-
Notifications
You must be signed in to change notification settings - Fork 1.1k
/
attach.c
1868 lines (1532 loc) · 48.7 KB
/
attach.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
/* SPDX-License-Identifier: LGPL-2.1+ */
#ifndef _GNU_SOURCE
#define _GNU_SOURCE 1
#endif
#include <errno.h>
#include <fcntl.h>
#include <grp.h>
#include <linux/unistd.h>
#include <pwd.h>
#include <pthread.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/mount.h>
#include <sys/param.h>
#include <sys/prctl.h>
#include <sys/socket.h>
#include <sys/syscall.h>
#include <sys/wait.h>
#include <termios.h>
#include <unistd.h>
#include <lxc/lxccontainer.h>
#include "af_unix.h"
#include "attach.h"
#include "caps.h"
#include "cgroups/cgroup.h"
#include "cgroups/cgroup_utils.h"
#include "commands.h"
#include "conf.h"
#include "config.h"
#include "confile.h"
#include "log.h"
#include "lsm/lsm.h"
#include "lxclock.h"
#include "lxcseccomp.h"
#include "macro.h"
#include "mainloop.h"
#include "memory_utils.h"
#include "mount_utils.h"
#include "namespace.h"
#include "process_utils.h"
#include "sync.h"
#include "syscall_wrappers.h"
#include "terminal.h"
#include "utils.h"
#if HAVE_SYS_PERSONALITY_H
#include <sys/personality.h>
#endif
lxc_log_define(attach, lxc);
/* Define default options if no options are supplied by the user. */
static lxc_attach_options_t attach_static_default_options = LXC_ATTACH_OPTIONS_DEFAULT;
/*
* The context used to attach to the container.
* @attach_flags : the attach flags specified in lxc_attach_options_t
* @init_pid : the PID of the container's init process
* @dfd_init_pid : file descriptor to /proc/@init_pid
* __Must be closed in attach_context_security_barrier()__!
* @dfd_self_pid : file descriptor to /proc/self
* __Must be closed in attach_context_security_barrier()__!
* @setup_ns_uid : if CLONE_NEWUSER is specified will contain the uid used
* during attach setup.
* @setup_ns_gid : if CLONE_NEWUSER is specified will contain the gid used
* during attach setup.
* @target_ns_uid : if CLONE_NEWUSER is specified the uid that the final
* program will be run with.
* @target_ns_gid : if CLONE_NEWUSER is specified the gid that the final
* program will be run with.
* @target_host_uid : if CLONE_NEWUSER is specified the uid that the final
* program will be run with on the host.
* @target_host_gid : if CLONE_NEWUSER is specified the gid that the final
* program will be run with on the host.
* @lsm_label : LSM label to be used for the attaching process
* @container : the container we're attaching o
* @personality : the personality to use for the final program
* @capability : the capability mask of the @init_pid
* @ns_inherited : flags of namespaces that the final program will inherit
* from @init_pid
* @ns_fd : file descriptors to @init_pid's namespaces
*/
struct attach_context {
unsigned int attach_flags;
int init_pid;
int init_pidfd;
int dfd_init_pid;
int dfd_self_pid;
uid_t setup_ns_uid;
gid_t setup_ns_gid;
uid_t target_ns_uid;
gid_t target_ns_gid;
uid_t target_host_uid;
uid_t target_host_gid;
char *lsm_label;
struct lxc_container *container;
signed long personality;
unsigned long long capability_mask;
int ns_inherited;
int ns_fd[LXC_NS_MAX];
struct lsm_ops *lsm_ops;
};
static pid_t pidfd_get_pid(int dfd_init_pid, int pidfd)
{
__do_free char *line = NULL;
__do_fclose FILE *f = NULL;
size_t len = 0;
char path[STRLITERALLEN("fdinfo/") + INTTYPE_TO_STRLEN(int) + 1 ] = "fdinfo/";
int ret;
if (dfd_init_pid < 0 || pidfd < 0)
return ret_errno(EBADF);
ret = strnprintf(path + STRLITERALLEN("fdinfo/"), INTTYPE_TO_STRLEN(int), "%d", pidfd);
if (ret < 0)
return ret_errno(EIO);
f = fdopen_at(dfd_init_pid, path, "re", PROTECT_OPEN, PROTECT_LOOKUP_BENEATH);
if (!f)
return -errno;
while (getline(&line, &len, f) != -1) {
const char *prefix = "Pid:\t";
const size_t prefix_len = STRLITERALLEN("Pid:\t");
int pid = -ESRCH;
char *slider = line;
if (!strnequal(slider, prefix, prefix_len))
continue;
slider += prefix_len;
slider = lxc_trim_whitespace_in_place(slider);
ret = lxc_safe_int(slider, &pid);
if (ret)
return -ret;
return pid;
}
return ret_errno(ENOENT);
}
static inline bool sync_wake_pid(int fd, pid_t pid)
{
return lxc_write_nointr(fd, &pid, sizeof(pid_t)) == sizeof(pid_t);
}
static inline bool sync_wait_pid(int fd, pid_t *pid)
{
return lxc_read_nointr(fd, pid, sizeof(pid_t)) == sizeof(pid_t);
}
static inline bool sync_wake_fd(int fd, int fd_send)
{
return lxc_abstract_unix_send_fds(fd, &fd_send, 1, NULL, 0) > 0;
}
static inline bool sync_wait_fd(int fd, int *fd_recv)
{
return lxc_abstract_unix_recv_one_fd(fd, fd_recv, NULL, 0) > 0;
}
static bool attach_lsm(lxc_attach_options_t *options)
{
return (options->namespaces & CLONE_NEWNS) &&
(options->attach_flags & (LXC_ATTACH_LSM | LXC_ATTACH_LSM_LABEL));
}
static struct attach_context *alloc_attach_context(void)
{
struct attach_context *ctx;
ctx = zalloc(sizeof(struct attach_context));
if (!ctx)
return ret_set_errno(NULL, ENOMEM);
ctx->dfd_self_pid = -EBADF;
ctx->dfd_init_pid = -EBADF;
ctx->init_pidfd = -EBADF;
ctx->init_pid = -ESRCH;
ctx->setup_ns_uid = LXC_INVALID_UID;
ctx->setup_ns_gid = LXC_INVALID_GID;
ctx->target_ns_uid = LXC_INVALID_UID;
ctx->target_ns_gid = LXC_INVALID_GID;
ctx->target_host_uid = LXC_INVALID_UID;
ctx->target_host_gid = LXC_INVALID_GID;
for (int i = 0; i < LXC_NS_MAX; i++)
ctx->ns_fd[i] = -EBADF;
return ctx;
}
static int get_personality(const char *name, const char *lxcpath,
signed long *personality)
{
__do_free char *p = NULL;
signed long per;
p = lxc_cmd_get_config_item(name, "lxc.arch", lxcpath);
if (!p) {
*personality = LXC_ARCH_UNCHANGED;
return 0;
}
per = lxc_config_parse_arch(p);
if (per == LXC_ARCH_UNCHANGED)
return ret_errno(EINVAL);
*personality = per;
return 0;
}
static int userns_setup_ids(struct attach_context *ctx,
lxc_attach_options_t *options)
{
__do_free char *line = NULL;
__do_fclose FILE *f_gidmap = NULL, *f_uidmap = NULL;
size_t len = 0;
uid_t init_ns_uid = LXC_INVALID_UID;
gid_t init_ns_gid = LXC_INVALID_GID;
uid_t nsuid, hostuid, range_uid;
gid_t nsgid, hostgid, range_gid;
if (!(options->namespaces & CLONE_NEWUSER))
return 0;
f_uidmap = fdopen_at(ctx->dfd_init_pid, "uid_map", "re", PROTECT_OPEN, PROTECT_LOOKUP_BENEATH);
if (!f_uidmap)
return log_error_errno(-errno, errno, "Failed to open uid_map");
while (getline(&line, &len, f_uidmap) != -1) {
if (sscanf(line, "%u %u %u", &nsuid, &hostuid, &range_uid) != 3)
continue;
if (0 >= nsuid && 0 < nsuid + range_uid) {
ctx->setup_ns_uid = 0;
TRACE("Container has mapping for uid 0");
break;
}
if (ctx->target_host_uid >= hostuid && ctx->target_host_uid < hostuid + range_uid) {
init_ns_uid = (ctx->target_host_uid - hostuid) + nsuid;
TRACE("Container runs with uid %d", init_ns_uid);
}
}
f_gidmap = fdopen_at(ctx->dfd_init_pid, "gid_map", "re", PROTECT_OPEN, PROTECT_LOOKUP_BENEATH);
if (!f_gidmap)
return log_error_errno(-errno, errno, "Failed to open gid_map");
while (getline(&line, &len, f_gidmap) != -1) {
if (sscanf(line, "%u %u %u", &nsgid, &hostgid, &range_gid) != 3)
continue;
if (0 >= nsgid && 0 < nsgid + range_gid) {
ctx->setup_ns_gid = 0;
TRACE("Container has mapping for gid 0");
break;
}
if (ctx->target_host_gid >= hostgid && ctx->target_host_gid < hostgid + range_gid) {
init_ns_gid = (ctx->target_host_gid - hostgid) + nsgid;
TRACE("Container runs with gid %d", init_ns_gid);
}
}
if (ctx->setup_ns_uid == LXC_INVALID_UID)
ctx->setup_ns_uid = init_ns_uid;
if (ctx->setup_ns_gid == LXC_INVALID_UID)
ctx->setup_ns_gid = init_ns_gid;
return 0;
}
static void userns_target_ids(struct attach_context *ctx, lxc_attach_options_t *options)
{
if (options->uid != LXC_INVALID_UID)
ctx->target_ns_uid = options->uid;
else if (options->namespaces & CLONE_NEWUSER)
ctx->target_ns_uid = ctx->setup_ns_uid;
else
ctx->target_ns_uid = 0;
if (ctx->target_ns_uid == LXC_INVALID_UID)
WARN("Invalid uid specified");
if (options->gid != LXC_INVALID_GID)
ctx->target_ns_gid = options->gid;
else if (options->namespaces & CLONE_NEWUSER)
ctx->target_ns_gid = ctx->setup_ns_gid;
else
ctx->target_ns_gid = 0;
if (ctx->target_ns_gid == LXC_INVALID_GID)
WARN("Invalid gid specified");
}
static int parse_init_status(struct attach_context *ctx, lxc_attach_options_t *options)
{
__do_free char *line = NULL;
__do_fclose FILE *f = NULL;
size_t len = 0;
bool caps_found = false;
int ret;
f = fdopen_at(ctx->dfd_init_pid, "status", "re", PROTECT_OPEN, PROTECT_LOOKUP_BENEATH);
if (!f)
return log_error_errno(-errno, errno, "Failed to open status file");
while (getline(&line, &len, f) != -1) {
signed long value = -1;
/*
* Format is: real, effective, saved set user, fs we only care
* about real uid.
*/
ret = sscanf(line, "Uid: %ld", &value);
if (ret != EOF && ret == 1) {
ctx->target_host_uid = (uid_t)value;
TRACE("Container's init process runs with hostuid %d", ctx->target_host_uid);
goto next;
}
ret = sscanf(line, "Gid: %ld", &value);
if (ret != EOF && ret == 1) {
ctx->target_host_gid = (gid_t)value;
TRACE("Container's init process runs with hostgid %d", ctx->target_host_gid);
goto next;
}
ret = sscanf(line, "CapBnd: %llx", &ctx->capability_mask);
if (ret != EOF && ret == 1) {
caps_found = true;
goto next;
}
next:
if (ctx->target_host_uid != LXC_INVALID_UID &&
ctx->target_host_gid != LXC_INVALID_GID &&
caps_found)
break;
}
ret = userns_setup_ids(ctx, options);
if (ret)
return log_error_errno(ret, errno, "Failed to get setup ids");
userns_target_ids(ctx, options);
return 0;
}
static bool pidfd_setns_supported(struct attach_context *ctx)
{
int ret;
/*
* The ability to attach to time namespaces came after the introduction
* of of using pidfds for attaching to namespaces. To avoid having to
* special-case both CLONE_NEWUSER and CLONE_NEWTIME handling, let's
* use CLONE_NEWTIME as gatekeeper.
*/
if (ctx->init_pidfd >= 0)
ret = setns(ctx->init_pidfd, CLONE_NEWTIME);
else
ret = -EOPNOTSUPP;
TRACE("Attaching to namespaces via pidfds %s",
ret ? "unsupported" : "supported");
return ret == 0;
}
static int get_attach_context(struct attach_context *ctx,
struct lxc_container *container,
lxc_attach_options_t *options)
{
__do_free char *lsm_label = NULL;
int ret;
char path[LXC_PROC_PID_LEN];
ctx->container = container;
ctx->attach_flags = options->attach_flags;
ctx->dfd_self_pid = open_at(-EBADF, "/proc/self",
PROTECT_OPATH_FILE & ~O_NOFOLLOW,
(PROTECT_LOOKUP_ABSOLUTE_WITH_SYMLINKS & ~RESOLVE_NO_XDEV), 0);
if (ctx->dfd_self_pid < 0)
return log_error_errno(-errno, errno, "Failed to open /proc/self");
ctx->init_pidfd = lxc_cmd_get_init_pidfd(container->name, container->config_path);
if (ctx->init_pidfd >= 0)
ctx->init_pid = pidfd_get_pid(ctx->dfd_self_pid, ctx->init_pidfd);
else
ctx->init_pid = lxc_cmd_get_init_pid(container->name, container->config_path);
if (ctx->init_pid < 0)
return log_error(-1, "Failed to get init pid");
ret = strnprintf(path, sizeof(path), "/proc/%d", ctx->init_pid);
if (ret < 0)
return ret_errno(EIO);
ctx->dfd_init_pid = open_at(-EBADF, path,
PROTECT_OPATH_DIRECTORY,
(PROTECT_LOOKUP_ABSOLUTE & ~RESOLVE_NO_XDEV), 0);
if (ctx->dfd_init_pid < 0)
return log_error_errno(-errno, errno, "Failed to open /proc/%d", ctx->init_pid);
if (ctx->init_pidfd >= 0) {
ret = lxc_raw_pidfd_send_signal(ctx->init_pidfd, 0, NULL, 0);
if (ret)
return log_error_errno(-errno, errno, "Container process exited or PID has been recycled");
else
TRACE("Container process still running and PID was not recycled");
if (!pidfd_setns_supported(ctx)) {
/* We can't risk leaking file descriptors during attach. */
if (close(ctx->init_pidfd))
return log_error_errno(-errno, errno, "Failed to close pidfd");
ctx->init_pidfd = -EBADF;
TRACE("Attaching to namespaces via pidfds not supported");
}
}
/* Determine which namespaces the container was created with. */
if (options->namespaces == -1) {
options->namespaces = lxc_cmd_get_clone_flags(container->name, container->config_path);
if (options->namespaces == -1)
return log_error_errno(-EINVAL, EINVAL, "Failed to automatically determine the namespaces which the container uses");
for (int i = 0; i < LXC_NS_MAX; i++) {
if (ns_info[i].clone_flag & CLONE_NEWCGROUP)
if (!(options->attach_flags & LXC_ATTACH_MOVE_TO_CGROUP) ||
!cgns_supported())
continue;
if (ns_info[i].clone_flag & options->namespaces)
continue;
ctx->ns_inherited |= ns_info[i].clone_flag;
}
}
ret = parse_init_status(ctx, options);
if (ret)
return log_error_errno(-errno, errno, "Failed to open parse file");
ctx->lsm_ops = lsm_init_static();
if (attach_lsm(options)) {
if (ctx->attach_flags & LXC_ATTACH_LSM_LABEL)
lsm_label = options->lsm_label;
else
lsm_label = ctx->lsm_ops->process_label_get_at(ctx->lsm_ops, ctx->dfd_init_pid);
if (!lsm_label)
WARN("No security context received");
else
INFO("Retrieved security context %s", lsm_label);
}
ret = get_personality(container->name, container->config_path, &ctx->personality);
if (ret)
return log_error_errno(ret, errno, "Failed to get personality of the container");
if (!ctx->container->lxc_conf) {
ctx->container->lxc_conf = lxc_conf_init();
if (!ctx->container->lxc_conf)
return log_error_errno(-ENOMEM, ENOMEM, "Failed to allocate new lxc config");
}
ctx->lsm_label = move_ptr(lsm_label);
return 0;
}
static int same_nsfd(int dfd_pid1, int dfd_pid2, const char *ns_path)
{
int ret;
struct stat ns_st1, ns_st2;
ret = fstatat(dfd_pid1, ns_path, &ns_st1, 0);
if (ret)
return -errno;
ret = fstatat(dfd_pid2, ns_path, &ns_st2, 0);
if (ret)
return -errno;
/* processes are in the same namespace */
if ((ns_st1.st_dev == ns_st2.st_dev) &&
(ns_st1.st_ino == ns_st2.st_ino))
return 1;
return 0;
}
static int same_ns(int dfd_pid1, int dfd_pid2, const char *ns_path)
{
__do_close int ns_fd2 = -EBADF;
int ret = -1;
ns_fd2 = open_at(dfd_pid2, ns_path, PROTECT_OPEN_WITH_TRAILING_SYMLINKS,
(PROTECT_LOOKUP_BENEATH_WITH_MAGICLINKS &
~(RESOLVE_NO_XDEV | RESOLVE_BENEATH)), 0);
if (ns_fd2 < 0) {
if (errno == ENOENT)
return -ENOENT;
return syserror("Failed to open %d(%s)", dfd_pid2, ns_path);
}
ret = same_nsfd(dfd_pid1, dfd_pid2, ns_path);
switch (ret) {
case -ENOENT:
__fallthrough;
case 1:
return ret_errno(ENOENT);
case 0:
/* processes are in different namespaces */
return move_fd(ns_fd2);
}
return ret;
}
static int __prepare_namespaces_pidfd(struct attach_context *ctx)
{
for (int i = 0; i < LXC_NS_MAX; i++) {
int ret;
ret = same_nsfd(ctx->dfd_self_pid,
ctx->dfd_init_pid,
ns_info[i].proc_path);
switch (ret) {
case -ENOENT:
__fallthrough;
case 1:
ctx->ns_inherited &= ~ns_info[i].clone_flag;
TRACE("Shared %s namespace doesn't need attach", ns_info[i].proc_name);
continue;
case 0:
TRACE("Different %s namespace needs attach", ns_info[i].proc_name);
continue;
}
return syserror("Failed to determine whether %s namespace is shared",
ns_info[i].proc_name);
}
return 0;
}
static int __prepare_namespaces_nsfd(struct attach_context *ctx,
lxc_attach_options_t *options)
{
for (int i = 0; i < LXC_NS_MAX; i++) {
int j;
if (options->namespaces & ns_info[i].clone_flag)
ctx->ns_fd[i] = open_at(ctx->dfd_init_pid,
ns_info[i].proc_path,
PROTECT_OPEN_WITH_TRAILING_SYMLINKS,
(PROTECT_LOOKUP_BENEATH_WITH_MAGICLINKS &
~(RESOLVE_NO_XDEV | RESOLVE_BENEATH)),
0);
else if (ctx->ns_inherited & ns_info[i].clone_flag)
ctx->ns_fd[i] = same_ns(ctx->dfd_self_pid,
ctx->dfd_init_pid,
ns_info[i].proc_path);
else
continue;
if (ctx->ns_fd[i] >= 0)
continue;
if (ctx->ns_fd[i] == -ENOENT) {
ctx->ns_inherited &= ~ns_info[i].clone_flag;
continue;
}
/* We failed to preserve the namespace. */
SYSERROR("Failed to preserve %s namespace of %d",
ns_info[i].proc_name, ctx->init_pid);
/* Close all already opened file descriptors before we return an
* error, so we don't leak them.
*/
for (j = 0; j < i; j++)
close_prot_errno_disarm(ctx->ns_fd[j]);
return -1;
}
return 0;
}
static int prepare_namespaces(struct attach_context *ctx,
lxc_attach_options_t *options)
{
if (ctx->init_pidfd < 0)
return __prepare_namespaces_nsfd(ctx, options);
return __prepare_namespaces_pidfd(ctx);
}
static inline void put_namespaces(struct attach_context *ctx)
{
if (ctx->init_pidfd < 0) {
for (int i = 0; i < LXC_NS_MAX; i++)
close_prot_errno_disarm(ctx->ns_fd[i]);
}
}
static int __attach_namespaces_pidfd(struct attach_context *ctx,
lxc_attach_options_t *options)
{
unsigned int ns_flags = options->namespaces | ctx->ns_inherited;
int ret;
/* The common case is to attach to all namespaces. */
ret = setns(ctx->init_pidfd, ns_flags);
if (ret)
return log_error_errno(-errno, errno,
"Failed to attach to namespaces via pidfd");
/* We can't risk leaking file descriptors into the container. */
if (close(ctx->init_pidfd))
return log_error_errno(-errno, errno, "Failed to close pidfd");
ctx->init_pidfd = -EBADF;
return log_trace(0, "Attached to container namespaces via pidfd");
}
static int __attach_namespaces_nsfd(struct attach_context *ctx,
lxc_attach_options_t *options)
{
int fret = 0;
for (int i = 0; i < LXC_NS_MAX; i++) {
int ret;
if (ctx->ns_fd[i] < 0)
continue;
ret = setns(ctx->ns_fd[i], ns_info[i].clone_flag);
if (ret)
return log_error_errno(-errno, errno,
"Failed to attach to %s namespace of %d",
ns_info[i].proc_name,
ctx->init_pid);
if (close(ctx->ns_fd[i])) {
fret = -errno;
SYSERROR("Failed to close file descriptor for %s namespace",
ns_info[i].proc_name);
}
ctx->ns_fd[i] = -EBADF;
}
return fret;
}
static int attach_namespaces(struct attach_context *ctx,
lxc_attach_options_t *options)
{
if (lxc_log_trace()) {
for (int i = 0; i < LXC_NS_MAX; i++) {
if (ns_info[i].clone_flag & options->namespaces) {
TRACE("Attaching to %s namespace", ns_info[i].proc_name);
continue;
}
if (ns_info[i].clone_flag & ctx->ns_inherited) {
TRACE("Sharing %s namespace", ns_info[i].proc_name);
continue;
}
TRACE("Inheriting %s namespace", ns_info[i].proc_name);
}
}
if (ctx->init_pidfd < 0)
return __attach_namespaces_nsfd(ctx, options);
return __attach_namespaces_pidfd(ctx, options);
}
static void put_attach_context(struct attach_context *ctx)
{
if (ctx) {
if (!(ctx->attach_flags & LXC_ATTACH_LSM_LABEL))
free_disarm(ctx->lsm_label);
close_prot_errno_disarm(ctx->dfd_init_pid);
if (ctx->container) {
lxc_container_put(ctx->container);
ctx->container = NULL;
}
put_namespaces(ctx);
free(ctx);
}
}
/*
* Place anything in here that needs to be get rid of before we move into the
* container's context and fail hard if we can't.
*/
static bool attach_context_security_barrier(struct attach_context *ctx)
{
if (ctx) {
if (close(ctx->dfd_self_pid))
return false;
ctx->dfd_self_pid = -EBADF;
if (close(ctx->dfd_init_pid))
return false;
ctx->dfd_init_pid = -EBADF;
}
return true;
}
int lxc_attach_remount_sys_proc(void)
{
int ret;
ret = unshare(CLONE_NEWNS);
if (ret < 0)
return log_error_errno(-1, errno, "Failed to unshare mount namespace");
if (detect_shared_rootfs() && mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL))
SYSERROR("Failed to recursively turn root mount tree into dependent mount. Continuing...");
/* Assume /proc is always mounted, so remount it. */
ret = umount2("/proc", MNT_DETACH);
if (ret < 0)
return log_error_errno(-1, errno, "Failed to unmount /proc");
ret = mount("none", "/proc", "proc", 0, NULL);
if (ret < 0)
return log_error_errno(-1, errno, "Failed to remount /proc");
/*
* Try to umount /sys. If it's not a mount point, we'll get EINVAL, then
* we ignore it because it may not have been mounted in the first place.
*/
ret = umount2("/sys", MNT_DETACH);
if (ret < 0 && errno != EINVAL)
return log_error_errno(-1, errno, "Failed to unmount /sys");
/* Remount it. */
if (ret == 0 && mount("none", "/sys", "sysfs", 0, NULL))
return log_error_errno(-1, errno, "Failed to remount /sys");
return 0;
}
static int drop_capabilities(struct attach_context *ctx)
{
int last_cap;
last_cap = lxc_caps_last_cap();
for (int cap = 0; cap <= last_cap; cap++) {
if (ctx->capability_mask & (1LL << cap))
continue;
if (prctl(PR_CAPBSET_DROP, prctl_arg(cap), prctl_arg(0),
prctl_arg(0), prctl_arg(0)))
return log_error_errno(-1, errno, "Failed to drop capability %d", cap);
TRACE("Dropped capability %d", cap);
}
return 0;
}
static int lxc_attach_set_environment(struct attach_context *ctx,
enum lxc_attach_env_policy_t policy,
char **extra_env, char **extra_keep)
{
int ret;
struct lxc_list *iterator;
if (policy == LXC_ATTACH_CLEAR_ENV) {
int path_kept = 0;
char **extra_keep_store = NULL;
if (extra_keep) {
size_t count, i;
for (count = 0; extra_keep[count]; count++)
;
extra_keep_store = zalloc(count * sizeof(char *));
if (!extra_keep_store)
return -1;
for (i = 0; i < count; i++) {
char *v = getenv(extra_keep[i]);
if (v) {
extra_keep_store[i] = strdup(v);
if (!extra_keep_store[i]) {
while (i > 0)
free(extra_keep_store[--i]);
free(extra_keep_store);
return -1;
}
if (strequal(extra_keep[i], "PATH"))
path_kept = 1;
}
}
}
if (clearenv()) {
if (extra_keep_store) {
char **p;
for (p = extra_keep_store; *p; p++)
free(*p);
free(extra_keep_store);
}
return log_error(-1, "Failed to clear environment");
}
if (extra_keep_store) {
size_t i;
for (i = 0; extra_keep[i]; i++) {
if (extra_keep_store[i]) {
ret = setenv(extra_keep[i], extra_keep_store[i], 1);
if (ret < 0)
SYSWARN("Failed to set environment variable");
}
free(extra_keep_store[i]);
}
free(extra_keep_store);
}
/* Always set a default path; shells and execlp tend to be fine
* without it, but there is a disturbing number of C programs
* out there that just assume that getenv("PATH") is never NULL
* and then die a painful segfault death.
*/
if (!path_kept) {
ret = setenv("PATH", "/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin", 1);
if (ret < 0)
SYSWARN("Failed to set environment variable");
}
}
ret = putenv("container=lxc");
if (ret < 0)
return log_warn(-1, "Failed to set environment variable");
/* Set container environment variables.*/
if (ctx->container->lxc_conf) {
lxc_list_for_each(iterator, &ctx->container->lxc_conf->environment) {
char *env_tmp;
env_tmp = strdup((char *)iterator->elem);
if (!env_tmp)
return -1;
ret = putenv(env_tmp);
if (ret < 0)
return log_error_errno(-1, errno, "Failed to set environment variable: %s", (char *)iterator->elem);
}
}
/* Set extra environment variables. */
if (extra_env) {
for (; *extra_env; extra_env++) {
char *p;
/* We just assume the user knows what they are doing, so
* we don't do any checks.
*/
p = strdup(*extra_env);
if (!p)
return -1;
ret = putenv(p);
if (ret < 0)
SYSWARN("Failed to set environment variable");
}
}
return 0;
}
static char *lxc_attach_getpwshell(uid_t uid)
{
__do_free char *line = NULL, *result = NULL;
__do_fclose FILE *pipe_f = NULL;
int fd, ret;
pid_t pid;
int pipes[2];
bool found = false;
size_t line_bufsz = 0;
/* We need to fork off a process that runs the getent program, and we
* need to capture its output, so we use a pipe for that purpose.
*/
ret = pipe2(pipes, O_CLOEXEC);
if (ret < 0)
return NULL;
pid = fork();
if (pid < 0) {
close(pipes[0]);
close(pipes[1]);
return NULL;
}
if (!pid) {
char uid_buf[32];
char *arguments[] = {
"getent",
"passwd",
uid_buf,
NULL
};
close(pipes[0]);
/* We want to capture stdout. */
ret = dup2(pipes[1], STDOUT_FILENO);
close(pipes[1]);
if (ret < 0)
_exit(EXIT_FAILURE);
/* Get rid of stdin/stderr, so we try to associate it with
* /dev/null.
*/
fd = open_devnull();
if (fd < 0) {
close(STDIN_FILENO);
close(STDERR_FILENO);
} else {
(void)dup3(fd, STDIN_FILENO, O_CLOEXEC);
(void)dup3(fd, STDERR_FILENO, O_CLOEXEC);
close(fd);
}
/* Finish argument list. */
ret = strnprintf(uid_buf, sizeof(uid_buf), "%ld", (long)uid);
if (ret <= 0)
_exit(EXIT_FAILURE);
/* Try to run getent program. */
(void)execvp("getent", arguments);
_exit(EXIT_FAILURE);
}
close(pipes[1]);
pipe_f = fdopen(pipes[0], "re");
if (!pipe_f) {
close(pipes[0]);
goto reap_child;
}
/* Transfer ownership of pipes[0] to pipe_f. */
move_fd(pipes[0]);
while (getline(&line, &line_bufsz, pipe_f) != -1) {
int i;
long value;
char *token;
char *endptr = NULL, *saveptr = NULL;
/* If we already found something, just continue to read
* until the pipe doesn't deliver any more data, but
* don't modify the existing data structure.
*/
if (found)
continue;
if (!line)
continue;
/* Trim line on the right hand side. */
for (i = strlen(line); i > 0 && (line[i - 1] == '\n' || line[i - 1] == '\r'); --i)
line[i - 1] = '\0';
/* Split into tokens: first: user name. */
token = strtok_r(line, ":", &saveptr);
if (!token)
continue;