forked from aosp-mirror/platform_system_core
-
-
Notifications
You must be signed in to change notification settings - Fork 19
/
fs_mgr.cpp
2369 lines (2087 loc) · 89 KB
/
fs_mgr.cpp
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
/*
* Copyright (C) 2012 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "fs_mgr.h"
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <inttypes.h>
#include <libgen.h>
#include <selinux/selinux.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/ioctl.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/swap.h>
#include <sys/types.h>
#include <sys/utsname.h>
#include <sys/wait.h>
#include <time.h>
#include <unistd.h>
#include <array>
#include <chrono>
#include <functional>
#include <map>
#include <memory>
#include <string>
#include <string_view>
#include <thread>
#include <utility>
#include <vector>
#include <android-base/chrono_utils.h>
#include <android-base/file.h>
#include <android-base/properties.h>
#include <android-base/stringprintf.h>
#include <android-base/strings.h>
#include <android-base/unique_fd.h>
#include <cutils/android_filesystem_config.h>
#include <cutils/android_reboot.h>
#include <cutils/partition_utils.h>
#include <cutils/properties.h>
#include <ext4_utils/ext4.h>
#include <ext4_utils/ext4_sb.h>
#include <ext4_utils/ext4_utils.h>
#include <ext4_utils/wipe.h>
#include <fs_avb/fs_avb.h>
#include <fs_mgr/file_wait.h>
#include <fs_mgr_overlayfs.h>
#include <fscrypt/fscrypt.h>
#include <libdm/dm.h>
#include <libdm/loop_control.h>
#include <liblp/metadata_format.h>
#include <linux/fs.h>
#include <linux/loop.h>
#include <linux/magic.h>
#include <log/log_properties.h>
#include <logwrap/logwrap.h>
#include "blockdev.h"
#include "fs_mgr_priv.h"
#define E2FSCK_BIN "/system/bin/e2fsck"
#define F2FS_FSCK_BIN "/system/bin/fsck.f2fs"
#define MKSWAP_BIN "/system/bin/mkswap"
#define TUNE2FS_BIN "/system/bin/tune2fs"
#define RESIZE2FS_BIN "/system/bin/resize2fs"
#define FSCK_LOG_FILE "/dev/fscklogs/log"
#define ZRAM_CONF_DEV "/sys/block/zram0/disksize"
#define ZRAM_CONF_MCS "/sys/block/zram0/max_comp_streams"
#define ZRAM_BACK_DEV "/sys/block/zram0/backing_dev"
#define SYSFS_EXT4_VERITY "/sys/fs/ext4/features/verity"
#define SYSFS_EXT4_CASEFOLD "/sys/fs/ext4/features/casefold"
#define ARRAY_SIZE(a) (sizeof(a) / sizeof(*(a)))
using android::base::Basename;
using android::base::GetBoolProperty;
using android::base::GetUintProperty;
using android::base::Realpath;
using android::base::SetProperty;
using android::base::StartsWith;
using android::base::StringPrintf;
using android::base::Timer;
using android::base::unique_fd;
using android::dm::DeviceMapper;
using android::dm::DmDeviceState;
using android::dm::DmTargetLinear;
using android::dm::LoopControl;
// Realistically, this file should be part of the android::fs_mgr namespace;
using namespace android::fs_mgr;
using namespace std::literals;
// record fs stat
enum FsStatFlags {
FS_STAT_IS_EXT4 = 0x0001,
FS_STAT_NEW_IMAGE_VERSION = 0x0002,
FS_STAT_E2FSCK_F_ALWAYS = 0x0004,
FS_STAT_UNCLEAN_SHUTDOWN = 0x0008,
FS_STAT_QUOTA_ENABLED = 0x0010,
FS_STAT_RO_MOUNT_FAILED = 0x0040,
FS_STAT_RO_UNMOUNT_FAILED = 0x0080,
FS_STAT_FULL_MOUNT_FAILED = 0x0100,
FS_STAT_FSCK_FAILED = 0x0200,
FS_STAT_FSCK_FS_FIXED = 0x0400,
FS_STAT_INVALID_MAGIC = 0x0800,
FS_STAT_TOGGLE_QUOTAS_FAILED = 0x10000,
FS_STAT_SET_RESERVED_BLOCKS_FAILED = 0x20000,
FS_STAT_ENABLE_ENCRYPTION_FAILED = 0x40000,
FS_STAT_ENABLE_VERITY_FAILED = 0x80000,
FS_STAT_ENABLE_CASEFOLD_FAILED = 0x100000,
FS_STAT_ENABLE_METADATA_CSUM_FAILED = 0x200000,
};
static void log_fs_stat(const std::string& blk_device, int fs_stat) {
std::string msg =
android::base::StringPrintf("\nfs_stat,%s,0x%x\n", blk_device.c_str(), fs_stat);
android::base::unique_fd fd(TEMP_FAILURE_RETRY(open(FSCK_LOG_FILE, O_WRONLY | O_CLOEXEC |
O_APPEND | O_CREAT, 0664)));
if (fd == -1 || !android::base::WriteStringToFd(msg, fd)) {
LWARNING << __FUNCTION__ << "() cannot log " << msg;
}
}
static bool is_extfs(const std::string& fs_type) {
return fs_type == "ext4" || fs_type == "ext3" || fs_type == "ext2";
}
static bool is_f2fs(const std::string& fs_type) {
return fs_type == "f2fs";
}
static std::string realpath(const std::string& blk_device) {
std::string real_path;
if (!Realpath(blk_device, &real_path)) {
real_path = blk_device;
}
return real_path;
}
static bool should_force_check(int fs_stat) {
return fs_stat &
(FS_STAT_E2FSCK_F_ALWAYS | FS_STAT_UNCLEAN_SHUTDOWN | FS_STAT_QUOTA_ENABLED |
FS_STAT_RO_MOUNT_FAILED | FS_STAT_RO_UNMOUNT_FAILED | FS_STAT_FULL_MOUNT_FAILED |
FS_STAT_FSCK_FAILED | FS_STAT_TOGGLE_QUOTAS_FAILED |
FS_STAT_SET_RESERVED_BLOCKS_FAILED | FS_STAT_ENABLE_ENCRYPTION_FAILED);
}
static bool umount_retry(const std::string& mount_point) {
int retry_count = 5;
bool umounted = false;
while (retry_count-- > 0) {
umounted = umount(mount_point.c_str()) == 0;
if (umounted) {
LINFO << __FUNCTION__ << "(): unmount(" << mount_point << ") succeeded";
break;
}
PERROR << __FUNCTION__ << "(): umount(" << mount_point << ") failed";
if (retry_count) sleep(1);
}
return umounted;
}
static void check_fs(const std::string& blk_device, const std::string& fs_type,
const std::string& target, int* fs_stat) {
int status;
int ret;
long tmpmnt_flags = MS_NOATIME | MS_NOEXEC | MS_NOSUID;
auto tmpmnt_opts = "errors=remount-ro"s;
const char* e2fsck_argv[] = {E2FSCK_BIN, "-y", blk_device.c_str()};
const char* e2fsck_forced_argv[] = {E2FSCK_BIN, "-f", "-y", blk_device.c_str()};
if (*fs_stat & FS_STAT_INVALID_MAGIC) { // will fail, so do not try
return;
}
Timer t;
/* Check for the types of filesystems we know how to check */
if (is_extfs(fs_type)) {
/*
* First try to mount and unmount the filesystem. We do this because
* the kernel is more efficient than e2fsck in running the journal and
* processing orphaned inodes, and on at least one device with a
* performance issue in the emmc firmware, it can take e2fsck 2.5 minutes
* to do what the kernel does in about a second.
*
* After mounting and unmounting the filesystem, run e2fsck, and if an
* error is recorded in the filesystem superblock, e2fsck will do a full
* check. Otherwise, it does nothing. If the kernel cannot mount the
* filesytsem due to an error, e2fsck is still run to do a full check
* fix the filesystem.
*/
if (!(*fs_stat & FS_STAT_FULL_MOUNT_FAILED)) { // already tried if full mount failed
errno = 0;
if (fs_type == "ext4") {
// This option is only valid with ext4
tmpmnt_opts += ",nomblk_io_submit";
}
ret = mount(blk_device.c_str(), target.c_str(), fs_type.c_str(), tmpmnt_flags,
tmpmnt_opts.c_str());
PINFO << __FUNCTION__ << "(): mount(" << blk_device << "," << target << "," << fs_type
<< ")=" << ret;
if (ret) {
*fs_stat |= FS_STAT_RO_MOUNT_FAILED;
} else if (!umount_retry(target)) {
// boot may fail but continue and leave it to later stage for now.
PERROR << __FUNCTION__ << "(): umount(" << target << ") timed out";
*fs_stat |= FS_STAT_RO_UNMOUNT_FAILED;
}
}
/*
* Some system images do not have e2fsck for licensing reasons
* (e.g. recent SDK system images). Detect these and skip the check.
*/
if (access(E2FSCK_BIN, X_OK)) {
LINFO << "Not running " << E2FSCK_BIN << " on " << realpath(blk_device)
<< " (executable not in system image)";
} else {
LINFO << "Running " << E2FSCK_BIN << " on " << realpath(blk_device);
if (should_force_check(*fs_stat)) {
ret = logwrap_fork_execvp(ARRAY_SIZE(e2fsck_forced_argv), e2fsck_forced_argv,
&status, false, LOG_KLOG | LOG_FILE, false,
FSCK_LOG_FILE);
} else {
ret = logwrap_fork_execvp(ARRAY_SIZE(e2fsck_argv), e2fsck_argv, &status, false,
LOG_KLOG | LOG_FILE, false, FSCK_LOG_FILE);
}
if (ret < 0) {
/* No need to check for error in fork, we can't really handle it now */
LERROR << "Failed trying to run " << E2FSCK_BIN;
*fs_stat |= FS_STAT_FSCK_FAILED;
} else if (status != 0) {
LINFO << "e2fsck returned status 0x" << std::hex << status;
*fs_stat |= FS_STAT_FSCK_FS_FIXED;
}
}
} else if (is_f2fs(fs_type)) {
const char* f2fs_fsck_argv[] = {F2FS_FSCK_BIN, "-a", "-c", "10000", "--debug-cache",
blk_device.c_str()};
const char* f2fs_fsck_forced_argv[] = {
F2FS_FSCK_BIN, "-f", "-c", "10000", "--debug-cache", blk_device.c_str()};
if (access(F2FS_FSCK_BIN, X_OK)) {
LINFO << "Not running " << F2FS_FSCK_BIN << " on " << realpath(blk_device)
<< " (executable not in system image)";
} else {
if (should_force_check(*fs_stat)) {
LINFO << "Running " << F2FS_FSCK_BIN << " -f -c 10000 --debug-cache "
<< realpath(blk_device);
ret = logwrap_fork_execvp(ARRAY_SIZE(f2fs_fsck_forced_argv), f2fs_fsck_forced_argv,
&status, false, LOG_KLOG | LOG_FILE, false,
FSCK_LOG_FILE);
} else {
LINFO << "Running " << F2FS_FSCK_BIN << " -a -c 10000 --debug-cache "
<< realpath(blk_device);
ret = logwrap_fork_execvp(ARRAY_SIZE(f2fs_fsck_argv), f2fs_fsck_argv, &status,
false, LOG_KLOG | LOG_FILE, false, FSCK_LOG_FILE);
}
if (ret < 0) {
/* No need to check for error in fork, we can't really handle it now */
LERROR << "Failed trying to run " << F2FS_FSCK_BIN;
*fs_stat |= FS_STAT_FSCK_FAILED;
} else if (status != 0) {
LINFO << F2FS_FSCK_BIN << " returned status 0x" << std::hex << status;
*fs_stat |= FS_STAT_FSCK_FS_FIXED;
}
}
}
android::base::SetProperty("ro.boottime.init.fsck." + Basename(target),
std::to_string(t.duration().count()));
return;
}
static ext4_fsblk_t ext4_blocks_count(const struct ext4_super_block* es) {
return ((ext4_fsblk_t)le32_to_cpu(es->s_blocks_count_hi) << 32) |
le32_to_cpu(es->s_blocks_count_lo);
}
static ext4_fsblk_t ext4_r_blocks_count(const struct ext4_super_block* es) {
return ((ext4_fsblk_t)le32_to_cpu(es->s_r_blocks_count_hi) << 32) |
le32_to_cpu(es->s_r_blocks_count_lo);
}
static bool is_ext4_superblock_valid(const struct ext4_super_block* es) {
if (es->s_magic != EXT4_SUPER_MAGIC) return false;
if (es->s_rev_level != EXT4_DYNAMIC_REV && es->s_rev_level != EXT4_GOOD_OLD_REV) return false;
if (EXT4_INODES_PER_GROUP(es) == 0) return false;
return true;
}
// Read the primary superblock from an ext4 filesystem. On failure return
// false. If it's not an ext4 filesystem, also set FS_STAT_INVALID_MAGIC.
static bool read_ext4_superblock(const std::string& blk_device, struct ext4_super_block* sb,
int* fs_stat) {
android::base::unique_fd fd(TEMP_FAILURE_RETRY(open(blk_device.c_str(), O_RDONLY | O_CLOEXEC)));
if (fd < 0) {
PERROR << "Failed to open '" << blk_device << "'";
return false;
}
if (TEMP_FAILURE_RETRY(pread(fd, sb, sizeof(*sb), 1024)) != sizeof(*sb)) {
PERROR << "Can't read '" << blk_device << "' superblock";
return false;
}
if (!is_ext4_superblock_valid(sb)) {
LINFO << "Invalid ext4 superblock on '" << blk_device << "'";
// not a valid fs, tune2fs, fsck, and mount will all fail.
*fs_stat |= FS_STAT_INVALID_MAGIC;
return false;
}
*fs_stat |= FS_STAT_IS_EXT4;
LINFO << "superblock s_max_mnt_count:" << sb->s_max_mnt_count << "," << blk_device;
if (sb->s_max_mnt_count == 0xffff) { // -1 (int16) in ext2, but uint16 in ext4
*fs_stat |= FS_STAT_NEW_IMAGE_VERSION;
}
return true;
}
// exported silent version of the above that just answer the question is_ext4
bool fs_mgr_is_ext4(const std::string& blk_device) {
android::base::ErrnoRestorer restore;
android::base::unique_fd fd(TEMP_FAILURE_RETRY(open(blk_device.c_str(), O_RDONLY | O_CLOEXEC)));
if (fd < 0) return false;
ext4_super_block sb;
if (TEMP_FAILURE_RETRY(pread(fd, &sb, sizeof(sb), 1024)) != sizeof(sb)) return false;
if (!is_ext4_superblock_valid(&sb)) return false;
return true;
}
// Some system images do not have tune2fs for licensing reasons.
// Detect these and skip running it.
static bool tune2fs_available(void) {
return access(TUNE2FS_BIN, X_OK) == 0;
}
static bool run_command(const char* argv[], int argc) {
int ret;
ret = logwrap_fork_execvp(argc, argv, nullptr, false, LOG_KLOG, false, nullptr);
return ret == 0;
}
// Enable/disable quota support on the filesystem if needed.
static void tune_quota(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
bool has_quota = (sb->s_feature_ro_compat & cpu_to_le32(EXT4_FEATURE_RO_COMPAT_QUOTA)) != 0;
bool want_quota = entry.fs_mgr_flags.quota;
// Enable projid support by default
bool want_projid = true;
if (has_quota == want_quota) {
return;
}
if (!tune2fs_available()) {
LERROR << "Unable to " << (want_quota ? "enable" : "disable") << " quotas on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
const char* argv[] = {TUNE2FS_BIN, nullptr, nullptr, blk_device.c_str()};
if (want_quota) {
LINFO << "Enabling quotas on " << blk_device;
argv[1] = "-Oquota";
// Once usr/grp unneeded, make just prjquota to save overhead
if (want_projid)
argv[2] = "-Qusrquota,grpquota,prjquota";
else
argv[2] = "-Qusrquota,grpquota";
*fs_stat |= FS_STAT_QUOTA_ENABLED;
} else {
LINFO << "Disabling quotas on " << blk_device;
argv[1] = "-O^quota";
argv[2] = "-Q^usrquota,^grpquota,^prjquota";
}
if (!run_command(argv, ARRAY_SIZE(argv))) {
LERROR << "Failed to run " TUNE2FS_BIN " to " << (want_quota ? "enable" : "disable")
<< " quotas on " << blk_device;
*fs_stat |= FS_STAT_TOGGLE_QUOTAS_FAILED;
}
}
// Set the number of reserved filesystem blocks if needed.
static void tune_reserved_size(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
if (entry.reserved_size == 0) {
return;
}
// The size to reserve is given in the fstab, but we won't reserve more
// than 2% of the filesystem.
const uint64_t max_reserved_blocks = ext4_blocks_count(sb) * 0.02;
uint64_t reserved_blocks = entry.reserved_size / EXT4_BLOCK_SIZE(sb);
if (reserved_blocks > max_reserved_blocks) {
LWARNING << "Reserved blocks " << reserved_blocks << " is too large; "
<< "capping to " << max_reserved_blocks;
reserved_blocks = max_reserved_blocks;
}
if ((ext4_r_blocks_count(sb) == reserved_blocks) && (sb->s_def_resgid == AID_RESERVED_DISK)) {
return;
}
if (!tune2fs_available()) {
LERROR << "Unable to set the number of reserved blocks on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
LINFO << "Setting reserved block count on " << blk_device << " to " << reserved_blocks;
auto reserved_blocks_str = std::to_string(reserved_blocks);
auto reserved_gid_str = std::to_string(AID_RESERVED_DISK);
const char* argv[] = {
TUNE2FS_BIN, "-r", reserved_blocks_str.c_str(), "-g", reserved_gid_str.c_str(),
blk_device.c_str()};
if (!run_command(argv, ARRAY_SIZE(argv))) {
LERROR << "Failed to run " TUNE2FS_BIN " to set the number of reserved blocks on "
<< blk_device;
*fs_stat |= FS_STAT_SET_RESERVED_BLOCKS_FAILED;
}
}
// Enable file-based encryption if needed.
static void tune_encrypt(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
if (!entry.fs_mgr_flags.file_encryption) {
return; // Nothing needs done.
}
std::vector<std::string> features_needed;
if ((sb->s_feature_incompat & cpu_to_le32(EXT4_FEATURE_INCOMPAT_ENCRYPT)) == 0) {
features_needed.emplace_back("encrypt");
}
android::fscrypt::EncryptionOptions options;
if (!android::fscrypt::ParseOptions(entry.encryption_options, &options)) {
LERROR << "Unable to parse encryption options on " << blk_device << ": "
<< entry.encryption_options;
return;
}
if ((options.flags &
(FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64 | FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32)) != 0) {
// We can only use this policy on ext4 if the "stable_inodes" feature
// is set on the filesystem, otherwise shrinking will break encrypted files.
if ((sb->s_feature_compat & cpu_to_le32(EXT4_FEATURE_COMPAT_STABLE_INODES)) == 0) {
features_needed.emplace_back("stable_inodes");
}
}
if (features_needed.size() == 0) {
return;
}
if (!tune2fs_available()) {
LERROR << "Unable to enable ext4 encryption on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
auto flags = android::base::Join(features_needed, ',');
auto flag_arg = "-O"s + flags;
const char* argv[] = {TUNE2FS_BIN, flag_arg.c_str(), blk_device.c_str()};
LINFO << "Enabling ext4 flags " << flags << " on " << blk_device;
if (!run_command(argv, ARRAY_SIZE(argv))) {
LERROR << "Failed to run " TUNE2FS_BIN " to enable "
<< "ext4 flags " << flags << " on " << blk_device;
*fs_stat |= FS_STAT_ENABLE_ENCRYPTION_FAILED;
}
}
// Enable fs-verity if needed.
static void tune_verity(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
bool has_verity = (sb->s_feature_ro_compat & cpu_to_le32(EXT4_FEATURE_RO_COMPAT_VERITY)) != 0;
bool want_verity = entry.fs_mgr_flags.fs_verity;
if (has_verity || !want_verity) {
return;
}
std::string verity_support;
if (!android::base::ReadFileToString(SYSFS_EXT4_VERITY, &verity_support)) {
LERROR << "Failed to open " << SYSFS_EXT4_VERITY;
return;
}
if (!(android::base::Trim(verity_support) == "supported")) {
LERROR << "Current ext4 verity not supported by kernel";
return;
}
if (!tune2fs_available()) {
LERROR << "Unable to enable ext4 verity on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
LINFO << "Enabling ext4 verity on " << blk_device;
const char* argv[] = {TUNE2FS_BIN, "-O", "verity", blk_device.c_str()};
if (!run_command(argv, ARRAY_SIZE(argv))) {
LERROR << "Failed to run " TUNE2FS_BIN " to enable "
<< "ext4 verity on " << blk_device;
*fs_stat |= FS_STAT_ENABLE_VERITY_FAILED;
}
}
// Enable casefold if needed.
static void tune_casefold(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
bool has_casefold = (sb->s_feature_incompat & cpu_to_le32(EXT4_FEATURE_INCOMPAT_CASEFOLD)) != 0;
bool wants_casefold =
android::base::GetBoolProperty("external_storage.casefold.enabled", false);
if (entry.mount_point != "/data" || !wants_casefold || has_casefold) return;
std::string casefold_support;
if (!android::base::ReadFileToString(SYSFS_EXT4_CASEFOLD, &casefold_support)) {
LERROR << "Failed to open " << SYSFS_EXT4_CASEFOLD;
return;
}
if (!(android::base::Trim(casefold_support) == "supported")) {
LERROR << "Current ext4 casefolding not supported by kernel";
return;
}
if (!tune2fs_available()) {
LERROR << "Unable to enable ext4 casefold on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
LINFO << "Enabling ext4 casefold on " << blk_device;
const char* argv[] = {TUNE2FS_BIN, "-O", "casefold", "-E", "encoding=utf8", blk_device.c_str()};
if (!run_command(argv, ARRAY_SIZE(argv))) {
LERROR << "Failed to run " TUNE2FS_BIN " to enable "
<< "ext4 casefold on " << blk_device;
*fs_stat |= FS_STAT_ENABLE_CASEFOLD_FAILED;
}
}
static bool resize2fs_available(void) {
return access(RESIZE2FS_BIN, X_OK) == 0;
}
// Enable metadata_csum
static void tune_metadata_csum(const std::string& blk_device, const FstabEntry& entry,
const struct ext4_super_block* sb, int* fs_stat) {
bool has_meta_csum =
(sb->s_feature_ro_compat & cpu_to_le32(EXT4_FEATURE_RO_COMPAT_METADATA_CSUM)) != 0;
bool want_meta_csum = entry.fs_mgr_flags.ext_meta_csum;
if (has_meta_csum || !want_meta_csum) return;
if (!tune2fs_available()) {
LERROR << "Unable to enable metadata_csum on " << blk_device
<< " because " TUNE2FS_BIN " is missing";
return;
}
if (!resize2fs_available()) {
LERROR << "Unable to enable metadata_csum on " << blk_device
<< " because " RESIZE2FS_BIN " is missing";
return;
}
LINFO << "Enabling ext4 metadata_csum on " << blk_device;
// Must give `-T now` to prevent last_fsck_time from growing too large,
// otherwise, tune2fs won't enable metadata_csum.
const char* tune2fs_args[] = {TUNE2FS_BIN, "-O", "metadata_csum,64bit,extent",
"-T", "now", blk_device.c_str()};
const char* resize2fs_args[] = {RESIZE2FS_BIN, "-b", blk_device.c_str()};
if (!run_command(tune2fs_args, ARRAY_SIZE(tune2fs_args))) {
LERROR << "Failed to run " TUNE2FS_BIN " to enable "
<< "ext4 metadata_csum on " << blk_device;
*fs_stat |= FS_STAT_ENABLE_METADATA_CSUM_FAILED;
} else if (!run_command(resize2fs_args, ARRAY_SIZE(resize2fs_args))) {
LERROR << "Failed to run " RESIZE2FS_BIN " to enable "
<< "ext4 metadata_csum on " << blk_device;
*fs_stat |= FS_STAT_ENABLE_METADATA_CSUM_FAILED;
}
}
// Read the primary superblock from an f2fs filesystem. On failure return
// false. If it's not an f2fs filesystem, also set FS_STAT_INVALID_MAGIC.
#define F2FS_SUPER_OFFSET 1024
static bool read_f2fs_superblock(const std::string& blk_device, int* fs_stat) {
android::base::unique_fd fd(TEMP_FAILURE_RETRY(open(blk_device.c_str(), O_RDONLY | O_CLOEXEC)));
__le32 sb1, sb2;
if (fd < 0) {
PERROR << "Failed to open '" << blk_device << "'";
return false;
}
if (TEMP_FAILURE_RETRY(pread(fd, &sb1, sizeof(sb1), F2FS_SUPER_OFFSET)) != sizeof(sb1)) {
PERROR << "Can't read '" << blk_device << "' superblock1";
return false;
}
// F2FS only supports block_size=page_size case. So, it is safe to call
// `getpagesize()` and use that as size of super block.
if (TEMP_FAILURE_RETRY(pread(fd, &sb2, sizeof(sb2), getpagesize() + F2FS_SUPER_OFFSET)) !=
sizeof(sb2)) {
PERROR << "Can't read '" << blk_device << "' superblock2";
return false;
}
if (sb1 != cpu_to_le32(F2FS_SUPER_MAGIC) && sb2 != cpu_to_le32(F2FS_SUPER_MAGIC)) {
LINFO << "Invalid f2fs superblock on '" << blk_device << "'";
*fs_stat |= FS_STAT_INVALID_MAGIC;
return false;
}
return true;
}
// exported silent version of the above that just answer the question is_f2fs
bool fs_mgr_is_f2fs(const std::string& blk_device) {
android::base::ErrnoRestorer restore;
android::base::unique_fd fd(TEMP_FAILURE_RETRY(open(blk_device.c_str(), O_RDONLY | O_CLOEXEC)));
if (fd < 0) return false;
__le32 sb;
if (TEMP_FAILURE_RETRY(pread(fd, &sb, sizeof(sb), F2FS_SUPER_OFFSET)) != sizeof(sb)) {
return false;
}
if (sb == cpu_to_le32(F2FS_SUPER_MAGIC)) return true;
if (TEMP_FAILURE_RETRY(pread(fd, &sb, sizeof(sb), getpagesize() + F2FS_SUPER_OFFSET)) !=
sizeof(sb)) {
return false;
}
return sb == cpu_to_le32(F2FS_SUPER_MAGIC);
}
static void SetReadAheadSize(const std::string& entry_block_device, off64_t size_kb) {
std::string block_device;
if (!Realpath(entry_block_device, &block_device)) {
PERROR << "Failed to realpath " << entry_block_device;
return;
}
static constexpr std::string_view kDevBlockPrefix("/dev/block/");
if (!android::base::StartsWith(block_device, kDevBlockPrefix)) {
LWARNING << block_device << " is not a block device";
return;
}
DeviceMapper& dm = DeviceMapper::Instance();
while (true) {
std::string block_name = block_device;
if (android::base::StartsWith(block_device, kDevBlockPrefix)) {
block_name = block_device.substr(kDevBlockPrefix.length());
}
std::string sys_partition =
android::base::StringPrintf("/sys/class/block/%s/partition", block_name.c_str());
struct stat info;
if (lstat(sys_partition.c_str(), &info) == 0) {
// it has a partition like "sda12".
block_name += "/..";
}
std::string sys_ra = android::base::StringPrintf("/sys/class/block/%s/queue/read_ahead_kb",
block_name.c_str());
std::string size = android::base::StringPrintf("%llu", (long long)size_kb);
android::base::WriteStringToFile(size, sys_ra.c_str());
LINFO << "Set readahead_kb: " << size << " on " << sys_ra;
auto parent = dm.GetParentBlockDeviceByPath(block_device);
if (!parent) {
return;
}
block_device = *parent;
}
}
//
// Prepare the filesystem on the given block device to be mounted.
//
// If the "check" option was given in the fstab record, or it seems that the
// filesystem was uncleanly shut down, we'll run fsck on the filesystem.
//
// If needed, we'll also enable (or disable) filesystem features as specified by
// the fstab record.
//
static int prepare_fs_for_mount(const std::string& blk_device, const FstabEntry& entry,
const std::string& alt_mount_point = "") {
auto& mount_point = alt_mount_point.empty() ? entry.mount_point : alt_mount_point;
// We need this because sometimes we have legacy symlinks that are
// lingering around and need cleaning up.
struct stat info;
if (lstat(mount_point.c_str(), &info) == 0 && (info.st_mode & S_IFMT) == S_IFLNK) {
unlink(mount_point.c_str());
}
mkdir(mount_point.c_str(), 0755);
// Don't need to return error, since it's a salt
if (entry.readahead_size_kb != -1) {
SetReadAheadSize(blk_device, entry.readahead_size_kb);
}
int fs_stat = 0;
if (is_extfs(entry.fs_type)) {
struct ext4_super_block sb;
if (read_ext4_superblock(blk_device, &sb, &fs_stat)) {
if ((sb.s_feature_incompat & EXT4_FEATURE_INCOMPAT_RECOVER) != 0 ||
(sb.s_state & EXT4_VALID_FS) == 0) {
LINFO << "Filesystem on " << blk_device << " was not cleanly shutdown; "
<< "state flags: 0x" << std::hex << sb.s_state << ", "
<< "incompat feature flags: 0x" << std::hex << sb.s_feature_incompat;
fs_stat |= FS_STAT_UNCLEAN_SHUTDOWN;
}
// Note: quotas should be enabled before running fsck.
tune_quota(blk_device, entry, &sb, &fs_stat);
} else {
return fs_stat;
}
} else if (is_f2fs(entry.fs_type)) {
if (!read_f2fs_superblock(blk_device, &fs_stat)) {
return fs_stat;
}
}
if (entry.fs_mgr_flags.check ||
(fs_stat & (FS_STAT_UNCLEAN_SHUTDOWN | FS_STAT_QUOTA_ENABLED))) {
check_fs(blk_device, entry.fs_type, mount_point, &fs_stat);
}
if (is_extfs(entry.fs_type) &&
(entry.reserved_size != 0 || entry.fs_mgr_flags.file_encryption ||
entry.fs_mgr_flags.fs_verity || entry.fs_mgr_flags.ext_meta_csum)) {
struct ext4_super_block sb;
if (read_ext4_superblock(blk_device, &sb, &fs_stat)) {
tune_reserved_size(blk_device, entry, &sb, &fs_stat);
tune_encrypt(blk_device, entry, &sb, &fs_stat);
tune_verity(blk_device, entry, &sb, &fs_stat);
tune_casefold(blk_device, entry, &sb, &fs_stat);
tune_metadata_csum(blk_device, entry, &sb, &fs_stat);
}
}
return fs_stat;
}
// Mark the given block device as read-only, using the BLKROSET ioctl.
bool fs_mgr_set_blk_ro(const std::string& blockdev, bool readonly) {
unique_fd fd(TEMP_FAILURE_RETRY(open(blockdev.c_str(), O_RDONLY | O_CLOEXEC)));
if (fd < 0) {
return false;
}
int ON = readonly;
return ioctl(fd, BLKROSET, &ON) == 0;
}
// Orange state means the device is unlocked, see the following link for details.
// https://source.android.com/security/verifiedboot/verified-boot#device_state
bool fs_mgr_is_device_unlocked() {
std::string verified_boot_state;
if (fs_mgr_get_boot_config("verifiedbootstate", &verified_boot_state)) {
return verified_boot_state == "orange";
}
return false;
}
// __mount(): wrapper around the mount() system call which also
// sets the underlying block device to read-only if the mount is read-only.
// See "man 2 mount" for return values.
static int __mount(const std::string& source, const std::string& target, const FstabEntry& entry) {
errno = 0;
unsigned long mountflags = entry.flags;
int ret = 0;
int save_errno = 0;
int gc_allowance = 0;
std::string opts;
std::string checkpoint_opts;
bool try_f2fs_gc_allowance = is_f2fs(entry.fs_type) && entry.fs_checkpoint_opts.length() > 0;
bool try_f2fs_fallback = false;
Timer t;
do {
if (save_errno == EINVAL && (try_f2fs_gc_allowance || try_f2fs_fallback)) {
PINFO << "Kernel does not support " << checkpoint_opts << ", trying without.";
try_f2fs_gc_allowance = false;
// Attempt without gc allowance before dropping.
try_f2fs_fallback = !try_f2fs_fallback;
}
if (try_f2fs_gc_allowance) {
checkpoint_opts = entry.fs_checkpoint_opts + ":" + std::to_string(gc_allowance) + "%";
} else if (try_f2fs_fallback) {
checkpoint_opts = entry.fs_checkpoint_opts;
} else {
checkpoint_opts = "";
}
opts = entry.fs_options + checkpoint_opts;
if (save_errno == EAGAIN) {
PINFO << "Retrying mount (source=" << source << ",target=" << target
<< ",type=" << entry.fs_type << ", gc_allowance=" << gc_allowance << "%)=" << ret
<< "(" << save_errno << ")";
}
ret = mount(source.c_str(), target.c_str(), entry.fs_type.c_str(), mountflags,
opts.c_str());
save_errno = errno;
if (try_f2fs_gc_allowance) gc_allowance += 10;
} while ((ret && save_errno == EAGAIN && gc_allowance <= 100) ||
(ret && save_errno == EINVAL && (try_f2fs_gc_allowance || try_f2fs_fallback)));
const char* target_missing = "";
const char* source_missing = "";
if (save_errno == ENOENT) {
if (access(target.c_str(), F_OK)) {
target_missing = "(missing)";
} else if (access(source.c_str(), F_OK)) {
source_missing = "(missing)";
}
errno = save_errno;
}
PINFO << __FUNCTION__ << "(source=" << source << source_missing << ",target=" << target
<< target_missing << ",type=" << entry.fs_type << ")=" << ret;
if ((ret == 0) && (mountflags & MS_RDONLY) != 0) {
fs_mgr_set_blk_ro(source);
}
if (ret == 0) {
android::base::SetProperty("ro.boottime.init.mount." + Basename(target),
std::to_string(t.duration().count()));
}
errno = save_errno;
return ret;
}
static bool fs_match(const std::string& in1, const std::string& in2) {
if (in1.empty() || in2.empty()) {
return false;
}
auto in1_end = in1.size() - 1;
while (in1_end > 0 && in1[in1_end] == '/') {
in1_end--;
}
auto in2_end = in2.size() - 1;
while (in2_end > 0 && in2[in2_end] == '/') {
in2_end--;
}
if (in1_end != in2_end) {
return false;
}
for (size_t i = 0; i <= in1_end; ++i) {
if (in1[i] != in2[i]) {
return false;
}
}
return true;
}
// Tries to mount any of the consecutive fstab entries that match
// the mountpoint of the one given by fstab[start_idx].
//
// end_idx: On return, will be the last entry that was looked at.
// attempted_idx: On return, will indicate which fstab entry
// succeeded. In case of failure, it will be the start_idx.
// Sets errno to match the 1st mount failure on failure.
static bool mount_with_alternatives(Fstab& fstab, int start_idx, int* end_idx,
int* attempted_idx) {
unsigned long i;
int mount_errno = 0;
bool mounted = false;
// Hunt down an fstab entry for the same mount point that might succeed.
for (i = start_idx;
// We required that fstab entries for the same mountpoint be consecutive.
i < fstab.size() && fstab[start_idx].mount_point == fstab[i].mount_point; i++) {
// Don't try to mount/encrypt the same mount point again.
// Deal with alternate entries for the same point which are required to be all following
// each other.
if (mounted) {
LINFO << __FUNCTION__ << "(): skipping fstab dup mountpoint=" << fstab[i].mount_point
<< " rec[" << i << "].fs_type=" << fstab[i].fs_type << " already mounted as "
<< fstab[*attempted_idx].fs_type;
continue;
}
// fstab[start_idx].blk_device is already updated to /dev/dm-<N> by
// AVB related functions. Copy it from start_idx to the current index i.
if ((i != start_idx) && fstab[i].fs_mgr_flags.logical &&
fstab[start_idx].fs_mgr_flags.logical &&
(fstab[i].logical_partition_name == fstab[start_idx].logical_partition_name)) {
fstab[i].blk_device = fstab[start_idx].blk_device;
}
int fs_stat = prepare_fs_for_mount(fstab[i].blk_device, fstab[i]);
if (fs_stat & FS_STAT_INVALID_MAGIC) {
LERROR << __FUNCTION__
<< "(): skipping mount due to invalid magic, mountpoint=" << fstab[i].mount_point
<< " blk_dev=" << realpath(fstab[i].blk_device) << " rec[" << i
<< "].fs_type=" << fstab[i].fs_type;
mount_errno = EINVAL; // continue bootup for metadata encryption
continue;
}
int retry_count = 2;
while (retry_count-- > 0) {
if (!__mount(fstab[i].blk_device, fstab[i].mount_point, fstab[i])) {
*attempted_idx = i;
mounted = true;
if (i != start_idx) {
LINFO << __FUNCTION__ << "(): Mounted " << fstab[i].blk_device << " on "
<< fstab[i].mount_point << " with fs_type=" << fstab[i].fs_type
<< " instead of " << fstab[start_idx].fs_type;
}
fs_stat &= ~FS_STAT_FULL_MOUNT_FAILED;
mount_errno = 0;
break;
} else {
if (retry_count <= 0) break; // run check_fs only once
fs_stat |= FS_STAT_FULL_MOUNT_FAILED;
// back up the first errno for crypto decisions.
if (mount_errno == 0) {
mount_errno = errno;
}
// retry after fsck
check_fs(fstab[i].blk_device, fstab[i].fs_type, fstab[i].mount_point, &fs_stat);
}
}
log_fs_stat(fstab[i].blk_device, fs_stat);
}
/* Adjust i for the case where it was still withing the recs[] */
if (i < fstab.size()) --i;
*end_idx = i;
if (!mounted) {
*attempted_idx = start_idx;
errno = mount_errno;
return false;
}
return true;
}
static bool TranslateExtLabels(FstabEntry* entry) {
if (!StartsWith(entry->blk_device, "LABEL=")) {
return true;
}
std::string label = entry->blk_device.substr(6);
if (label.size() > 16) {
LERROR << "FS label is longer than allowed by filesystem";
return false;
}
auto blockdir = std::unique_ptr<DIR, decltype(&closedir)>{opendir("/dev/block"), closedir};
if (!blockdir) {
LERROR << "couldn't open /dev/block";
return false;
}
struct dirent* ent;
while ((ent = readdir(blockdir.get()))) {
if (ent->d_type != DT_BLK)
continue;
unique_fd fd(TEMP_FAILURE_RETRY(
openat(dirfd(blockdir.get()), ent->d_name, O_RDONLY | O_CLOEXEC)));
if (fd < 0) {
LERROR << "Cannot open block device /dev/block/" << ent->d_name;
return false;
}