-
Notifications
You must be signed in to change notification settings - Fork 596
/
files.c
1804 lines (1490 loc) · 39.5 KB
/
files.c
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
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include <linux/limits.h>
#include <linux/major.h>
#include <sys/types.h>
#include <sys/prctl.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <stdlib.h>
#include "types.h"
#include "files.h"
#include "file-ids.h"
#include "files-reg.h"
#include "file-lock.h"
#include "image.h"
#include "common/list.h"
#include "rst-malloc.h"
#include "util-caps.h"
#include "common/lock.h"
#include "sockets.h"
#include "pstree.h"
#include "tty.h"
#include "pipes.h"
#include "fifo.h"
#include "eventfd.h"
#include "eventpoll.h"
#include "fsnotify.h"
#include "sk-packet.h"
#include "mount.h"
#include "signalfd.h"
#include "memfd.h"
#include "namespaces.h"
#include "tun.h"
#include "timerfd.h"
#include "imgset.h"
#include "fs-magic.h"
#include "fdinfo.h"
#include "cr_options.h"
#include "autofs.h"
#include "parasite.h"
#include "parasite-syscall.h"
#include "string.h"
#include "kerndat.h"
#include "fdstore.h"
#include "bpfmap.h"
#include "protobuf.h"
#include "util.h"
#include "images/fs.pb-c.h"
#include "images/ext-file.pb-c.h"
#include "plugin.h"
#define FDESC_HASH_SIZE 64
static struct hlist_head file_desc_hash[FDESC_HASH_SIZE];
/* file_desc's, which fle is not owned by a process, that is able to open them */
static LIST_HEAD(fake_master_head);
static u32 max_file_desc_id = 0;
static void init_fdesc_hash(void)
{
int i;
for (i = 0; i < FDESC_HASH_SIZE; i++)
INIT_HLIST_HEAD(&file_desc_hash[i]);
}
void file_desc_init(struct file_desc *d, u32 id, struct file_desc_ops *ops)
{
INIT_LIST_HEAD(&d->fd_info_head);
INIT_LIST_HEAD(&d->fake_master_list);
INIT_HLIST_NODE(&d->hash);
d->id = id;
d->ops = ops;
}
int file_desc_add(struct file_desc *d, u32 id, struct file_desc_ops *ops)
{
file_desc_init(d, id, ops);
hlist_add_head(&d->hash, &file_desc_hash[id % FDESC_HASH_SIZE]);
if (id > max_file_desc_id)
max_file_desc_id = id;
return 0; /* this is to make tail-calls in collect_one_foo look nice */
}
struct file_desc *find_file_desc_raw(int type, u32 id)
{
struct file_desc *d;
struct hlist_head *chain;
chain = &file_desc_hash[id % FDESC_HASH_SIZE];
hlist_for_each_entry(d, chain, hash)
if ((d->id == id) && (d->ops->type == type || type == FD_TYPES__UND))
/*
* Warning -- old CRIU might generate matching IDs
* for different file types! So any code that uses
* FD_TYPES__UND for fdesc search MUST make sure it's
* dealing with the merged files images where all
* descs are forced to have different IDs.
*/
return d;
return NULL;
}
static inline struct file_desc *find_file_desc(FdinfoEntry *fe)
{
return find_file_desc_raw(fe->type, fe->id);
}
u32 find_unused_file_desc_id(void)
{
return max_file_desc_id + 1;
}
struct fdinfo_list_entry *find_used_fd(struct pstree_item *task, int fd)
{
struct list_head *head;
struct fdinfo_list_entry *fle;
head = &rsti(task)->fds;
list_for_each_entry_reverse(fle, head, ps_list) {
if (fle->fe->fd == fd)
return fle;
/* List is ordered, so let's stop */
if (fle->fe->fd < fd)
break;
}
return NULL;
}
static void collect_task_fd(struct fdinfo_list_entry *new_fle, struct rst_info *ri)
{
struct fdinfo_list_entry *fle;
/*
* fles in fds list are ordered by fd. Fds are restored from img files
* in ascending order, so it is faster to insert them from the end of
* the list.
*/
list_for_each_entry_reverse(fle, &ri->fds, ps_list) {
if (fle->fe->fd < new_fle->fe->fd)
break;
}
list_add(&new_fle->ps_list, &fle->ps_list);
}
unsigned int find_unused_fd(struct pstree_item *task, int hint_fd)
{
struct list_head *head;
struct fdinfo_list_entry *fle;
int fd = 0, prev_fd;
if ((hint_fd >= 0) && (!find_used_fd(task, hint_fd))) {
fd = hint_fd;
goto out;
}
prev_fd = service_fd_min_fd(task) - 1;
head = &rsti(task)->fds;
list_for_each_entry_reverse(fle, head, ps_list) {
fd = fle->fe->fd;
if (prev_fd > fd) {
fd++;
goto out;
}
prev_fd = fd - 1;
}
BUG();
out:
return fd;
}
int find_unused_fd_pid(pid_t pid)
{
struct pstree_item *task;
task = pstree_item_by_virt(pid);
if (!task) {
pr_err("Invalid pid:%d\n", pid);
return -1;
}
return find_unused_fd(task, -1);
}
int set_fds_event(pid_t virt)
{
struct pstree_item *item;
bool is_set;
item = pstree_item_by_virt(virt);
BUG_ON(!item);
is_set = !!test_and_set_bit_le(FDS_EVENT_BIT, &item->task_st_le_bits);
if (!is_set)
futex_wake(&item->task_st);
return 0;
}
void clear_fds_event(void)
{
clear_bit_le(FDS_EVENT_BIT, ¤t->task_st_le_bits);
}
void wait_fds_event(void)
{
futex_t *f = ¤t->task_st;
int value;
value = htole32(FDS_EVENT);
futex_wait_if_cond(f, value, &);
clear_fds_event();
}
struct fdinfo_list_entry *try_file_master(struct file_desc *d)
{
if (list_empty(&d->fd_info_head))
return NULL;
return list_first_entry(&d->fd_info_head, struct fdinfo_list_entry, desc_list);
}
struct fdinfo_list_entry *file_master(struct file_desc *d)
{
struct fdinfo_list_entry *fle;
fle = try_file_master(d);
if (!fle) {
pr_err("Empty list on file desc id %#x(%d)\n", d->id, d->ops ? d->ops->type : -1);
BUG();
}
return fle;
}
void show_saved_files(void)
{
int i;
struct file_desc *fd;
pr_info("File descs:\n");
for (i = 0; i < FDESC_HASH_SIZE; i++)
hlist_for_each_entry(fd, &file_desc_hash[i], hash) {
struct fdinfo_list_entry *le;
pr_info(" `- type %d ID %#x\n", fd->ops->type, fd->id);
list_for_each_entry(le, &fd->fd_info_head, desc_list)
pr_info(" `- FD %d pid %d\n", le->fe->fd, le->pid);
}
}
/*
* Workaround for the OverlayFS bug present before Kernel 4.2
*
* This is here only to support the Linux Kernel between versions
* 3.18 and 4.2. After that, this workaround is not needed anymore,
* but it will work properly on both a kernel with and without the bug.
*
* When a process has a file open in an OverlayFS directory,
* the information in /proc/<pid>/fd/<fd> and /proc/<pid>/fdinfo/<fd>
* is wrong. We can't even rely on stat()-ing /proc/<pid>/fd/<fd> since
* this will show us the wrong filesystem type.
*
* So we grab that information from the mountinfo table instead. This is done
* every time fill_fdlink is called. See lookup_overlayfs for more details.
*
*/
static int fixup_overlayfs(struct fd_parms *p, struct fd_link *link)
{
struct mount_info *m;
if (!link)
return 0;
m = lookup_overlayfs(link->name, p->stat.st_dev, p->stat.st_ino, p->mnt_id);
if (IS_ERR(m))
return -1;
if (!m)
return 0;
p->mnt_id = m->mnt_id;
/*
* If the bug is present, the file path from /proc/<pid>/fd
* does not include the mountpoint, so we prepend it ourselves.
*/
if (strcmp("./", m->ns_mountpoint) != 0) {
char buf[PATH_MAX];
int n;
__strlcpy(buf, link->name, PATH_MAX);
n = snprintf(link->name, PATH_MAX, "%s/%s", m->ns_mountpoint, buf + 2);
if (n >= PATH_MAX) {
pr_err("Not enough space to replace %s\n", buf);
return -1;
}
}
return 0;
}
/*
* The gen_id thing is used to optimize the comparison of shared files.
* If two files have different gen_ids, then they are different for sure.
* If it matches, we don't know it and have to call sys_kcmp().
*
* The kcmp-ids.c engine does this trick, see comments in it for more info.
*/
uint32_t make_gen_id(uint32_t st_dev, uint32_t st_ino, uint64_t pos)
{
uint32_t pos_hi = pos >> 32;
uint32_t pos_low = pos & 0xffffffff;
return st_dev ^ st_ino ^ pos_hi ^ pos_low;
}
int do_dump_gen_file(struct fd_parms *p, int lfd, const struct fdtype_ops *ops, FdinfoEntry *e)
{
int ret = -1;
e->type = ops->type;
e->id = make_gen_id((uint32_t)p->stat.st_dev, (uint32_t)p->stat.st_ino, (uint64_t)p->pos);
e->fd = p->fd;
e->flags = p->fd_flags;
ret = fd_id_generate(p->pid, e, p);
if (ret == 1) /* new ID generated */
ret = ops->dump(lfd, e->id, p);
else
/* Remove locks generated by the fd before going to the next */
discard_dup_locks_tail(p->pid, e->fd);
return ret;
}
int fill_fdlink(int lfd, const struct fd_parms *p, struct fd_link *link)
{
int len;
link->name[0] = '.';
len = read_fd_link(lfd, &link->name[1], sizeof(link->name) - 1);
if (len < 0) {
pr_err("Can't read link for pid %d fd %d\n", p->pid, p->fd);
return -1;
}
link->len = len + 1;
if (opts.overlayfs)
if (fixup_overlayfs((struct fd_parms *)p, link) < 0)
return -1;
return 0;
}
static int fill_fd_params(struct pid *owner_pid, int fd, int lfd, struct fd_opts *opts, struct fd_parms *p)
{
int ret;
struct statfs fsbuf;
struct fdinfo_common fdinfo = { .mnt_id = -1, .owner = owner_pid->ns[0].virt };
if (fstat(lfd, &p->stat) < 0) {
pr_perror("Can't stat fd %d", lfd);
return -1;
}
if (fstatfs(lfd, &fsbuf) < 0) {
pr_perror("Can't statfs fd %d", lfd);
return -1;
}
if (parse_fdinfo_pid(owner_pid->real, fd, FD_TYPES__UND, &fdinfo))
return -1;
p->fs_type = fsbuf.f_type;
p->fd = fd;
p->pos = fdinfo.pos;
/*
* The kernel artificially adds the O_CLOEXEC flag on the file pointer
* flags by looking at the flags on the file descriptor (see kernel
* code fs/proc/fd.c). FD_CLOEXEC is a file descriptor property, which
* is saved in fd_flags.
*/
p->flags = fdinfo.flags & ~O_CLOEXEC;
p->mnt_id = fdinfo.mnt_id;
p->pid = owner_pid->real;
p->fd_flags = opts->flags;
fown_entry__init(&p->fown);
pr_info("%d fdinfo %d: pos: %#16" PRIx64 " flags: %16o/%#x\n", owner_pid->real, fd, p->pos, p->flags,
(int)p->fd_flags);
if (p->flags & O_PATH)
ret = 0;
else
ret = fcntl(lfd, F_GETSIG, 0);
if (ret < 0) {
pr_perror("Can't get owner signum on %d", lfd);
return -1;
}
p->fown.signum = ret;
if (opts->fown.pid == 0)
return 0;
p->fown.pid = opts->fown.pid;
p->fown.pid_type = opts->fown.pid_type;
p->fown.uid = opts->fown.uid;
p->fown.euid = opts->fown.euid;
return 0;
}
static const struct fdtype_ops *get_misc_dev_ops(int minor)
{
switch (minor) {
case TUN_MINOR:
return &tunfile_dump_ops;
case AUTOFS_MINOR:
return ®file_dump_ops;
};
return NULL;
}
static const struct fdtype_ops *get_mem_dev_ops(struct fd_parms *p, int minor)
{
const struct fdtype_ops *ops = NULL;
/*
* If /dev/kmsg is opened in write-only mode the file position
* should not be set up upon restore, kernel doesn't allow that.
*/
if (minor == 11 && (p->flags & O_ACCMODE) == O_WRONLY && p->pos == 0)
p->pos = -1ULL;
ops = ®file_dump_ops;
return ops;
}
static int dump_chrdev(struct fd_parms *p, int lfd, FdinfoEntry *e)
{
struct fd_link *link_old = p->link;
int maj = major(p->stat.st_rdev);
const struct fdtype_ops *ops;
struct fd_link link;
int err;
switch (maj) {
case MEM_MAJOR:
ops = get_mem_dev_ops(p, minor(p->stat.st_rdev));
break;
case MISC_MAJOR:
ops = get_misc_dev_ops(minor(p->stat.st_rdev));
if (ops)
break;
/* fallthrough */
default: {
char more[32];
if (is_tty(p->stat.st_rdev, p->stat.st_dev)) {
if (fill_fdlink(lfd, p, &link))
return -1;
p->link = &link;
ops = &tty_dump_ops;
break;
}
sprintf(more, "%d:%d", maj, minor(p->stat.st_rdev));
err = dump_unsupp_fd(p, lfd, "chr", more, e);
p->link = link_old;
return err;
}
}
err = do_dump_gen_file(p, lfd, ops, e);
p->link = link_old;
return err;
}
static int dump_one_file(struct pid *pid, int fd, int lfd, struct fd_opts *opts, struct parasite_ctl *ctl,
FdinfoEntry *e, struct parasite_drain_fd *dfds)
{
struct fd_parms p = FD_PARMS_INIT;
const struct fdtype_ops *ops;
struct fd_link link;
if (fill_fd_params(pid, fd, lfd, opts, &p) < 0) {
pr_err("Can't get stat on %d\n", fd);
return -1;
}
if (note_file_lock(pid, fd, lfd, &p))
return -1;
/* Lease can be set only on regular file */
if (S_ISREG(p.stat.st_mode)) {
int ret = correct_file_leases_type(pid, fd, lfd);
if (ret < 0)
return ret;
}
p.fd_ctl = ctl; /* Some dump_opts require this to talk to parasite */
p.dfds = dfds; /* epoll needs to verify if target fd exist */
if (S_ISSOCK(p.stat.st_mode))
return dump_socket(&p, lfd, e);
if (S_ISCHR(p.stat.st_mode))
return dump_chrdev(&p, lfd, e);
if (p.fs_type == ANON_INODE_FS_MAGIC) {
char link[32];
if (read_fd_link(lfd, link, sizeof(link)) < 0)
return -1;
if (is_eventfd_link(link))
ops = &eventfd_dump_ops;
else if (is_eventpoll_link(link))
ops = &eventpoll_dump_ops;
else if (is_inotify_link(link))
ops = &inotify_dump_ops;
else if (is_fanotify_link(link))
ops = &fanotify_dump_ops;
else if (is_signalfd_link(link))
ops = &signalfd_dump_ops;
else if (is_timerfd_link(link))
ops = &timerfd_dump_ops;
#ifdef CONFIG_HAS_LIBBPF
else if (is_bpfmap_link(link))
ops = &bpfmap_dump_ops;
#endif
else
return dump_unsupp_fd(&p, lfd, "anon", link, e);
return do_dump_gen_file(&p, lfd, ops, e);
}
if (S_ISREG(p.stat.st_mode) || S_ISDIR(p.stat.st_mode) || S_ISLNK(p.stat.st_mode)) {
if (fill_fdlink(lfd, &p, &link))
return -1;
p.link = &link;
/* TODO: Dump for hugetlb fd when memfd hugetlb is not supported */
if (is_memfd(p.stat.st_dev) || (kdat.has_memfd_hugetlb && is_hugetlb_dev(p.stat.st_dev, NULL)))
ops = &memfd_dump_ops;
else if (link.name[1] == '/')
ops = ®file_dump_ops;
else if (check_ns_proc(&link))
ops = &nsfile_dump_ops;
else
return dump_unsupp_fd(&p, lfd, "reg", link.name + 1, e);
return do_dump_gen_file(&p, lfd, ops, e);
}
if (S_ISFIFO(p.stat.st_mode)) {
if (p.fs_type == PIPEFS_MAGIC)
ops = &pipe_dump_ops;
else
ops = &fifo_dump_ops;
return do_dump_gen_file(&p, lfd, ops, e);
}
/*
* For debug purpose -- at least show the link
* file pointing to when reporting unsupported file.
* On error simply empty string here.
*/
if (fill_fdlink(lfd, &p, &link))
memzero(&link, sizeof(link));
return dump_unsupp_fd(&p, lfd, "unknown", link.name + 1, e);
}
int dump_my_file(int lfd, u32 *id, int *type)
{
struct pid me = {};
struct fd_opts fdo = {};
FdinfoEntry e = FDINFO_ENTRY__INIT;
me.real = getpid();
me.ns[0].virt = -1; /* FIXME */
if (dump_one_file(&me, lfd, lfd, &fdo, NULL, &e, NULL))
return -1;
*id = e.id;
*type = e.type;
return 0;
}
int dump_task_files_seized(struct parasite_ctl *ctl, struct pstree_item *item, struct parasite_drain_fd *dfds)
{
int *lfds = NULL;
struct cr_img *img = NULL;
struct fd_opts *opts = NULL;
int i, ret = -1;
int off, nr_fds = min((int)PARASITE_MAX_FDS, dfds->nr_fds);
pr_info("\n");
pr_info("Dumping opened files (pid: %d)\n", item->pid->real);
pr_info("----------------------------------------\n");
lfds = xmalloc(nr_fds * sizeof(int));
if (!lfds)
goto err;
opts = xmalloc(nr_fds * sizeof(struct fd_opts));
if (!opts)
goto err;
img = open_image(CR_FD_FDINFO, O_DUMP, item->ids->files_id);
if (!img)
goto err;
ret = 0; /* Don't fail if nr_fds == 0 */
for (off = 0; ret == 0 && off < dfds->nr_fds; off += nr_fds) {
if (nr_fds + off > dfds->nr_fds)
nr_fds = dfds->nr_fds - off;
ret = parasite_drain_fds_seized(ctl, dfds, nr_fds, off, lfds, opts);
if (ret)
goto err;
for (i = 0; i < nr_fds; i++) {
FdinfoEntry e = FDINFO_ENTRY__INIT;
ret = dump_one_file(item->pid, dfds->fds[i + off], lfds[i], opts + i, ctl, &e, dfds);
if (ret)
break;
ret = pb_write_one(img, &e, PB_FDINFO);
if (ret)
break;
}
for (i = 0; i < nr_fds; i++)
close(lfds[i]);
}
pr_info("----------------------------------------\n");
err:
if (img)
close_image(img);
xfree(opts);
xfree(lfds);
return ret;
}
static int predump_one_fd(int pid, int fd)
{
const struct fdtype_ops *ops;
char link[PATH_MAX], t[32];
int ret = 0;
snprintf(t, sizeof(t), "/proc/%d/fd/%d", pid, fd);
ret = readlink(t, link, sizeof(link));
if (ret < 0) {
pr_perror("Can't read link of fd %d", fd);
return -1;
} else if ((size_t)ret == sizeof(link)) {
pr_err("Buffer for read link of fd %d is too small\n", fd);
return -1;
}
link[ret] = 0;
ret = 0;
if (is_inotify_link(link))
ops = &inotify_dump_ops;
else if (is_fanotify_link(link))
ops = &fanotify_dump_ops;
else
goto out;
pr_debug("Pre-dumping %d's %d fd\n", pid, fd);
ret = ops->pre_dump(pid, fd);
out:
return ret;
}
int predump_task_files(int pid)
{
struct dirent *de;
DIR *fd_dir;
int ret = -1;
pr_info("Pre-dump fds for %d)\n", pid);
fd_dir = opendir_proc(pid, "fd");
if (!fd_dir)
return -1;
while ((de = readdir(fd_dir))) {
if (dir_dots(de))
continue;
if (predump_one_fd(pid, atoi(de->d_name)))
goto out;
}
ret = 0;
out:
closedir(fd_dir);
return ret;
}
int restore_fown(int fd, FownEntry *fown)
{
struct f_owner_ex owner;
uid_t uids[3];
if (fown->signum) {
if (fcntl(fd, F_SETSIG, fown->signum)) {
pr_perror("Can't set signal");
return -1;
}
}
/* May be untouched */
if (!fown->pid)
return 0;
if (getresuid(&uids[0], &uids[1], &uids[2])) {
pr_perror("Can't get current UIDs");
return -1;
}
if (setresuid(fown->uid, fown->euid, uids[2])) {
pr_perror("Can't set UIDs");
return -1;
}
owner.type = fown->pid_type;
owner.pid = fown->pid;
if (fcntl(fd, F_SETOWN_EX, &owner)) {
pr_perror("Can't setup %d file owner pid", fd);
return -1;
}
if (setresuid(uids[0], uids[1], uids[2])) {
pr_perror("Can't revert UIDs back");
return -1;
}
if (prctl(PR_SET_DUMPABLE, 1, 0))
pr_perror("Unable to set PR_SET_DUMPABLE");
return 0;
}
int rst_file_params(int fd, FownEntry *fown, int flags)
{
if (set_fd_flags(fd, flags) < 0)
return -1;
if (restore_fown(fd, fown) < 0)
return -1;
return 0;
}
static struct fdinfo_list_entry *alloc_fle(int pid, FdinfoEntry *fe)
{
struct fdinfo_list_entry *fle;
fle = shmalloc(sizeof(*fle));
if (!fle)
return NULL;
fle->pid = pid;
fle->fe = fe;
fle->received = 0;
fle->fake = 0;
fle->stage = FLE_INITIALIZED;
fle->task = pstree_item_by_virt(pid);
if (!fle->task) {
pr_err("Can't find task with pid %d\n", pid);
shfree_last(fle);
return NULL;
}
return fle;
}
static void __collect_desc_fle(struct fdinfo_list_entry *new_le, struct file_desc *fdesc)
{
struct fdinfo_list_entry *le;
list_for_each_entry_reverse(le, &fdesc->fd_info_head, desc_list)
if (pid_rst_prio_eq(le->pid, new_le->pid))
break;
list_add(&new_le->desc_list, &le->desc_list);
}
static void collect_desc_fle(struct fdinfo_list_entry *new_le, struct file_desc *fdesc, bool force_master)
{
new_le->desc = fdesc;
if (!force_master)
__collect_desc_fle(new_le, fdesc);
else {
/* Link as first entry */
list_add(&new_le->desc_list, &fdesc->fd_info_head);
}
}
struct fdinfo_list_entry *collect_fd_to(int pid, FdinfoEntry *e, struct rst_info *rst_info, struct file_desc *fdesc,
bool fake, bool force_master)
{
struct fdinfo_list_entry *new_le;
new_le = alloc_fle(pid, e);
if (new_le) {
new_le->fake = (!!fake);
collect_desc_fle(new_le, fdesc, force_master);
collect_task_fd(new_le, rst_info);
}
return new_le;
}
int collect_fd(int pid, FdinfoEntry *e, struct rst_info *rst_info, bool fake)
{
struct file_desc *fdesc;
pr_info("Collect fdinfo pid=%d fd=%d id=%#x\n", pid, e->fd, e->id);
fdesc = find_file_desc(e);
if (fdesc == NULL) {
pr_err("No file for fd %d id %#x\n", e->fd, e->id);
return -1;
}
if (!collect_fd_to(pid, e, rst_info, fdesc, fake, false))
return -1;
return 0;
}
FdinfoEntry *dup_fdinfo(FdinfoEntry *old, int fd, unsigned flags)
{
FdinfoEntry *e;
e = shmalloc(sizeof(*e));
if (!e)
return NULL;
fdinfo_entry__init(e);
e->id = old->id;
e->type = old->type;
e->fd = fd;
e->flags = flags;
return e;
}
int dup_fle(struct pstree_item *task, struct fdinfo_list_entry *ple, int fd, unsigned flags)
{
FdinfoEntry *e;
e = dup_fdinfo(ple->fe, fd, flags);
if (!e)
return -1;
return collect_fd(vpid(task), e, rsti(task), false);
}
int prepare_fd_pid(struct pstree_item *item)
{
int ret = 0;
struct cr_img *img;
pid_t pid = vpid(item);
struct rst_info *rst_info = rsti(item);
INIT_LIST_HEAD(&rst_info->fds);
if (item->ids == NULL) /* zombie */
return 0;
if (rsti(item)->fdt && rsti(item)->fdt->pid != vpid(item))
return 0;
img = open_image(CR_FD_FDINFO, O_RSTR, item->ids->files_id);
if (!img)
return -1;
while (1) {
FdinfoEntry *e;
ret = pb_read_one_eof(img, &e, PB_FDINFO);
if (ret <= 0)
break;
if (e->fd >= kdat.sysctl_nr_open) {
ret = -1;
pr_err("Too big FD number to restore %d\n", e->fd);
break;
}
ret = collect_fd(pid, e, rst_info, false);
if (ret < 0) {
fdinfo_entry__free_unpacked(e, NULL);
break;
}
}
close_image(img);
return ret;
}
#define SETFL_MASK (O_APPEND | O_ASYNC | O_NONBLOCK | O_NDELAY | O_DIRECT | O_NOATIME)
int set_fd_flags(int fd, int flags)
{
int ret;
ret = fcntl(fd, F_GETFL, 0);
if (ret < 0)
goto err;
flags = (SETFL_MASK & flags) | (ret & ~SETFL_MASK);
ret = fcntl(fd, F_SETFL, flags);
if (ret < 0)
goto err;
/* Let's check, that now actual flags contains those we need */
ret = fcntl(fd, F_GETFL, 0);
if (ret < 0)
goto err;
if (ret != flags) {
pr_err("fcntl call on fd %d (flags %#o) succeeded, "
"but some flags were dropped: %#o\n",
fd, flags, ret);
return -1;
}
return 0;
err:
pr_perror("fcntl call on fd %d (flags %x) failed", fd, flags);
return -1;
}
struct fd_open_state {
char *name;
int (*cb)(int, struct fdinfo_list_entry *);
};
static int receive_fd(struct fdinfo_list_entry *fle);
static void transport_name_gen(struct sockaddr_un *addr, int *len, int pid)
{
addr->sun_family = AF_UNIX;
snprintf(addr->sun_path, UNIX_PATH_MAX, "x/crtools-fd-%d-%" PRIx64, pid, criu_run_id);
*len = SUN_LEN(addr);
*addr->sun_path = '\0';
}
static bool task_fle(struct pstree_item *task, struct fdinfo_list_entry *fle)
{
struct fdinfo_list_entry *tmp;
list_for_each_entry(tmp, &rsti(task)->fds, ps_list)
if (fle == tmp)
return true;
return false;
}
static int plant_fd(struct fdinfo_list_entry *fle, int fd)
{
BUG_ON(fle->received);
fle->received = 1;
return reopen_fd_as(fle->fe->fd, fd);
}
static int recv_fd_from_peer(struct fdinfo_list_entry *fle)
{
struct fdinfo_list_entry *tmp;
int fd, ret, tsock;
if (fle->received)