.join(c) + '\n'
def find_required(fixed, reordered, i, opndtab):
known = fixed
used = opndtab[reordered[i][2]].used
req = []
for j in range(i):
if used & ~known == 0:
break
if opndtab[reordered[j][2]].gen & used & ~known != 0:
req = req + ['%s%d' % (reordered[j][0], reordered[j][1])]
known = known | opndtab[reordered[j][2]].gen
return 'enc' if req == [] else '(enc | %s)' % ' | '.join(req)
def make_enc(n, reordered, f, opndtab):
(ds, i, ot) = reordered[n]
instr_arg_if_required = 'instr, ' if ot == 'imm16' else ''
encode_method_format_str = ('encode_opnd_%s(%s, opcode, pc, '
'instr_get_%s(instr, %d), ' +
instr_arg_if_required + '&%s%d)')
ret_str = (encode_method_format_str %
(ot, ('0' if opndtab[ot].used == 0 else
'enc & 0x%08x' % opndtab[ot].used
if opndtab[ot].used & ~f == 0 else
'%s & 0x%08x' % (find_required(f, reordered, n, opndtab),
opndtab[ot].used)),
ds, i, ds, i))
return ret_str
def generate_opndset_encoders(opndsettab, opndtab):
c = []
for name in sorted(opndsettab):
os = opndsettab[name]
(fixed, dsts, srcs, enc_order) = (os.fixed, os.dsts, os.srcs, os.enc_order)
c += ['/* %s <- %s */' % (os.dsts, os.srcs)]
c += ['static uint',
('encode_opnds%s' % name) +
'(byte *pc, instr_t *instr, uint enc, decode_info_t *di)',
'{']
if dsts + srcs == []:
c.append(' return enc;')
else:
vars = (['dst%d' % i for i in range(len(dsts))] +
['src%d' % i for i in range(len(srcs))])
c += [' int opcode = instr->opcode;']
# The initial values are only required to silence a bad compiler warning:
c += [' uint ' + ' = 0, '.join(vars) + ' = 0;']
tests = (['instr_num_dsts(instr) == %d && instr_num_srcs(instr) == %d' %
(len(dsts), len(srcs))] +
[make_enc(i, enc_order, fixed, opndtab)
for i in range(len(enc_order))])
tests2 = (['dst%d == (enc & 0x%08x)' % (i, opndtab[dsts[i]].gen)
for i in range(len(dsts))] +
['src%d == (enc & 0x%08x)' % (i, opndtab[srcs[i]].gen)
for i in range(len(srcs))])
c += [' if (' + ' &&\n '.join(tests) + ') {']
c += [' ASSERT((dst%d & 0x%08x) == 0);' %
(i, ONES & ~opndtab[dsts[i]].gen) for i in range(len(dsts))]
c += [' ASSERT((src%d & 0x%08x) == 0);' %
(i, ONES & ~opndtab[srcs[i]].gen) for i in range(len(srcs))]
c += [' enc |= ' + ' | '.join(vars) + ';']
c += [' if (' + ' &&\n '.join(tests2) + ')']
c += [' return enc;']
c += [' }']
c += [' return ENCFAIL;']
c.append('}')
c.append('')
return '\n'.join(c) + '\n'
def generate_encoder(patterns, opndsettab, opndtab, opc_props, curr_isa, next_isa):
c = []
case = dict()
for p in patterns:
(opcode_bits, opnd_bits, opcode, opndset) = p
if opcode not in case:
case[opcode] = []
case[opcode].append(p)
c += ['static uint',
'encoder_' + curr_isa + '(byte *pc, instr_t *instr, decode_info_t *di)',
'{',
' uint enc;',
' (void)enc;',
' switch (instr->opcode) {']
def reorder_key(t):
opcode_bits, opnd_bits, opcode, opndset = t
return (opcode, opndset, opcode_bits, opnd_bits)
for opcode in sorted(case):
c.append(' case OP_%s:' % opcode)
if opc_props[opcode].feat != 'BASE':
c.append('# if !defined(DR_HOST_NOT_TARGET) && !defined(STANDALONE_DECODER)')
c.append(' if (!proc_has_feature(FEATURE_%s))' % opc_props[opcode].feat)
c.append(' return ENCFAIL;')
c.append('# endif')
patterns = sorted(case[opcode], key=reorder_key)
last_pattern = patterns.pop()
for pattern in patterns:
c.append(' enc = encode_opnds%s(pc, instr, 0x%08x, di);' % (
pattern.opndset, pattern.set_bits()))
c.append(' if (enc != ENCFAIL)')
c.append(' return enc;')
# Fallthrough to call the next version of the encoder if defined.
if next_isa != '':
c.append(' enc = encode_opnds%s(pc, instr, 0x%08x, di);' % (
last_pattern.opndset, last_pattern.set_bits()))
c.append(' if (enc != ENCFAIL)')
c.append(' return enc;')
c += [' break;']
else:
c.append(' return encode_opnds%s(pc, instr, 0x%08x, di);' % (
last_pattern.opndset, last_pattern.set_bits()))
c += [' }']
# Call the next version of the encoder if defined.
if next_isa != '':
c += [' return encoder_' + next_isa + '(pc, instr, di);']
else:
c += [' return ENCFAIL;']
c += ['}']
return '\n'.join(c) + '\n'
def generate_opcodes(patterns):
c = ['#ifndef _DR_IR_OPCODES_AARCH64_H_',
'#define _DR_IR_OPCODES_AARCH64_H_ 1',
'',
'/****************************************************************************',
' * OPCODES',
' */',
'/**',
' * @file dr_ir_opcodes_aarch64.h',
' * @brief Instruction opcode constants for AArch64.',
' */',
'/** Opcode constants for use in the instr_t data structure. */',
'enum {',
'/* 0 */ OP_INVALID, /* NULL, */ /**< INVALID opcode */',
'/* 1 */ OP_UNDECODED, /* NULL, */ /**< UNDECODED opcode */',
'/* 2 */ OP_CONTD, /* NULL, */ /**< CONTD opcode */',
'/* 3 */ OP_LABEL, /* NULL, */ /**< LABEL opcode */',
'/* 4 */ OP_xx, /* placeholder for undecoded instructions */',
'/* 5 */ OP_ldstex, /* single-entry single-exit block with exclusive load/store */',
'']
pattern_d = {int(p.enum): p.opcode for p in patterns}
for i in range(6, max(pattern_d.keys()) + 1):
try:
c.append('/* {i:>4} */ OP_{opcode} = {i}, /**< AArch64 {opcode} opcode. */'.format(
i=i, opcode=pattern_d[i]))
except KeyError:
pass
c += ['',
'',
' OP_AFTER_LAST,',
' OP_FIRST = OP_LABEL + 1, /**< First real opcode. */',
' OP_LAST = OP_AFTER_LAST - 1, /**< Last real opcode. */',
'};',
'',
'/* alternative names */',
'#define OP_jmp OP_b '
'/**< Platform-independent opcode name for jump. */',
'#define OP_jmp_short OP_b '
'/**< Platform-independent opcode name for short jump. */',
'#define OP_load OP_ldr '
'/**< Platform-independent opcode name for load. */',
'#define OP_store OP_str '
'/**< Platform-independent opcode name for store. */',
'',
'/******************************'
'**********************************************/',
'',
'#endif /* _DR_IR_OPCODES_AARCH64_H */']
return '\n'.join(c) + '\n'
def generate_opcode_names(patterns):
c = ['#ifndef OPCODE_NAMES_H',
'#define OPCODE_NAMES_H 1',
'',
'const char *opcode_names[] = {',
'/* 0 */ "<invalid>",',
'/* 1 */ "<undecoded>",',
'/* 2 */ "<contd>",',
'/* 3 */ "<label>",',
'/* 4 */ "xx",',
'/* 5 */ "ldstex",']
pattern_d = {int(p.enum): p for p in patterns}
for i in range(6, max(pattern_d.keys()) + 1):
try:
name = pattern_d[i].opcode
except KeyError:
name = "<invalid>"
c.append('/*{:>4} */ "{}",'.format(i, name))
c += [
'};',
'',
'#endif /* OPCODE_NAMES_H */']
return '\n'.join(c) + '\n'
# Generates pairs of opcodes and masks for their operands for all the
# side-effect-free and non-branch instructions suppported by the decoder.
# The generated file is used by the drstatecmp-fuzz-app.
def generate_opcode_opnd_pairs(patterns):
c = ['#ifndef DR_OPCODE_OPND_PAIRS_H',
'#define DR_OPCODE_OPND_PAIRS_H 1',
'',
'#include <stdint.h>',
'',
'typedef struct {',
' uint32_t opcode;',
' uint32_t opnd;',
'} dr_opcode_opnd_pair_t;',
'',
'const dr_opcode_opnd_pair_t dr_fuzz_opcode_opnd_pairs[] = {']
# Exclude instructions with side-effects and branch instructions.
# Particularly, exclude: i) Load/Stores (opcode: x1x0);
# ii) Branches, Exception Generating and System instructions (opcode: 101x);
# iii) SVE memory (opcode: 0010 and 1 for the most significant bit).
# The allowed instructions include all the Data Processing instructions (including
# the FP and SIMD ones) and the non-memory SVE instructions.
excluded_opcodes = re.compile('....1.0|...101.|1..0010')
cnt = 0
for p in patterns:
if excluded_opcodes.match(p.pattern):
continue
cnt += 1
c.append('/* %s */ {%s, %s},' % (p.opcode, bin(p.opcode_bits), bin(p.opnd_bits)))
c += ['};',
'']
c.append('#define DR_FUZZ_INST_CNT %d' % cnt)
c += ['',
'#endif /* DR_OPCODE_OPND_PAIRS_H */']
return '\n'.join(c) + '\n'
def write_if_changed(file, data):
try:
if open(file, 'r').read() == data:
return
except IOError:
pass
open(file, 'w').write(data)
def read_opnd_defs_file(path):
opndtab = dict()
file_msg = 'operand definitions file'
try:
with open(path, 'r') as file:
for line in (l.split('#')[0].strip() for l in file):
if not line:
continue
if not re.match('^[x\?\-\+]{32} +[a-zA-Z_0-9]+$', line):
raise Exception('Cannot parse line: %s in %s' % (line, file_msg))
# Syntax: mask opndtype
mask, opndtype = line.split()
if opndtype in opndtab:
raise Exception('Repeated definition of opndtype %s in %s' % (opndtype, file_msg))
opndtab[opndtype] = Opnd(int(re.sub('[x\+]', '1', re.sub('[^x^\+]', '0', mask)), 2),
int(re.sub('\?', '1', re.sub('[^\?]', '0', mask)), 2),
int(re.sub('\+', '1', re.sub('[^\+]', '0', mask)), 2))
except IOError as e:
raise Exception('Unable to read operand definitions file, {}: {}'.format(path, e.strerror))
return opndtab
def read_codec_file(path):
opc_props = dict()
patterns = []
try:
with open(path, 'r') as file:
for line in (l.split('#')[0].strip() for l in file):
if not line:
continue
if re.match('^[01x\^]{32} +[n|r|w|rw|wr|er|ew]+ +[0-9]+ +[a-zA-Z0-9]* +[a-zA-Z_0-9][a-zA-Z_0-9 ]*:[a-zA-Z_0-9 ]*$', line):
# Syntax: pattern opcode opndtype* : opndtype*
pattern, nzcv_rw_flag, enum, feat, opcode, args = line.split(None, 5)
dsts, srcs = [a.split() for a in args.split(':')]
opcode_bits = int(re.sub('[\^x]', '0', pattern), 2)
opnd_bits = int(re.sub('x', '1', re.sub('[1\^]', '0', pattern)), 2)
high_soft_bits = int(re.sub('\^', '1', re.sub('[10x]', '0', pattern)), 2)
patterns.append(Pattern(pattern, opcode_bits, opnd_bits, high_soft_bits, opcode, (dsts, srcs), enum, feat))
opc_props[opcode] = Opcode(opcode, nzcv_rw_flag, feat)
continue
if re.match('^[01x\^]{32} +[n|r|w|rw|wr|er|ew]+ +[0-9]+ +[a-zA-Z0-9]* +[a-zA-Z_0-9]+ +[a-zA-Z_0-9]+', line):
# Syntax: pattern opcode opndset
pattern, nzcv_rw_flag, enum, feat, opcode, opndset = line.split()
opcode_bits = int(re.sub('x', '0', pattern), 2)
opnd_bits = int(re.sub('x', '1', re.sub('1', '0', pattern)), 2)
high_soft_bits = int(re.sub('\^', '1', re.sub('[10x]', '0', pattern)), 2)
patterns.append(Pattern(pattern, opcode_bits, opnd_bits, high_soft_bits, opcode, opndset, enum, feat))
opc_props[opcode] = Opcode(opcode, nzcv_rw_flag, feat)
continue
raise Exception('Cannot parse line: %s in %s' % (line, path))
except IOError as e:
print('Unable to read instruction definitions file, %s: %s' % (path, e.strerror))
sys.exit(1)
return patterns, opc_props
def pattern_to_str(opcode_bits, opnd_bits, opcode, opndset):
p = ''
for i in range(N - 1, -1, -1):
p += 'x' if (opnd_bits >> i & 1) else '%d' % (opcode_bits >> i & 1)
t = opndset
if not type(t) is str:
(dsts, srcs) = t
t = ' '.join(dsts) + ' : ' + ' '.join(srcs)
return '%s %s %s' % (p, opcode, t)
def consistency_check(patterns, opndtab):
for p in patterns:
if not isinstance(p.opndset, str):
dsts, srcs = p.opndset
unhandled_bits = p.opnd_bits
for ot in dsts + srcs:
try:
unhandled_bits &= ~opndtab[ot].gen
except KeyError:
raise Exception('Undefined opndtype %s in:\n%s' %
(ot, pattern_to_str(*p)))
if unhandled_bits:
raise Exception('Unhandled bits:\n%32s in:\n%s' %
(re.sub('1', 'x', re.sub('0', ' ', bin(unhandled_bits)[2:])),
pattern_to_str(*p)))
# Detect and mark overlapping patterns for special handling. Named as
# 'fallthrough' because the special handling is done at the end of the
# decoder's main if/then/else clauses block.
for i, pattern_a in enumerate(patterns):
for pattern_b in patterns[:i]:
non_zero_bits_a = 0
non_zero_bits_b = 0
try:
for opnd in (opndtab[op] for op in pattern_a.all_opnds()):
non_zero_bits_a &= opnd.non_zero
except KeyError:
pass
try:
for opnd in (opndtab[op] for op in pattern_b.all_opnds()):
non_zero_bits_b &= opnd.non_zero
except KeyError:
pass
zero_overlap = (
non_zero_bits_a & pattern_b.opnd_bits == 0 or
non_zero_bits_b & pattern_b.opnd_bits == 0)
if ((pattern_b.opcode_bits ^ pattern_a.opcode_bits) &
~pattern_b.opnd_bits & ~pattern_a.opnd_bits == 0 and
not zero_overlap):
print('Overlap found between:\n%s\nand\n%s' %
(pattern_to_str(*pattern_b),
pattern_to_str(*pattern_a)))
enc_key = fallthrough_instr_id(pattern_a.opcode, pattern_a.opcode_bits, pattern_a.opnd_bits)
if enc_key in FALLTHROUGH:
raise Exception('Error: multiple overlaps detected for '
'%s. Unable to resolve.\n' % enc_key)
print('Resolving overlap.')
FALLTHROUGH[enc_key] = FallthroughDecode(enc_key)
# This function reorders the operands for encoding so that no operand encoder
# requires bits that are generated by an operand encoder that has not yet
# been executed.
def reorder_opnds(fixed, dsts, srcs, opndtab):
def reorder(known, opnds):
if opnds == []:
return []
for i in range(len(opnds)):
(_, _, _, gen, used) = opnds[i]
if used & ~known == 0:
return [opnds[i]] + reorder(known | gen, opnds[0:i] + opnds[i + 1:])
raise Exception('Cyclic dependency: %s' %
' '.join([x for (_, _, x, _, _) in opnds]))
opnds = ([('dst', i, dsts[i]) for i in range(len(dsts))] +
[('src', i, srcs[i]) for i in range(len(srcs))])
opnds_expanded = [(ds, i, ot, opndtab[ot].gen, opndtab[ot].used)
for (ds, i, ot) in opnds]
opnds_reordered = [(ds, i, ot)
for (ds, i, ot, _, _) in reorder(fixed, opnds_expanded)]
return Opndset(fixed, dsts, srcs, opnds_reordered)
# Here we give the opndsets names, which will be used in function names.
# Opndsets specified in 'codec_<version>.txt' are prefixed with '_', while
# generated names have the form 'gen_X', where X is the hex representation of
# the smallest matching instruction word.
def opndset_naming(patterns, opndtab):
opndsets = dict() # maps (dst, src, opnd_bits) to smallest pattern seen so far
for opcode_bits, opnd_bits, opcode, opndset in patterns:
if type(opndset) is not str:
dsts, srcs = opndset
h = (' '.join(dsts), ' '.join(srcs), opnd_bits)
if h not in opndsets or opcode_bits < opndsets[h]:
opndsets[h] = opcode_bits
opndsettab = dict() # maps generated name to original opndsets
for p in patterns:
if type(p.opndset) is str:
new_opndset = '_' + p.opndset
else:
(dsts, srcs) = p.opndset
h = (' '.join(dsts), ' '.join(srcs), p.opnd_bits)
new_opndset = 'gen_%08x_%08x' % (opndsets[h], p.opnd_bits)
reordered = reorder_opnds(ONES & ~p.opnd_bits, dsts, srcs, opndtab)
if not new_opndset in opndsettab:
opndsettab[new_opndset] = reordered
p.opndset = new_opndset
enc_key = fallthrough_instr_id(p.opcode, p.opcode_bits, p.opnd_bits)
if enc_key in FALLTHROUGH:
FALLTHROUGH[enc_key].opndset = new_opndset
return (patterns, opndsettab)
def main():
if len(sys.argv) != 3:
print('Usage: codec.py path/to/input_dir path/to/output_dir')
sys.exit(1)
input_dir = sys.argv[1]
output_dir = sys.argv[2]
# The Arm AArch64's architecture versions supported by the DynamoRIO codec.
# Currently, v8.0 is fully supported, while v8.1, v8.2 and SVE are
# partially supported. The null terminator element at the end is required
# by some generator functions to correctly generate links between each
# version's decode/encode logic.
isa_versions = ['v80', 'v81', 'v82', 'sve', '']
# Read the instruction operand definitions. Used by the codec when
# generating code to decode and encode instructions.
opndtab = read_opnd_defs_file(os.path.join(input_dir, 'opnd_defs.txt'))
# Read all instruction definitions and use the instructions' operand
# signatures (opndsets) to generate decoder and encoder functions.
opndset_decode_funcs = opnd_header
opndset_encode_funcs = opnd_header
for isa_version in isa_versions[:-1]:
patterns, opc_props = read_codec_file(os.path.join(input_dir, 'codec_' + isa_version + '.txt'))
# Consistency check within each codec_<version>.txt file. Global check
# across all codec_<version>.txt files is done by
# aarch64_check_codec_order.py and codecsort.py.
consistency_check(patterns, opndtab)
(patterns, opndsettab) = opndset_naming(patterns, opndtab)
opndset_decode_funcs += generate_opndset_decoders(opndsettab, opndtab)
opndset_encode_funcs += generate_opndset_encoders(opndsettab, opndtab)
write_if_changed(os.path.join(output_dir, 'opnd_decode_funcs.h'),
opndset_decode_funcs)
write_if_changed(os.path.join(output_dir, 'opnd_encode_funcs.h'),
opndset_encode_funcs)
# Read all instruction definitions and use the instructions' bitmask to
# generate decode and encode logic.
for idx, isa_version in enumerate(isa_versions):
(patterns, opc_props) = read_codec_file(os.path.join(input_dir, 'codec_' + isa_version + '.txt'))
(patterns, opndsettab) = opndset_naming(patterns, opndtab)
write_if_changed(os.path.join(output_dir, 'decode_gen_' + isa_version +'.h'),
codec_header(isa_version) + generate_decoder(patterns, opndsettab, opndtab, opc_props, isa_version, isa_versions[idx + 1]))
write_if_changed(os.path.join(output_dir, 'encode_gen_' + isa_version +'.h'),
codec_header(isa_version) + generate_encoder(patterns, opndsettab, opndtab, opc_props, isa_version, isa_versions[idx + 1]))
if isa_versions[idx + 1] == '':
break
# Generate opcode declarations and definitions for the API and fuzz
# testing.
opcode_api_patterns = []
opcode_names_patterns = []
opcode_opnd_pairs_patterns = []
for isa_version in isa_versions[:-1]:
(patterns, opc_props) = read_codec_file(os.path.join(input_dir, 'codec_' + isa_version + '.txt'))
opcode_api_patterns += patterns
opcode_names_patterns += patterns
opcode_opnd_pairs_patterns += patterns
write_if_changed(os.path.join(output_dir, 'opcode_api.h'),
opcode_header + generate_opcodes(opcode_api_patterns))
write_if_changed(os.path.join(output_dir, 'opcode_names.h'),
opcode_header + generate_opcode_names(opcode_names_patterns))
opcode_opnd_pairs_patterns.sort(key=lambda pat: pat.opcode)
write_if_changed(os.path.join(output_dir, 'opcode_opnd_pairs.h'),
opcode_header + generate_opcode_opnd_pairs(opcode_opnd_pairs_patterns))
if __name__ == '__main__':
main()