summaryrefslogtreecommitdiffstats
path: root/src/ukify/ukify.py
blob: 0ddf56ae5e2dabab308608ff9d8bfa7f9b70a36d (plain)
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
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
#!/usr/bin/env python3
# SPDX-License-Identifier: LGPL-2.1-or-later
#
# This file is part of systemd.
#
# systemd is free software; you can redistribute it and/or modify it
# under the terms of the GNU Lesser General Public License as published by
# the Free Software Foundation; either version 2.1 of the License, or
# (at your option) any later version.
#
# systemd 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 Lesser General Public License
# along with systemd; If not, see <https://www.gnu.org/licenses/>.

# pylint: disable=import-outside-toplevel,consider-using-with,unused-argument
# pylint: disable=unnecessary-lambda-assignment

import argparse
import configparser
import contextlib
import collections
import dataclasses
import datetime
import fnmatch
import itertools
import json
import os
import pathlib
import pprint
import pydoc
import re
import shlex
import shutil
import socket
import subprocess
import sys
import tempfile
import textwrap
from hashlib import sha256
from typing import (Any,
                    Callable,
                    IO,
                    Optional,
                    Sequence,
                    Union)

import pefile  # type: ignore

__version__ = '{{PROJECT_VERSION}} ({{GIT_VERSION}})'

EFI_ARCH_MAP = {
    # host_arch glob : [efi_arch, 32_bit_efi_arch if mixed mode is supported]
    'x86_64'       : ['x64', 'ia32'],
    'i[3456]86'    : ['ia32'],
    'aarch64'      : ['aa64'],
    'armv[45678]*l': ['arm'],
    'loongarch32'  : ['loongarch32'],
    'loongarch64'  : ['loongarch64'],
    'riscv32'      : ['riscv32'],
    'riscv64'      : ['riscv64'],
}
EFI_ARCHES: list[str] = sum(EFI_ARCH_MAP.values(), [])

def guess_efi_arch():
    arch = os.uname().machine

    for glob, mapping in EFI_ARCH_MAP.items():
        if fnmatch.fnmatch(arch, glob):
            efi_arch, *fallback = mapping
            break
    else:
        raise ValueError(f'Unsupported architecture {arch}')

    # This makes sense only on some architectures, but it also probably doesn't
    # hurt on others, so let's just apply the check everywhere.
    if fallback:
        fw_platform_size = pathlib.Path('/sys/firmware/efi/fw_platform_size')
        try:
            size = fw_platform_size.read_text().strip()
        except FileNotFoundError:
            pass
        else:
            if int(size) == 32:
                efi_arch = fallback[0]

    # print(f'Host arch {arch!r}, EFI arch {efi_arch!r}')
    return efi_arch


def page(text: str, enabled: Optional[bool]) -> None:
    if enabled:
        # Initialize less options from $SYSTEMD_LESS or provide a suitable fallback.
        os.environ['LESS'] = os.getenv('SYSTEMD_LESS', 'FRSXMK')
        pydoc.pager(text)
    else:
        print(text)


def shell_join(cmd):
    # TODO: drop in favour of shlex.join once shlex.join supports pathlib.Path.
    return ' '.join(shlex.quote(str(x)) for x in cmd)


def round_up(x, blocksize=4096):
    return (x + blocksize - 1) // blocksize * blocksize


def try_import(modname, name=None):
    try:
        return __import__(modname)
    except ImportError as e:
        raise ValueError(f'Kernel is compressed with {name or modname}, but module unavailable') from e


def maybe_decompress(filename):
    """Decompress file if compressed. Return contents."""
    f = open(filename, 'rb')
    start = f.read(4)
    f.seek(0)

    if start.startswith(b'\x7fELF'):
        # not compressed
        return f.read()

    if start.startswith(b'MZ'):
        # not compressed aarch64 and riscv64
        return f.read()

    if start.startswith(b'\x1f\x8b'):
        gzip = try_import('gzip')
        return gzip.open(f).read()

    if start.startswith(b'\x28\xb5\x2f\xfd'):
        zstd = try_import('zstd')
        return zstd.uncompress(f.read())

    if start.startswith(b'\x02\x21\x4c\x18'):
        lz4 = try_import('lz4.frame', 'lz4')
        return lz4.frame.decompress(f.read())

    if start.startswith(b'\x04\x22\x4d\x18'):
        print('Newer lz4 stream format detected! This may not boot!')
        lz4 = try_import('lz4.frame', 'lz4')
        return lz4.frame.decompress(f.read())

    if start.startswith(b'\x89LZO'):
        # python3-lzo is not packaged for Fedora
        raise NotImplementedError('lzo decompression not implemented')

    if start.startswith(b'BZh'):
        bz2 = try_import('bz2', 'bzip2')
        return bz2.open(f).read()

    if start.startswith(b'\x5d\x00\x00'):
        lzma = try_import('lzma')
        return lzma.open(f).read()

    raise NotImplementedError(f'unknown file format (starts with {start})')


class Uname:
    # This class is here purely as a namespace for the functions

    VERSION_PATTERN = r'(?P<version>[a-z0-9._-]+) \([^ )]+\) (?:#.*)'

    NOTES_PATTERN = r'^\s+Linux\s+0x[0-9a-f]+\s+OPEN\n\s+description data: (?P<version>[0-9a-f ]+)\s*$'

    # Linux version 6.0.8-300.fc37.ppc64le (mockbuild@buildvm-ppc64le-03.iad2.fedoraproject.org)
    # (gcc (GCC) 12.2.1 20220819 (Red Hat 12.2.1-2), GNU ld version 2.38-24.fc37)
    # #1 SMP Fri Nov 11 14:39:11 UTC 2022
    TEXT_PATTERN = rb'Linux version (?P<version>\d\.\S+) \('

    @classmethod
    def scrape_x86(cls, filename, opts=None):
        # Based on https://gitlab.archlinux.org/archlinux/mkinitcpio/mkinitcpio/-/blob/master/functions#L136
        # and https://www.kernel.org/doc/html/latest/x86/boot.html#the-real-mode-kernel-header
        with open(filename, 'rb') as f:
            f.seek(0x202)
            magic = f.read(4)
            if magic != b'HdrS':
                raise ValueError('Real-Mode Kernel Header magic not found')
            f.seek(0x20E)
            offset = f.read(1)[0] + f.read(1)[0]*256  # Pointer to kernel version string
            f.seek(0x200 + offset)
            text = f.read(128)
        text = text.split(b'\0', maxsplit=1)[0]
        text = text.decode()

        if not (m := re.match(cls.VERSION_PATTERN, text)):
            raise ValueError(f'Cannot parse version-host-release uname string: {text!r}')
        return m.group('version')

    @classmethod
    def scrape_elf(cls, filename, opts=None):
        readelf = find_tool('readelf', opts=opts)

        cmd = [
            readelf,
            '--notes',
            filename,
        ]

        print('+', shell_join(cmd))
        try:
            notes = subprocess.check_output(cmd, stderr=subprocess.PIPE, text=True)
        except subprocess.CalledProcessError as e:
            raise ValueError(e.stderr.strip()) from e

        if not (m := re.search(cls.NOTES_PATTERN, notes, re.MULTILINE)):
            raise ValueError('Cannot find Linux version note')

        text = ''.join(chr(int(c, 16)) for c in m.group('version').split())
        return text.rstrip('\0')

    @classmethod
    def scrape_generic(cls, filename, opts=None):
        # import libarchive
        # libarchive-c fails with
        # ArchiveError: Unrecognized archive format (errno=84, retcode=-30, archive_p=94705420454656)

        # Based on https://gitlab.archlinux.org/archlinux/mkinitcpio/mkinitcpio/-/blob/master/functions#L209

        text = maybe_decompress(filename)
        if not (m := re.search(cls.TEXT_PATTERN, text)):
            raise ValueError(f'Cannot find {cls.TEXT_PATTERN!r} in {filename}')

        return m.group('version').decode()

    @classmethod
    def scrape(cls, filename, opts=None):
        for func in (cls.scrape_x86, cls.scrape_elf, cls.scrape_generic):
            try:
                version = func(filename, opts=opts)
                print(f'Found uname version: {version}')
                return version
            except ValueError as e:
                print(str(e))
        return None

DEFAULT_SECTIONS_TO_SHOW = {
        '.linux'    : 'binary',
        '.initrd'   : 'binary',
        '.splash'   : 'binary',
        '.dt'       : 'binary',
        '.cmdline'  : 'text',
        '.osrel'    : 'text',
        '.uname'    : 'text',
        '.pcrpkey'  : 'text',
        '.pcrsig'   : 'text',
        '.sbat'     : 'text',
}

@dataclasses.dataclass
class Section:
    name: str
    content: Optional[pathlib.Path]
    tmpfile: Optional[IO] = None
    measure: bool = False
    output_mode: Optional[str] = None

    @classmethod
    def create(cls, name, contents, **kwargs):
        if isinstance(contents, (str, bytes)):
            mode = 'wt' if isinstance(contents, str) else 'wb'
            tmp = tempfile.NamedTemporaryFile(mode=mode, prefix=f'tmp{name}')
            tmp.write(contents)
            tmp.flush()
            contents = pathlib.Path(tmp.name)
        else:
            tmp = None

        return cls(name, contents, tmpfile=tmp, **kwargs)

    @classmethod
    def parse_input(cls, s):
        try:
            name, contents, *rest = s.split(':')
        except ValueError as e:
            raise ValueError(f'Cannot parse section spec (name or contents missing): {s!r}') from e
        if rest:
            raise ValueError(f'Cannot parse section spec (extraneous parameters): {s!r}')

        if contents.startswith('@'):
            contents = pathlib.Path(contents[1:])

        sec = cls.create(name, contents)
        sec.check_name()
        return sec

    @classmethod
    def parse_output(cls, s):
        if not (m := re.match(r'([a-zA-Z0-9_.]+):(text|binary)(?:@(.+))?', s)):
            raise ValueError(f'Cannot parse section spec: {s!r}')

        name, ttype, out = m.groups()
        out = pathlib.Path(out) if out else None

        return cls.create(name, out, output_mode=ttype)

    def size(self):
        return self.content.stat().st_size

    def check_name(self):
        # PE section names with more than 8 characters are legal, but our stub does
        # not support them.
        if not self.name.isascii() or not self.name.isprintable():
            raise ValueError(f'Bad section name: {self.name!r}')
        if len(self.name) > 8:
            raise ValueError(f'Section name too long: {self.name!r}')


@dataclasses.dataclass
class UKI:
    executable: list[Union[pathlib.Path, str]]
    sections: list[Section] = dataclasses.field(default_factory=list, init=False)

    def add_section(self, section):
        if section.name in [s.name for s in self.sections]:
            raise ValueError(f'Duplicate section {section.name}')

        self.sections += [section]


def parse_banks(s):
    banks = re.split(r',|\s+', s)
    # TODO: do some sanity checking here
    return banks


KNOWN_PHASES = (
    'enter-initrd',
    'leave-initrd',
    'sysinit',
    'ready',
    'shutdown',
    'final',
)

def parse_phase_paths(s):
    # Split on commas or whitespace here. Commas might be hard to parse visually.
    paths = re.split(r',|\s+', s)

    for path in paths:
        for phase in path.split(':'):
            if phase not in KNOWN_PHASES:
                raise argparse.ArgumentTypeError(f'Unknown boot phase {phase!r} ({path=})')

    return paths


def check_splash(filename):
    if filename is None:
        return

    # import is delayed, to avoid import when the splash image is not used
    try:
        from PIL import Image
    except ImportError:
        return

    img = Image.open(filename, formats=['BMP'])
    print(f'Splash image {filename} is {img.width}×{img.height} pixels')


def check_inputs(opts):
    for name, value in vars(opts).items():
        if name in {'output', 'tools'}:
            continue

        if isinstance(value, pathlib.Path):
            # Open file to check that we can read it, or generate an exception
            value.open().close()
        elif isinstance(value, list):
            for item in value:
                if isinstance(item, pathlib.Path):
                    item.open().close()

    check_splash(opts.splash)


def check_cert_and_keys_nonexistent(opts):
    # Raise if any of the keys and certs are found on disk
    paths = itertools.chain(
        (opts.sb_key, opts.sb_cert),
        *((priv_key, pub_key)
          for priv_key, pub_key, _ in key_path_groups(opts)))
    for path in paths:
        if path and path.exists():
            raise ValueError(f'{path} is present')


def find_tool(name, fallback=None, opts=None):
    if opts and opts.tools:
        for d in opts.tools:
            tool = d / name
            if tool.exists():
                return tool

    if shutil.which(name) is not None:
        return name

    if fallback is None:
        print(f"Tool {name} not installed!")

    return fallback

def combine_signatures(pcrsigs):
    combined = collections.defaultdict(list)
    for pcrsig in pcrsigs:
        for bank, sigs in pcrsig.items():
            for sig in sigs:
                if sig not in combined[bank]:
                    combined[bank] += [sig]
    return json.dumps(combined)


def key_path_groups(opts):
    if not opts.pcr_private_keys:
        return

    n_priv = len(opts.pcr_private_keys)
    pub_keys = opts.pcr_public_keys or [None] * n_priv
    pp_groups = opts.phase_path_groups or [None] * n_priv

    yield from zip(opts.pcr_private_keys,
                   pub_keys,
                   pp_groups)


def call_systemd_measure(uki, linux, opts):
    measure_tool = find_tool('systemd-measure',
                             '/usr/lib/systemd/systemd-measure',
                             opts=opts)

    banks = opts.pcr_banks or ()

    # PCR measurement

    if opts.measure:
        pp_groups = opts.phase_path_groups or []

        cmd = [
            measure_tool,
            'calculate',
            f'--linux={linux}',
            *(f"--{s.name.removeprefix('.')}={s.content}"
              for s in uki.sections
              if s.measure),
            *(f'--bank={bank}'
              for bank in banks),
            # For measurement, the keys are not relevant, so we can lump all the phase paths
            # into one call to systemd-measure calculate.
            *(f'--phase={phase_path}'
              for phase_path in itertools.chain.from_iterable(pp_groups)),
        ]

        print('+', shell_join(cmd))
        subprocess.check_call(cmd)

    # PCR signing

    if opts.pcr_private_keys:
        pcrsigs = []

        cmd = [
            measure_tool,
            'sign',
            f'--linux={linux}',
            *(f"--{s.name.removeprefix('.')}={s.content}"
              for s in uki.sections
              if s.measure),
            *(f'--bank={bank}'
              for bank in banks),
        ]

        for priv_key, pub_key, group in key_path_groups(opts):
            extra = [f'--private-key={priv_key}']
            if pub_key:
                extra += [f'--public-key={pub_key}']
            extra += [f'--phase={phase_path}' for phase_path in group or ()]

            print('+', shell_join(cmd + extra))
            pcrsig = subprocess.check_output(cmd + extra, text=True)
            pcrsig = json.loads(pcrsig)
            pcrsigs += [pcrsig]

        combined = combine_signatures(pcrsigs)
        uki.add_section(Section.create('.pcrsig', combined))


def join_initrds(initrds):
    if not initrds:
        return None
    if len(initrds) == 1:
        return initrds[0]

    seq = []
    for file in initrds:
        initrd = file.read_bytes()
        n = len(initrd)
        padding = b'\0' * (round_up(n, 4) - n)  # pad to 32 bit alignment
        seq += [initrd, padding]

    return b''.join(seq)


def pairwise(iterable):
    a, b = itertools.tee(iterable)
    next(b, None)
    return zip(a, b)


class PEError(Exception):
    pass


def pe_add_sections(uki: UKI, output: str):
    pe = pefile.PE(uki.executable, fast_load=True)

    # Old stubs do not have the symbol/string table stripped, even though image files should not have one.
    if symbol_table := pe.FILE_HEADER.PointerToSymbolTable:
        symbol_table_size = 18 * pe.FILE_HEADER.NumberOfSymbols
        if string_table_size := pe.get_dword_from_offset(symbol_table + symbol_table_size):
            symbol_table_size += string_table_size

        # Let's be safe and only strip it if it's at the end of the file.
        if symbol_table + symbol_table_size == len(pe.__data__):
            pe.__data__ = pe.__data__[:symbol_table]
            pe.FILE_HEADER.PointerToSymbolTable = 0
            pe.FILE_HEADER.NumberOfSymbols = 0
            pe.FILE_HEADER.IMAGE_FILE_LOCAL_SYMS_STRIPPED = True

    # Old stubs might have been stripped, leading to unaligned raw data values, so let's fix them up here.
    # pylint thinks that Structure doesn't have various members that it has…
    # pylint: disable=no-member

    for i, section in enumerate(pe.sections):
        oldp = section.PointerToRawData
        oldsz = section.SizeOfRawData
        section.PointerToRawData = round_up(oldp, pe.OPTIONAL_HEADER.FileAlignment)
        section.SizeOfRawData = round_up(oldsz, pe.OPTIONAL_HEADER.FileAlignment)
        padp = section.PointerToRawData - oldp
        padsz = section.SizeOfRawData - oldsz

        for later_section in pe.sections[i+1:]:
            later_section.PointerToRawData += padp + padsz

        pe.__data__ = pe.__data__[:oldp] + bytes(padp) + pe.__data__[oldp:oldp+oldsz] + bytes(padsz) + pe.__data__[oldp+oldsz:]

    # We might not have any space to add new sections. Let's try our best to make some space by padding the
    # SizeOfHeaders to a multiple of the file alignment. This is safe because the first section's data starts
    # at a multiple of the file alignment, so all space before that is unused.
    pe.OPTIONAL_HEADER.SizeOfHeaders = round_up(pe.OPTIONAL_HEADER.SizeOfHeaders, pe.OPTIONAL_HEADER.FileAlignment)
    pe = pefile.PE(data=pe.write(), fast_load=True)

    warnings = pe.get_warnings()
    if warnings:
        raise PEError(f'pefile warnings treated as errors: {warnings}')

    security = pe.OPTIONAL_HEADER.DATA_DIRECTORY[pefile.DIRECTORY_ENTRY['IMAGE_DIRECTORY_ENTRY_SECURITY']]
    if security.VirtualAddress != 0:
        # We could strip the signatures, but why would anyone sign the stub?
        raise PEError('Stub image is signed, refusing.')

    for section in uki.sections:
        new_section = pefile.SectionStructure(pe.__IMAGE_SECTION_HEADER_format__, pe=pe)
        new_section.__unpack__(b'\0' * new_section.sizeof())

        offset = pe.sections[-1].get_file_offset() + new_section.sizeof()
        if offset + new_section.sizeof() > pe.OPTIONAL_HEADER.SizeOfHeaders:
            raise PEError(f'Not enough header space to add section {section.name}.')

        assert section.content
        data = section.content.read_bytes()

        new_section.set_file_offset(offset)
        new_section.Name = section.name.encode()
        new_section.Misc_VirtualSize = len(data)
        # Non-stripped stubs might still have an unaligned symbol table at the end, making their size
        # unaligned, so we make sure to explicitly pad the pointer to new sections to an aligned offset.
        new_section.PointerToRawData = round_up(len(pe.__data__), pe.OPTIONAL_HEADER.FileAlignment)
        new_section.SizeOfRawData = round_up(len(data), pe.OPTIONAL_HEADER.FileAlignment)
        new_section.VirtualAddress = round_up(
            pe.sections[-1].VirtualAddress + pe.sections[-1].Misc_VirtualSize,
            pe.OPTIONAL_HEADER.SectionAlignment,
        )

        new_section.IMAGE_SCN_MEM_READ = True
        if section.name == '.linux':
            # Old kernels that use EFI handover protocol will be executed inline.
            new_section.IMAGE_SCN_CNT_CODE = True
        else:
            new_section.IMAGE_SCN_CNT_INITIALIZED_DATA = True

        # Special case, mostly for .sbat: the stub will already have a .sbat section, but we want to append
        # the one from the kernel to it. It should be small enough to fit in the existing section, so just
        # swap the data.
        for i, s in enumerate(pe.sections):
            if s.Name.rstrip(b"\x00").decode() == section.name:
                if new_section.Misc_VirtualSize > s.SizeOfRawData:
                    raise PEError(f'Not enough space in existing section {section.name} to append new data.')

                padding = bytes(new_section.SizeOfRawData - new_section.Misc_VirtualSize)
                pe.__data__ = pe.__data__[:s.PointerToRawData] + data + padding + pe.__data__[pe.sections[i+1].PointerToRawData:]
                s.SizeOfRawData = new_section.SizeOfRawData
                s.Misc_VirtualSize = new_section.Misc_VirtualSize
                break
        else:
            pe.__data__ = pe.__data__[:] + bytes(new_section.PointerToRawData - len(pe.__data__)) + data + bytes(new_section.SizeOfRawData - len(data))

            pe.FILE_HEADER.NumberOfSections += 1
            pe.OPTIONAL_HEADER.SizeOfInitializedData += new_section.Misc_VirtualSize
            pe.__structures__.append(new_section)
            pe.sections.append(new_section)

    pe.OPTIONAL_HEADER.CheckSum = 0
    pe.OPTIONAL_HEADER.SizeOfImage = round_up(
        pe.sections[-1].VirtualAddress + pe.sections[-1].Misc_VirtualSize,
        pe.OPTIONAL_HEADER.SectionAlignment,
    )

    pe.write(output)

def merge_sbat(input_pe: [pathlib.Path], input_text: [str]) -> str:
    sbat = []

    for f in input_pe:
        try:
            pe = pefile.PE(f, fast_load=True)
        except pefile.PEFormatError:
            print(f"{f} is not a valid PE file, not extracting SBAT section.")
            continue

        for section in pe.sections:
            if section.Name.rstrip(b"\x00").decode() == ".sbat":
                split = section.get_data().rstrip(b"\x00").decode().splitlines()
                if not split[0].startswith('sbat,'):
                    print(f"{f} does not contain a valid SBAT section, skipping.")
                    continue
                # Filter out the sbat line, we'll add it back later, there needs to be only one and it
                # needs to be first.
                sbat += split[1:]

    for t in input_text:
        if t.startswith('@'):
            t = pathlib.Path(t[1:]).read_text()
        split = t.splitlines()
        if not split[0].startswith('sbat,'):
            print(f"{t} does not contain a valid SBAT section, skipping.")
            continue
        sbat += split[1:]

    return 'sbat,1,SBAT Version,sbat,1,https://github.com/rhboot/shim/blob/main/SBAT.md\n' + '\n'.join(sbat) + "\n\x00"

def signer_sign(cmd):
    print('+', shell_join(cmd))
    subprocess.check_call(cmd)

def find_sbsign(opts=None):
    return find_tool('sbsign', opts=opts)

def sbsign_sign(sbsign_tool, input_f, output_f, opts=None):
    sign_invocation = [
        sbsign_tool,
        '--key', opts.sb_key,
        '--cert', opts.sb_cert,
        input_f,
        '--output', output_f,
    ]
    if opts.signing_engine is not None:
        sign_invocation += ['--engine', opts.signing_engine]
    signer_sign(sign_invocation)

def find_pesign(opts=None):
    return find_tool('pesign', opts=opts)

def pesign_sign(pesign_tool, input_f, output_f, opts=None):
    sign_invocation = [
        pesign_tool, '-s', '--force',
        '-n', opts.sb_certdir,
        '-c', opts.sb_cert_name,
        '-i', input_f,
        '-o', output_f,
    ]
    signer_sign(sign_invocation)

SBVERIFY = {
    'name': 'sbverify',
    'option': '--list',
    'output': 'No signature table present',
}

PESIGCHECK = {
    'name': 'pesign',
    'option': '-i',
    'output': 'No signatures found.',
    'flags': '-S'
}

def verify(tool, opts):
    verify_tool = find_tool(tool['name'], opts=opts)
    cmd = [
        verify_tool,
        tool['option'],
        opts.linux,
    ]
    if 'flags' in tool:
        cmd.append(tool['flags'])

    print('+', shell_join(cmd))
    info = subprocess.check_output(cmd, text=True)

    return tool['output'] in info

def make_uki(opts):
    # kernel payload signing

    sign_tool = None
    sign_args_present = opts.sb_key or opts.sb_cert_name
    sign_kernel = opts.sign_kernel
    sign = None
    linux = opts.linux

    if sign_args_present:
        if opts.signtool == 'sbsign':
            sign_tool = find_sbsign(opts=opts)
            sign = sbsign_sign
            verify_tool = SBVERIFY
        else:
            sign_tool = find_pesign(opts=opts)
            sign = pesign_sign
            verify_tool = PESIGCHECK

        if sign_tool is None:
            raise ValueError(f'{opts.signtool}, required for signing, is not installed')

        if sign_kernel is None and opts.linux is not None:
            # figure out if we should sign the kernel
            sign_kernel = verify(verify_tool, opts)

        if sign_kernel:
            linux_signed = tempfile.NamedTemporaryFile(prefix='linux-signed')
            linux = pathlib.Path(linux_signed.name)
            sign(sign_tool, opts.linux, linux, opts=opts)

    if opts.uname is None and opts.linux is not None:
        print('Kernel version not specified, starting autodetection 😖.')
        opts.uname = Uname.scrape(opts.linux, opts=opts)

    uki = UKI(opts.stub)
    initrd = join_initrds(opts.initrd)

    pcrpkey = opts.pcrpkey
    if pcrpkey is None:
        if opts.pcr_public_keys and len(opts.pcr_public_keys) == 1:
            pcrpkey = opts.pcr_public_keys[0]
        elif opts.pcr_private_keys and len(opts.pcr_private_keys) == 1:
            import cryptography.hazmat.primitives.serialization as serialization
            privkey = serialization.load_pem_private_key(opts.pcr_private_keys[0].read_bytes(), password=None)
            pcrpkey = privkey.public_key().public_bytes(
                encoding=serialization.Encoding.PEM,
                format=serialization.PublicFormat.SubjectPublicKeyInfo,
            )

    sections = [
        # name,      content,         measure?
        ('.osrel',   opts.os_release, True ),
        ('.cmdline', opts.cmdline,    True ),
        ('.dtb',     opts.devicetree, True ),
        ('.uname',   opts.uname,      True ),
        ('.splash',  opts.splash,     True ),
        ('.pcrpkey', pcrpkey,         True ),
        ('.initrd',  initrd,          True ),

        # linux shall be last to leave breathing room for decompression.
        # We'll add it later.
    ]

    for name, content, measure in sections:
        if content:
            uki.add_section(Section.create(name, content, measure=measure))

    # systemd-measure doesn't know about those extra sections
    for section in opts.sections:
        uki.add_section(section)

    if linux is not None:
        # Merge the .sbat sections from stub, kernel and parameter, so that revocation can be done on either.
        uki.add_section(Section.create('.sbat', merge_sbat([opts.stub, linux], opts.sbat), measure=True))
    else:
        # Addons don't use the stub so we add SBAT manually
        if not opts.sbat:
            opts.sbat = ["""sbat,1,SBAT Version,sbat,1,https://github.com/rhboot/shim/blob/main/SBAT.md
uki,1,UKI,uki,1,https://www.freedesktop.org/software/systemd/man/systemd-stub.html
"""]
        uki.add_section(Section.create('.sbat', merge_sbat([], opts.sbat), measure=False))

    # PCR measurement and signing

    # We pass in the contents for .linux separately because we need them to do the measurement but can't add
    # the section yet because we want .linux to be the last section. Make sure any other sections are added
    # before this function is called.
    call_systemd_measure(uki, linux, opts=opts)

    # UKI creation

    if linux is not None:
        uki.add_section(Section.create('.linux', linux, measure=True))

    if sign_args_present:
        unsigned = tempfile.NamedTemporaryFile(prefix='uki')
        unsigned_output = unsigned.name
    else:
        unsigned_output = opts.output

    pe_add_sections(uki, unsigned_output)

    # UKI signing

    if sign_args_present:
        assert sign
        sign(sign_tool, unsigned_output, opts.output, opts=opts)

        # We end up with no executable bits, let's reapply them
        os.umask(umask := os.umask(0))
        os.chmod(opts.output, 0o777 & ~umask)

    print(f"Wrote {'signed' if sign_args_present else 'unsigned'} {opts.output}")


ONE_DAY = datetime.timedelta(1, 0, 0)


@contextlib.contextmanager
def temporary_umask(mask: int):
    # Drop <mask> bits from umask
    old = os.umask(0)
    os.umask(old | mask)
    try:
        yield
    finally:
        os.umask(old)


def generate_key_cert_pair(
        common_name: str,
        valid_days: int,
        keylength: int = 2048,
) -> tuple[bytes]:

    from cryptography import x509
    import cryptography.hazmat.primitives as hp

    # We use a keylength of 2048 bits. That is what Microsoft documents as
    # supported/expected:
    # https://learn.microsoft.com/en-us/windows-hardware/manufacture/desktop/windows-secure-boot-key-creation-and-management-guidance?view=windows-11#12-public-key-cryptography

    now = datetime.datetime.utcnow()

    key = hp.asymmetric.rsa.generate_private_key(
        public_exponent=65537,
        key_size=keylength,
    )
    cert = x509.CertificateBuilder(
    ).subject_name(
        x509.Name([x509.NameAttribute(x509.oid.NameOID.COMMON_NAME, common_name)])
    ).issuer_name(
        x509.Name([x509.NameAttribute(x509.oid.NameOID.COMMON_NAME, common_name)])
    ).not_valid_before(
        now,
    ).not_valid_after(
        now + ONE_DAY * valid_days
    ).serial_number(
        x509.random_serial_number()
    ).public_key(
        key.public_key()
    ).add_extension(
        x509.BasicConstraints(ca=False, path_length=None),
        critical=True,
    ).sign(
        private_key=key,
        algorithm=hp.hashes.SHA256(),
    )

    cert_pem = cert.public_bytes(
        encoding=hp.serialization.Encoding.PEM,
    )
    key_pem = key.private_bytes(
        encoding=hp.serialization.Encoding.PEM,
        format=hp.serialization.PrivateFormat.TraditionalOpenSSL,
        encryption_algorithm=hp.serialization.NoEncryption(),
    )

    return key_pem, cert_pem


def generate_priv_pub_key_pair(keylength : int = 2048) -> tuple[bytes]:
    import cryptography.hazmat.primitives as hp

    key = hp.asymmetric.rsa.generate_private_key(
        public_exponent=65537,
        key_size=keylength,
    )
    priv_key_pem = key.private_bytes(
        encoding=hp.serialization.Encoding.PEM,
        format=hp.serialization.PrivateFormat.TraditionalOpenSSL,
        encryption_algorithm=hp.serialization.NoEncryption(),
    )
    pub_key_pem = key.public_key().public_bytes(
        encoding=hp.serialization.Encoding.PEM,
        format=hp.serialization.PublicFormat.SubjectPublicKeyInfo,
    )

    return priv_key_pem, pub_key_pem


def generate_keys(opts):
    # This will generate keys and certificates and write them to the paths that
    # are specified as input paths.
    if opts.sb_key or opts.sb_cert:
        fqdn = socket.getfqdn()
        cn = f'SecureBoot signing key on host {fqdn}'
        key_pem, cert_pem = generate_key_cert_pair(
            common_name=cn,
            valid_days=opts.sb_cert_validity,
        )
        print(f'Writing SecureBoot private key to {opts.sb_key}')
        with temporary_umask(0o077):
            opts.sb_key.write_bytes(key_pem)
        print(f'Writing SecureBoot certificate to {opts.sb_cert}')
        opts.sb_cert.write_bytes(cert_pem)

    for priv_key, pub_key, _ in key_path_groups(opts):
        priv_key_pem, pub_key_pem = generate_priv_pub_key_pair()

        print(f'Writing private key for PCR signing to {priv_key}')
        with temporary_umask(0o077):
            priv_key.write_bytes(priv_key_pem)
        if pub_key:
            print(f'Writing public key for PCR signing to {pub_key}')
            pub_key.write_bytes(pub_key_pem)


def inspect_section(opts, section):
    name = section.Name.rstrip(b"\x00").decode()

    # find the config for this section in opts and whether to show it
    config = opts.sections_by_name.get(name, None)
    show = (config or
            opts.all or
            (name in DEFAULT_SECTIONS_TO_SHOW and not opts.sections))
    if not show:
        return name, None

    ttype = config.output_mode if config else DEFAULT_SECTIONS_TO_SHOW.get(name, 'binary')

    data = section.get_data(ignore_padding=True)
    size = section.Misc_VirtualSize
    digest = sha256(data).hexdigest()

    struct = {
        'size' : size,
        'sha256' : digest,
    }

    if ttype == 'text':
        try:
            struct['text'] = data.decode()
        except UnicodeDecodeError as e:
            print(f"Section {name!r} is not valid text: {e}")
            struct['text'] = '(not valid UTF-8)'

    if config and config.content:
        assert isinstance(config.content, pathlib.Path)
        config.content.write_bytes(data)

    if opts.json == 'off':
        print(f"{name}:\n  size: {size} bytes\n  sha256: {digest}")
        if ttype == 'text':
            text = textwrap.indent(struct['text'].rstrip(), ' ' * 4)
            print(f"  text:\n{text}")

    return name, struct


def inspect_sections(opts):
    indent = 4 if opts.json == 'pretty' else None

    for file in opts.files:
        pe = pefile.PE(file, fast_load=True)
        gen = (inspect_section(opts, section) for section in pe.sections)
        descs = {key:val for (key, val) in gen if val}
        if opts.json != 'off':
            json.dump(descs, sys.stdout, indent=indent)


@dataclasses.dataclass(frozen=True)
class ConfigItem:
    @staticmethod
    def config_list_prepend(
            namespace: argparse.Namespace,
            group: Optional[str],
            dest: str,
            value: Any,
    ) -> None:
        "Prepend value to namespace.<dest>"

        assert not group

        old = getattr(namespace, dest, [])
        if old is None:
            old = []
        setattr(namespace, dest, value + old)

    @staticmethod
    def config_set_if_unset(
            namespace: argparse.Namespace,
            group: Optional[str],
            dest: str,
            value: Any,
    ) -> None:
        "Set namespace.<dest> to value only if it was None"

        assert not group

        if getattr(namespace, dest) is None:
            setattr(namespace, dest, value)

    @staticmethod
    def config_set_group(
            namespace: argparse.Namespace,
            group: Optional[str],
            dest: str,
            value: Any,
    ) -> None:
        "Set namespace.<dest>[idx] to value, with idx derived from group"

        # pylint: disable=protected-access
        if group not in namespace._groups:
            namespace._groups += [group]
        idx = namespace._groups.index(group)

        old = getattr(namespace, dest, None)
        if old is None:
            old = []
        setattr(namespace, dest,
                old + ([None] * (idx - len(old))) + [value])

    @staticmethod
    def parse_boolean(s: str) -> bool:
        "Parse 1/true/yes/y/t/on as true and 0/false/no/n/f/off/None as false"
        s_l = s.lower()
        if s_l in {'1', 'true', 'yes', 'y', 't', 'on'}:
            return True
        if s_l in {'0', 'false', 'no', 'n', 'f', 'off'}:
            return False
        raise ValueError('f"Invalid boolean literal: {s!r}')

    # arguments for argparse.ArgumentParser.add_argument()
    name: Union[str, tuple[str, str]]
    dest: Optional[str] = None
    metavar: Optional[str] = None
    type: Optional[Callable] = None
    nargs: Optional[str] = None
    action: Optional[Union[str, Callable]] = None
    default: Any = None
    version: Optional[str] = None
    choices: Optional[tuple[str, ...]] = None
    const: Optional[Any] = None
    help: Optional[str] = None

    # metadata for config file parsing
    config_key: Optional[str] = None
    config_push: Callable[[argparse.Namespace, Optional[str], str, Any], None] = \
                    config_set_if_unset

    def _names(self) -> tuple[str, ...]:
        return self.name if isinstance(self.name, tuple) else (self.name,)

    def argparse_dest(self) -> str:
        # It'd be nice if argparse exported this, but I don't see that in the API
        if self.dest:
            return self.dest
        return self._names()[0].lstrip('-').replace('-', '_')

    def add_to(self, parser: argparse.ArgumentParser):
        kwargs = { key:val
                   for key in dataclasses.asdict(self)
                   if (key not in ('name', 'config_key', 'config_push') and
                       (val := getattr(self, key)) is not None) }
        args = self._names()
        parser.add_argument(*args, **kwargs)

    def apply_config(self, namespace, section, group, key, value) -> None:
        assert f'{section}/{key}' == self.config_key
        dest = self.argparse_dest()

        conv: Callable[[str], Any]
        if self.action == argparse.BooleanOptionalAction:
            # We need to handle this case separately: the options are called
            # --foo and --no-foo, and no argument is parsed. But in the config
            # file, we have Foo=yes or Foo=no.
            conv = self.parse_boolean
        elif self.type:
            conv = self.type
        else:
            conv = lambda s:s

        # This is a bit ugly, but --initrd is the only option which is specified
        # with multiple args on the command line and a space-separated list in the
        # config file.
        if self.name == '--initrd':
            value = [conv(v) for v in value.split()]
        else:
            value = conv(value)

        self.config_push(namespace, group, dest, value)

    def config_example(self) -> tuple[Optional[str], Optional[str], Optional[str]]:
        if not self.config_key:
            return None, None, None
        section_name, key = self.config_key.split('/', 1)
        if section_name.endswith(':'):
            section_name += 'NAME'
        if self.choices:
            value = '|'.join(self.choices)
        else:
            value = self.metavar or self.argparse_dest().upper()
        return (section_name, key, value)


VERBS = ('build', 'genkey', 'inspect')

CONFIG_ITEMS = [
    ConfigItem(
        'positional',
        metavar = 'VERB',
        nargs = '*',
        help = f"operation to perform ({','.join(VERBS)})",
    ),

    ConfigItem(
        '--version',
        action = 'version',
        version = f'ukify {__version__}',
    ),

    ConfigItem(
        '--summary',
        help = 'print parsed config and exit',
        action = 'store_true',
    ),

    ConfigItem(
        '--linux',
        type = pathlib.Path,
        help = 'vmlinuz file [.linux section]',
        config_key = 'UKI/Linux',
    ),

    ConfigItem(
        '--initrd',
        metavar = 'INITRD',
        type = pathlib.Path,
        action = 'append',
        help = 'initrd file [part of .initrd section]',
        config_key = 'UKI/Initrd',
        config_push = ConfigItem.config_list_prepend,
    ),

    ConfigItem(
        ('--config', '-c'),
        metavar = 'PATH',
        help = 'configuration file',
    ),

    ConfigItem(
        '--cmdline',
        metavar = 'TEXT|@PATH',
        help = 'kernel command line [.cmdline section]',
        config_key = 'UKI/Cmdline',
    ),

    ConfigItem(
        '--os-release',
        metavar = 'TEXT|@PATH',
        help = 'path to os-release file [.osrel section]',
        config_key = 'UKI/OSRelease',
    ),

    ConfigItem(
        '--devicetree',
        metavar = 'PATH',
        type = pathlib.Path,
        help = 'Device Tree file [.dtb section]',
        config_key = 'UKI/DeviceTree',
    ),
    ConfigItem(
        '--splash',
        metavar = 'BMP',
        type = pathlib.Path,
        help = 'splash image bitmap file [.splash section]',
        config_key = 'UKI/Splash',
    ),
    ConfigItem(
        '--pcrpkey',
        metavar = 'KEY',
        type = pathlib.Path,
        help = 'embedded public key to seal secrets to [.pcrpkey section]',
        config_key = 'UKI/PCRPKey',
    ),
    ConfigItem(
        '--uname',
        metavar='VERSION',
        help='"uname -r" information [.uname section]',
        config_key = 'UKI/Uname',
    ),

    ConfigItem(
        '--efi-arch',
        metavar = 'ARCH',
        choices = ('ia32', 'x64', 'arm', 'aa64', 'riscv64'),
        help = 'target EFI architecture',
        config_key = 'UKI/EFIArch',
    ),

    ConfigItem(
        '--stub',
        type = pathlib.Path,
        help = 'path to the sd-stub file [.text,.data,… sections]',
        config_key = 'UKI/Stub',
    ),

    ConfigItem(
        '--sbat',
        metavar = 'TEXT|@PATH',
        help = 'SBAT policy [.sbat section]',
        default = [],
        action = 'append',
        config_key = 'UKI/SBAT',
    ),

    ConfigItem(
        '--section',
        dest = 'sections',
        metavar = 'NAME:TEXT|@PATH',
        action = 'append',
        default = [],
        help = 'section as name and contents [NAME section] or section to print',
    ),

    ConfigItem(
        '--pcr-banks',
        metavar = 'BANK…',
        type = parse_banks,
        config_key = 'UKI/PCRBanks',
    ),

    ConfigItem(
        '--signing-engine',
        metavar = 'ENGINE',
        help = 'OpenSSL engine to use for signing',
        config_key = 'UKI/SigningEngine',
    ),
    ConfigItem(
        '--signtool',
        choices = ('sbsign', 'pesign'),
        dest = 'signtool',
        default = 'sbsign',
        help = 'whether to use sbsign or pesign. Default is sbsign.',
        config_key = 'UKI/SecureBootSigningTool',
    ),
    ConfigItem(
        '--secureboot-private-key',
        dest = 'sb_key',
        help = 'required by --signtool=sbsign. Path to key file or engine-specific designation for SB signing',
        config_key = 'UKI/SecureBootPrivateKey',
    ),
    ConfigItem(
        '--secureboot-certificate',
        dest = 'sb_cert',
        help = 'required by --signtool=sbsign. sbsign needs a path to certificate file or engine-specific designation for SB signing',
        config_key = 'UKI/SecureBootCertificate',
    ),
    ConfigItem(
        '--secureboot-certificate-dir',
        dest = 'sb_certdir',
        default = '/etc/pki/pesign',
        help = 'required by --signtool=pesign. Path to nss certificate database directory for PE signing. Default is /etc/pki/pesign',
        config_key = 'UKI/SecureBootCertificateDir',
    ),
    ConfigItem(
        '--secureboot-certificate-name',
        dest = 'sb_cert_name',
        help = 'required by --signtool=pesign. pesign needs a certificate nickname of nss certificate database entry to use for PE signing',
        config_key = 'UKI/SecureBootCertificateName',
    ),
    ConfigItem(
        '--secureboot-certificate-validity',
        metavar = 'DAYS',
        dest = 'sb_cert_validity',
        default = 365 * 10,
        help = "period of validity (in days) for a certificate created by 'genkey'",
        config_key = 'UKI/SecureBootCertificateValidity',
    ),

    ConfigItem(
        '--sign-kernel',
        action = argparse.BooleanOptionalAction,
        help = 'Sign the embedded kernel',
        config_key = 'UKI/SignKernel',
    ),

    ConfigItem(
        '--pcr-private-key',
        dest = 'pcr_private_keys',
        metavar = 'PATH',
        type = pathlib.Path,
        action = 'append',
        help = 'private part of the keypair for signing PCR signatures',
        config_key = 'PCRSignature:/PCRPrivateKey',
        config_push = ConfigItem.config_set_group,
    ),
    ConfigItem(
        '--pcr-public-key',
        dest = 'pcr_public_keys',
        metavar = 'PATH',
        type = pathlib.Path,
        action = 'append',
        help = 'public part of the keypair for signing PCR signatures',
        config_key = 'PCRSignature:/PCRPublicKey',
        config_push = ConfigItem.config_set_group,
    ),
    ConfigItem(
        '--phases',
        dest = 'phase_path_groups',
        metavar = 'PHASE-PATH…',
        type = parse_phase_paths,
        action = 'append',
        help = 'phase-paths to create signatures for',
        config_key = 'PCRSignature:/Phases',
        config_push = ConfigItem.config_set_group,
    ),

    ConfigItem(
        '--tools',
        type = pathlib.Path,
        action = 'append',
        help = 'Directories to search for tools (systemd-measure, …)',
    ),

    ConfigItem(
        ('--output', '-o'),
        type = pathlib.Path,
        help = 'output file path',
    ),

    ConfigItem(
        '--measure',
        action = argparse.BooleanOptionalAction,
        help = 'print systemd-measure output for the UKI',
    ),

    ConfigItem(
        '--json',
        choices = ('pretty', 'short', 'off'),
        default = 'off',
        help = 'generate JSON output',
    ),
    ConfigItem(
        '-j',
        dest='json',
        action='store_const',
        const='pretty',
        help='equivalent to --json=pretty',
    ),

    ConfigItem(
        '--all',
        help = 'print all sections',
        action = 'store_true',
    ),
]

CONFIGFILE_ITEMS = { item.config_key:item
                     for item in CONFIG_ITEMS
                     if item.config_key }


def apply_config(namespace, filename=None):
    if filename is None:
        filename = namespace.config
    if filename is None:
        return

    # Fill in ._groups based on --pcr-public-key=, --pcr-private-key=, and --phases=.
    assert '_groups' not in namespace
    n_pcr_priv = len(namespace.pcr_private_keys or ())
    namespace._groups = list(range(n_pcr_priv))  # pylint: disable=protected-access

    cp = configparser.ConfigParser(
        comment_prefixes='#',
        inline_comment_prefixes='#',
        delimiters='=',
        empty_lines_in_values=False,
        interpolation=None,
        strict=False)
    # Do not make keys lowercase
    cp.optionxform = lambda option: option

    cp.read(filename)

    for section_name, section in cp.items():
        idx = section_name.find(':')
        if idx >= 0:
            section_name, group = section_name[:idx+1], section_name[idx+1:]
            if not section_name or not group:
                raise ValueError('Section name components cannot be empty')
            if ':' in group:
                raise ValueError('Section name cannot contain more than one ":"')
        else:
            group = None
        for key, value in section.items():
            if item := CONFIGFILE_ITEMS.get(f'{section_name}/{key}'):
                item.apply_config(namespace, section_name, group, key, value)
            else:
                print(f'Unknown config setting [{section_name}] {key}=')


def config_example():
    prev_section = None
    for item in CONFIG_ITEMS:
        section, key, value = item.config_example()
        if section:
            if prev_section != section:
                if prev_section:
                    yield ''
                yield f'[{section}]'
                prev_section = section
            yield f'{key} = {value}'


class PagerHelpAction(argparse._HelpAction):  # pylint: disable=protected-access
    def __call__(
        self,
        parser: argparse.ArgumentParser,
        namespace: argparse.Namespace,
        values: Union[str, Sequence[Any], None] = None,
        option_string: Optional[str] = None
    ) -> None:
        page(parser.format_help(), True)
        parser.exit()


def create_parser():
    p = argparse.ArgumentParser(
        description='Build and sign Unified Kernel Images',
        allow_abbrev=False,
        add_help=False,
        usage='''\
ukify [options…] VERB
''',
        epilog='\n  '.join(('config file:', *config_example())),
        formatter_class=argparse.RawDescriptionHelpFormatter,
    )

    for item in CONFIG_ITEMS:
        item.add_to(p)

    # Suppress printing of usage synopsis on errors
    p.error = lambda message: p.exit(2, f'{p.prog}: error: {message}\n')

    # Make --help paged
    p.add_argument(
        '-h', '--help',
        action=PagerHelpAction,
        help='show this help message and exit',
    )

    return p


def finalize_options(opts):
    # Figure out which syntax is being used, one of:
    # ukify verb --arg --arg --arg
    # ukify linux initrd…
    if len(opts.positional) >= 1 and opts.positional[0] == 'inspect':
        opts.verb = opts.positional[0]
        opts.files = opts.positional[1:]
        if not opts.files:
            raise ValueError('file(s) to inspect must be specified')
        if len(opts.files) > 1 and opts.json != 'off':
            # We could allow this in the future, but we need to figure out the right structure
            raise ValueError('JSON output is not allowed with multiple files')
    elif len(opts.positional) == 1 and opts.positional[0] in VERBS:
        opts.verb = opts.positional[0]
    elif opts.linux or opts.initrd:
        raise ValueError('--linux/--initrd options cannot be used with positional arguments')
    else:
        print("Assuming obsolete commandline syntax with no verb. Please use 'build'.")
        if opts.positional:
            opts.linux = pathlib.Path(opts.positional[0])
        # If we have initrds from parsing config files, append our positional args at the end
        opts.initrd = (opts.initrd or []) + [pathlib.Path(arg) for arg in opts.positional[1:]]
        opts.verb = 'build'

    # Check that --pcr-public-key=, --pcr-private-key=, and --phases=
    # have either the same number of arguments are are not specified at all.
    n_pcr_pub = None if opts.pcr_public_keys is None else len(opts.pcr_public_keys)
    n_pcr_priv = None if opts.pcr_private_keys is None else len(opts.pcr_private_keys)
    n_phase_path_groups = None if opts.phase_path_groups is None else len(opts.phase_path_groups)
    if n_pcr_pub is not None and n_pcr_pub != n_pcr_priv:
        raise ValueError('--pcr-public-key= specifications must match --pcr-private-key=')
    if n_phase_path_groups is not None and n_phase_path_groups != n_pcr_priv:
        raise ValueError('--phases= specifications must match --pcr-private-key=')

    if opts.cmdline and opts.cmdline.startswith('@'):
        opts.cmdline = pathlib.Path(opts.cmdline[1:])
    elif opts.cmdline:
        # Drop whitespace from the commandline. If we're reading from a file,
        # we copy the contents verbatim. But configuration specified on the commandline
        # or in the config file may contain additional whitespace that has no meaning.
        opts.cmdline = ' '.join(opts.cmdline.split())

    if opts.os_release and opts.os_release.startswith('@'):
        opts.os_release = pathlib.Path(opts.os_release[1:])
    elif not opts.os_release and opts.linux:
        p = pathlib.Path('/etc/os-release')
        if not p.exists():
            p = pathlib.Path('/usr/lib/os-release')
        opts.os_release = p

    if opts.efi_arch is None:
        opts.efi_arch = guess_efi_arch()

    if opts.stub is None:
        if opts.linux is not None:
            opts.stub = pathlib.Path(f'/usr/lib/systemd/boot/efi/linux{opts.efi_arch}.efi.stub')
        else:
            opts.stub = pathlib.Path(f'/usr/lib/systemd/boot/efi/addon{opts.efi_arch}.efi.stub')

    if opts.signing_engine is None:
        if opts.sb_key:
            opts.sb_key = pathlib.Path(opts.sb_key)
        if opts.sb_cert:
            opts.sb_cert = pathlib.Path(opts.sb_cert)

    if opts.signtool == 'sbsign':
        if bool(opts.sb_key) ^ bool(opts.sb_cert):
            raise ValueError('--secureboot-private-key= and --secureboot-certificate= must be specified together when using --signtool=sbsign')
    else:
        if not bool(opts.sb_cert_name):
            raise ValueError('--secureboot-certificate-name must be specified when using --signtool=pesign')

    if opts.sign_kernel and not opts.sb_key and not opts.sb_cert_name:
        raise ValueError('--sign-kernel requires either --secureboot-private-key= and --secureboot-certificate= (for sbsign) or --secureboot-certificate-name= (for pesign) to be specified')

    if opts.verb == 'build' and opts.output is None:
        if opts.linux is None:
            raise ValueError('--output= must be specified when building a PE addon')
        suffix = '.efi' if opts.sb_key or opts.sb_cert_name else '.unsigned.efi'
        opts.output = opts.linux.name + suffix

    # Now that we know if we're inputting or outputting, really parse section config
    f = Section.parse_output if opts.verb == 'inspect' else Section.parse_input
    opts.sections = [f(s) for s in opts.sections]
    # A convenience dictionary to make it easy to look up sections
    opts.sections_by_name = {s.name:s for s in opts.sections}

    if opts.summary:
        # TODO: replace pprint() with some fancy formatting.
        pprint.pprint(vars(opts))
        sys.exit()


def parse_args(args=None):
    opts = create_parser().parse_args(args)
    apply_config(opts)
    finalize_options(opts)
    return opts


def main():
    opts = parse_args()
    if opts.verb == 'build':
        check_inputs(opts)
        make_uki(opts)
    elif opts.verb == 'genkey':
        check_cert_and_keys_nonexistent(opts)
        generate_keys(opts)
    elif opts.verb == 'inspect':
        inspect_sections(opts)
    else:
        assert False


if __name__ == '__main__':
    main()