-
Notifications
You must be signed in to change notification settings - Fork 5
/
test_instruction.py
283 lines (240 loc) · 8.61 KB
/
test_instruction.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
import binascii
import bz2
from collections import namedtuple
import csv
import os
import re
import pytest
from .instruction import Instruction, InstructionType, sign_extend
from .disassembly import disassemble_instruction, tokens_to_text
DIR = os.path.dirname(__file__)
def test_decode_abs():
# RRR type
# ABS ar, at
# 0110 0000 rrrr 0001 tttt 0000
# 60 r1 t0
# ABS a7, a9
# 60 71 90 => 907160
INSN_ABS = binascii.unhexlify("907160")
insn = Instruction.decode(INSN_ABS)
assert insn.op0 == 0
assert insn.op1 == 0
assert insn.op2 == 6
assert insn.r == 7
assert insn.t == 9
assert insn.s == 1
assert insn.length == 3
assert insn.mnem == "ABS"
assert insn.instruction_type == InstructionType.RRR
def test_decode_add():
"""
ADD ar, as, at
ADD a3, a2, a1
* bit 23
* 1000 # op2
* 0000 # op1
* 0011 # a3 is r
* 0010 # a2 is s
* 0001 # a1 is t
* 0000 # op0
* bit 0
Thus our insn is 80 32 10, which must be byte swapped to 10 32 80
"""
#EveryInstR Group
insn = Instruction.decode(binascii.unhexlify("103280"))
assert insn.op0 == 0
assert insn.op1 == 0
assert insn.op2 == 8
assert insn.r == 3
assert insn.s == 2
assert insn.t == 1
assert insn.length == 3
assert insn.mnem == "ADD"
assert insn.instruction_type == InstructionType.RRR
def test_add_narrow():
"""
ADD.N ar, as, at
* bit 15
* rrrr
* ssss
* tttt
* 1010 # op0
* bit 0
Requires Code Density Option
ADD.N a9, a5, a3
is then 1001 0101 0011 1010, or 953a, reversed to 3a95
"""
INSN_ADD_N = binascii.unhexlify("3a95")
insn = Instruction.decode(INSN_ADD_N)
assert insn.op0 == 0b1010
assert insn.t == 3
assert insn.s == 5
assert insn.r == 9
assert insn.length == 2
assert insn.mnem == "ADD.N"
def test_addi():
"""
RRI8 type
ADDI at, as, -128..127
* bit 23
* imm8 # check encoding of this
* 1100
* s
* t
* 0010
* bit 0
ADDI a11, a1, -2
is then
1111 1110 1100 0001 1011 0010, or fe c1 b2, reversed to b2c1fe
"""
insn = Instruction.decode(binascii.unhexlify("b2c1fe"))
assert insn.op0 == 0b0010
assert insn.r == 0b1100
assert insn.s == 1
assert insn.t == 11
# TODO: handle and test negative handling. I'd argue it should be a separate
# value, as the decoded imm8 doesn't seem like a signed value
#assert insn.imm8 == -2
assert insn.imm8 == 0b11111110
assert insn.length == 3
assert insn.mnem == "ADDI"
assert insn.instruction_type == InstructionType.RRI8
test_mnemonics_data = []
with bz2.open(os.path.join(DIR, "test_mnemonics.csv.bz2"), "rt") as fp:
reader = csv.reader(fp)
for row in reader:
opcode = row[0]
mnem = row[1]
opbytes = binascii.unhexlify(opcode)
test_mnemonics_data.append((opbytes, mnem.strip()))
def test_mnemonics_data_is_valid():
assert len(test_mnemonics_data) > 0
assert len(test_mnemonics_data[0]) == 2
def compare_mnem(one, two):
to_compare = []
for it in (one, two):
if (it.startswith("rsr.") or
it.startswith("wsr.") or
it.startswith("xsr.")):
# Work around not having the register names for special regs
it = it[:3]
it = it.lower().strip()
to_compare.append(it)
one, two = to_compare
return one == two
@pytest.mark.parametrize("opbytes,mnem_expected", test_mnemonics_data)
def test_mnem_from_file(opbytes, mnem_expected):
insn = Instruction.decode(opbytes)
assert insn.length == len(opbytes)
assert compare_mnem(insn.mnem, mnem_expected)
mtd_re = r'([0-9a-f]+):\s+([0-9a-f]+)\s+([a-z0-9.]+)\s+(.*)$'
mtd_rec = re.compile(mtd_re)
with bz2.open(os.path.join(DIR, "test_mnemonic_text.dump.bz2"), "rt") as fp:
mnem_text_dump = fp.readlines()
def bswap_opcode_string(opstr):
data = binascii.unhexlify(opstr)
reverse_data = bytearray(data)
reverse_data.reverse()
return binascii.hexlify(reverse_data).decode('utf-8')
DisassLine = namedtuple('DisassLine', ['addr', 'opcode', 'mnem', 'rest'])
def parse_test_data(data_lines):
newdata = []
for line in data_lines:
match_obj = mtd_rec.match(line)
assert match_obj
addr, opcode, mnem, rest = match_obj.groups()
opcode = bswap_opcode_string(opcode)
assert len(addr)
assert len(opcode)
assert len(mnem)
newdata.append(DisassLine(addr, opcode, mnem, rest))
return newdata
def test_mtd_re():
data = parse_test_data(mnem_text_dump)
assert len(data) > 0
assert len(data[0]) == 4
def _normalize_insn(it):
it = it.replace("\t", "").lower()
tokens = []
for tok in it.split():
tok = tok.replace(",", "")
if tok.startswith("0x"):
tokens.append(str(sign_extend(int(tok, 0), 32)))
else:
tokens.append(tok)
return ''.join(tokens)
def compare_insn(one, two):
one = _normalize_insn(one)
two = _normalize_insn(two)
return one == two
def test_tokens_to_text():
INSN_ABS = binascii.unhexlify("907160")
insn = Instruction.decode(INSN_ABS)
disass_text = tokens_to_text(disassemble_instruction(insn, 0))
assert compare_insn(disass_text, "ABS a7, a9")
assert compare_insn(disass_text, "abs a7, a9")
mtd_data = parse_test_data(mnem_text_dump)
# mnem_text_dump is a bunch of dumped disassembly, uniq'd on the mnem for
# brevity
@pytest.mark.parametrize("parsed_line", mtd_data)
def test_mnem_text_dump(parsed_line):
insn = Instruction.decode(binascii.unhexlify(parsed_line.opcode))
assert compare_mnem(insn.mnem, parsed_line.mnem)
addr = int(parsed_line.addr, 16)
disass_text = tokens_to_text(disassemble_instruction(insn, addr))
expected_insn_text = (parsed_line.mnem + " " + parsed_line.rest).strip()
assert compare_insn(expected_insn_text, disass_text)
with bz2.open(os.path.join(DIR, "torture_test.dump.bz2"), "rt") as fp:
lots_text_dump = fp.readlines()
lots_data = parse_test_data(lots_text_dump)
# lots_text_dump is a bunch of dumped disassembly, uniq'd on the mnem for
# brevity
@pytest.mark.parametrize("parsed_line", lots_data)
def test_lots_text_dump(parsed_line):
insn = Instruction.decode(binascii.unhexlify(parsed_line.opcode))
assert compare_mnem(insn.mnem, parsed_line.mnem)
addr = int(parsed_line.addr, 16)
disass_text = tokens_to_text(disassemble_instruction(insn, addr))
expected_insn_text = (parsed_line.mnem + " " + parsed_line.rest).strip()
assert compare_insn(expected_insn_text, disass_text)
with bz2.open( os.path.join(DIR, "esp32_torture_test.dump.bz2"), "rt") as fp:
esp32_lots_text_dump = fp.readlines()
esp32_lots_data = parse_test_data(esp32_lots_text_dump)
# lots_text_dump is a bunch of dumped disassembly, uniq'd on the mnem for
# brevity
@pytest.mark.parametrize("esp32_parsed_line", esp32_lots_data)
def test_lots_text_dump(esp32_parsed_line):
if esp32_parsed_line.mnem in ['rer', 'wer']:
# I disagree with objdump here; the manual states that these insns take
# arguments; objdump doesn't appear to think so? Also possible my
# cleanup of the output broke the objdump results?
pytest.xfail()
insn = Instruction.decode(binascii.unhexlify(esp32_parsed_line.opcode))
assert compare_mnem(insn.mnem, esp32_parsed_line.mnem)
addr = int(esp32_parsed_line.addr, 16)
disass_text = tokens_to_text(disassemble_instruction(insn, addr))
expected_insn_text = (esp32_parsed_line.mnem + " " +
esp32_parsed_line.rest).strip()
assert compare_insn(expected_insn_text, disass_text)
def test_rotw_positive():
rotw_insn = binascii.unhexlify("208040") # ROTW 2
insn = Instruction.decode(rotw_insn)
assert compare_mnem(insn.mnem, "ROTW")
assert insn.rotw_simm4() == 2
disass_text = tokens_to_text(disassemble_instruction(insn, 0x1000))
assert compare_insn(disass_text, "ROTW 2")
def test_rotw_negative():
rotw_insn = binascii.unhexlify("f08040") # ROTW -1
insn = Instruction.decode(rotw_insn)
assert compare_mnem(insn.mnem, "ROTW")
assert insn.rotw_simm4() == -1
disass_text = tokens_to_text(disassemble_instruction(insn, 0x1000))
assert compare_insn(disass_text, "ROTW -1")
# We didn't have any tests for the FPU, which lead to an undetected typo
def test_mov_s_fpu():
movs_insn = binascii.unhexlify("0012fa")
insn = Instruction.decode(movs_insn)
assert compare_mnem(insn.mnem, "MOV.S")
# Disassembly support does not yet exist
#disass_text = tokens_to_text(disassemble_instruction(insn, 0x1000))
#assert compare_insn(disass_text, "MOV.S f1, f2")