-
Notifications
You must be signed in to change notification settings - Fork 1
/
runguard.c
1427 lines (1212 loc) · 42 KB
/
runguard.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
/*
runguard -- run command with restrictions.
Part of the DOMjudge Programming Contest Jury System and licensed
under the GNU GPL. See README and COPYING for details.
Multiple minor improvements ported from the DOMjudge-ETH tree.
Based on an idea from the timeout program, written by Wietse Venema
as part of The Coroner's Toolkit.
Program specifications:
This program will run the specified command in a separate process
group (session) and apply the restrictions as specified after
forking, before executing the command.
The stdin and stdout streams are passed to the command and runguard
does not read or write to these. Error and verbose messages from
runguard are by default written to stderr, hence mixed with stderr
output of the command, unless that is optionally redirected to file.
The command and its children are sent a SIGTERM after the runtime
has passed, followed by a SIGKILL after 'killdelay'. The program is
considered to have finished when the main program thread exits. At
that time any children still running are killed.
*/
#include "config.h"
/* Some system/site specific config: VALID_USERS, CHROOT_PREFIX */
#include "runguard-config.h"
/* For chroot(), which is not POSIX. */
#define _DEFAULT_SOURCE
/* For unshare() used by cgroups. */
#define _GNU_SOURCE
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/param.h>
#include <sys/select.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <sys/times.h>
#include <sys/resource.h>
#include <sys/types.h>
#include <errno.h>
#include <fcntl.h>
#include <signal.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <stdarg.h>
#include <stdio.h>
#include <getopt.h>
#include <fnmatch.h>
#include <regex.h>
#include <pwd.h>
#include <grp.h>
#include <time.h>
#include <math.h>
#include <limits.h>
#include <inttypes.h>
#include <libcgroup.h>
#include <sched.h>
#include <sys/sysinfo.h>
#define PROGRAM "runguard"
#define VERSION DOMJUDGE_VERSION "/" REVISION
#define max(x,y) ((x) > (y) ? (x) : (y))
#define min(x,y) ((x) < (y) ? (x) : (y))
/* Array indices for input/output file descriptors as used by pipe() */
#define PIPE_IN 1
#define PIPE_OUT 0
#define BUF_SIZE 4*1024
/* Types of time for writing to file. */
#define WALL_TIME_TYPE 0
#define CPU_TIME_TYPE 1
/* Strings to write to file when exceeding no/soft/hard/both limits. */
const char output_timelimit_str[4][16] = {
"",
"soft-timelimit",
"hard-timelimit",
"hard-timelimit"
};
/* Bitmask of soft/hard timelimit (used in array above and
* {wall,cpu}timelimit_reached variabled below). */
const int soft_timelimit = 1;
const int hard_timelimit = 2;
const struct timespec killdelay = { 0, 100000000L }; /* 0.1 seconds */
const struct timespec cg_delete_delay = { 0, 10000000L }; /* 0.01 seconds */
extern int errno;
#ifndef _GNU_SOURCE
extern char **environ;
#endif
const int exit_failure = -1;
char *progname;
char *cmdname;
char **cmdargs;
char *rootdir;
char *stdoutfilename;
char *stderrfilename;
char *metafilename;
char *environment_variables;
FILE *metafile;
char cgroupname[255];
const char *cpuset;
/* Linux Out-Of-Memory adjustment for current process. */
#define OOM_PATH_NEW "/proc/self/oom_score_adj"
#define OOM_PATH_OLD "/proc/self/oom_adj"
#define OOM_RESET_VALUE 0
char *runuser;
char *rungroup;
int runuid;
int rungid;
int use_root;
int use_walltime;
int use_cputime;
int use_user;
int use_group;
int redir_stdout;
int redir_stderr;
int limit_streamsize;
int outputmeta;
int outputtimetype;
int no_coredump;
int preserve_environment;
int be_verbose;
int be_quiet;
int show_help;
int show_version;
double walltimelimit[2], cputimelimit[2]; /* in seconds, soft and hard limits */
int walllimit_reached, cpulimit_reached; /* 1=soft, 2=hard, 3=both limits reached */
int64_t memsize;
rlim_t filesize;
rlim_t nproc;
size_t streamsize;
int use_splice;
pid_t child_pid = -1;
static volatile sig_atomic_t received_SIGCHLD = 0;
static volatile sig_atomic_t received_signal = -1;
FILE *child_stdout;
FILE *child_stderr;
int child_pipefd[3][2];
int child_redirfd[3];
struct timeval progstarttime, starttime, endtime;
struct tms startticks, endticks;
struct option const long_opts[] = {
{"root", required_argument, NULL, 'r'},
{"user", required_argument, NULL, 'u'},
{"group", required_argument, NULL, 'g'},
{"walltime", required_argument, NULL, 't'},
{"cputime", required_argument, NULL, 'C'},
{"memsize", required_argument, NULL, 'm'},
{"filesize", required_argument, NULL, 'f'},
{"nproc", required_argument, NULL, 'p'},
{"cpuset", required_argument, NULL, 'P'},
{"no-core", no_argument, NULL, 'c'},
{"stdout", required_argument, NULL, 'o'},
{"stderr", required_argument, NULL, 'e'},
{"streamsize", required_argument, NULL, 's'},
{"environment",no_argument, NULL, 'E'},
{"variable", required_argument, NULL, 'V'},
{"outmeta", required_argument, NULL, 'M'},
{"verbose", no_argument, NULL, 'v'},
{"quiet", no_argument, NULL, 'q'},
{"help", no_argument, &show_help, 1 },
{"version", no_argument, &show_version, 1 },
{ NULL, 0, NULL, 0 }
};
void warning( const char *, ...) __attribute__((format (printf, 1, 2)));
void verbose( const char *, ...) __attribute__((format (printf, 1, 2)));
void error(int, const char *, ...) __attribute__((format (printf, 2, 3)));
void write_meta(const char*, const char *, ...) __attribute__((format (printf, 2, 3)));
void warning(const char *format, ...)
{
va_list ap;
va_start(ap,format);
if ( ! be_quiet ) {
fprintf(stderr,"%s: warning: ",progname);
vfprintf(stderr,format,ap);
fprintf(stderr,"\n");
}
va_end(ap);
}
void verbose(const char *format, ...)
{
va_list ap;
va_start(ap,format);
struct timeval currtime;
double runtime;
if ( ! be_quiet && be_verbose ) {
gettimeofday(&currtime,NULL);
runtime = (currtime.tv_sec - progstarttime.tv_sec ) +
(currtime.tv_usec - progstarttime.tv_usec)*1E-6;
fprintf(stderr,"%s [%d @ %10.6lf]: verbose: ",progname,getpid(),runtime);
vfprintf(stderr,format,ap);
fprintf(stderr,"\n");
}
va_end(ap);
}
void error(int errnum, const char *format, ...)
{
va_list ap;
va_start(ap,format);
sigset_t sigs;
char *errstr;
int errlen, errpos;
/*
* Make sure the signal handler for these (terminate()) does not
* interfere, we are exiting now anyway.
*/
sigaddset(&sigs, SIGALRM);
sigaddset(&sigs, SIGTERM);
sigprocmask(SIG_BLOCK, &sigs, NULL);
/* First print to string to be able to reuse the message. */
errlen = strlen(progname)+255;
if ( format!=NULL ) errlen += strlen(format);
errstr = (char *)malloc(errlen);
if ( errstr==NULL ) abort();
sprintf(errstr,"%s",progname);
errpos = strlen(errstr);
if ( format!=NULL ) {
snprintf(errstr+errpos,errlen-errpos,": ");
errpos += 2;
vsnprintf(errstr+errpos,errlen-errpos,format,ap);
errpos += strlen(errstr+errpos);
}
if ( errnum!=0 ) {
/* Special case libcgroup error codes. */
if ( errnum==ECGOTHER ) {
snprintf(errstr+errpos,errlen-errpos,": libcgroup");
errpos += strlen(errstr+errpos);
errnum = errno;
}
if ( errnum>=ECGROUPNOTCOMPILED && errnum<=ECGROUPNOTCOMPILED ) {
snprintf(errstr+errpos,errlen-errpos,": %s",cgroup_strerror(errnum));
} else {
snprintf(errstr+errpos,errlen-errpos,": %s",strerror(errnum));
}
errpos += strlen(errstr+errpos);
}
if ( format==NULL && errnum==0 ) {
snprintf(errstr+errpos,errlen-errpos,": unknown error");
}
fprintf(stderr,"%s\nTry `%s --help' for more information.\n",errstr,progname);
va_end(ap);
write_meta("internal-error","%s",errstr);
if ( outputmeta && metafile != NULL && fclose(metafile)!=0 ) {
fprintf(stderr,"\nError writing to metafile '%s'.\n",metafilename);
}
/* Make sure that all children are killed before terminating */
if ( child_pid > 0) {
verbose("sending SIGKILL");
if ( kill(-child_pid,SIGKILL)!=0 && errno!=ESRCH ) {
fprintf(stderr,"unable to send SIGKILL to children while terminating "
"due to previous error: %s\n", strerror(errno));
/*
* continue, there is not much we can do here.
* In the worst case, this will trigger an error
* in testcase_run.sh, as the runuser may still be
* running processes
*/
}
/* Wait a while to make sure the process is killed by now. */
nanosleep(&killdelay,NULL);
}
exit(exit_failure);
}
void write_meta(const char *key, const char *format, ...)
{
va_list ap;
if ( !outputmeta ) return;
va_start(ap,format);
if ( fprintf(metafile,"%s: ",key)<=0 ) {
error(0,"cannot write to file `%s'",metafilename);
}
if ( vfprintf(metafile,format,ap)<0 ) {
error(0,"cannot write to file `%s'(vfprintf)",metafilename);
}
if ( fprintf(metafile,"\n")<=0 ) {
error(0,"cannot write to file `%s'",metafilename);
}
va_end(ap);
}
void version(const char *prog, const char *vers)
{
printf("\
%s -- part of DOMjudge version %s\n\
Written by the DOMjudge developers\n\n\
DOMjudge comes with ABSOLUTELY NO WARRANTY. This is free software, and you\n\
are welcome to redistribute it under certain conditions. See the GNU\n\
General Public Licence for details.\n", prog, vers);
exit(0);
}
void usage()
{
printf("\
Usage: %s [OPTION]... COMMAND...\n\
Run COMMAND with restrictions.\n\
\n", progname);
printf("\
-r, --root=ROOT run COMMAND with root directory set to ROOT\n\
-u, --user=USER run COMMAND as user with username or ID USER\n\
-g, --group=GROUP run COMMAND under group with name or ID GROUP\n\
-t, --walltime=TIME kill COMMAND after TIME wallclock seconds\n\
-C, --cputime=TIME set maximum CPU time to TIME seconds\n\
-m, --memsize=SIZE set total memory limit to SIZE kB\n\
-f, --filesize=SIZE set maximum created filesize to SIZE kB;\n");
printf("\
-p, --nproc=N set maximum no. processes to N\n\
-P, --cpuset=ID use only processor number ID (or set, e.g. \"0,2-3\")\n\
-c, --no-core disable core dumps\n\
-o, --stdout=FILE redirect COMMAND stdout output to FILE\n\
-e, --stderr=FILE redirect COMMAND stderr output to FILE\n\
-s, --streamsize=SIZE truncate COMMAND stdout/stderr streams at SIZE kB\n\
-E, --environment preserve environment variables (default only PATH)\n\
-V, --variable add additonal environment variables (in form KEY=VALUE;KEY2=VALUE2)\n\
-M, --outmeta=FILE write metadata (runtime, exitcode, etc.) to FILE\n");
printf("\
-v, --verbose display some extra warnings and information\n\
-q, --quiet suppress all warnings and verbose output\n\
--help display this help and exit\n\
--version output version information and exit\n");
printf("\n\
Note that root privileges are needed for the `root' and `user' options.\n\
If `user' is set, then `group' defaults to the same to prevent security\n\
issues, since otherwise the process would retain group root permissions.\n\
Additionally, Linux cgroup support is required for the `memsize' and\n\
`cputime' options, and to report actual memory usage.\n\
The COMMAND path is relative to the changed ROOT directory if specified.\n\
TIME may be specified as a float; two floats separated by `:' are treated\n\
as soft and hard limits. The runtime written to file is that of the last\n\
of wall/cpu time options set, and defaults to CPU time when neither is set.\n\
When run setuid without the `user' option, the user ID is set to the\n\
real user ID.\n");
exit(0);
}
void output_exit_time(int exitcode, double cpudiff)
{
double walldiff, userdiff, sysdiff;
int timelimit_reached = 0;
unsigned long ticks_per_second = sysconf(_SC_CLK_TCK);
verbose("command exited with exitcode %d",exitcode);
write_meta("exitcode","%d",exitcode);
if (received_signal != -1) {
write_meta("signal", "%d", received_signal);
}
walldiff = (endtime.tv_sec - starttime.tv_sec ) +
(endtime.tv_usec - starttime.tv_usec)*1E-6;
userdiff = (double)(endticks.tms_cutime - startticks.tms_cutime) / ticks_per_second;
sysdiff = (double)(endticks.tms_cstime - startticks.tms_cstime) / ticks_per_second;
write_meta("wall-time","%.3f", walldiff);
write_meta("user-time","%.3f", userdiff);
write_meta("sys-time", "%.3f", sysdiff);
write_meta("cpu-time", "%.3f", cpudiff);
verbose("runtime is %.3f seconds real, %.3f user, %.3f sys",
walldiff, userdiff, sysdiff);
if ( use_walltime && walldiff > walltimelimit[0] ) {
walllimit_reached |= soft_timelimit;
warning("timelimit exceeded (soft wall time)");
}
if ( use_cputime && cpudiff > cputimelimit[0] ) {
cpulimit_reached |= soft_timelimit;
warning("timelimit exceeded (soft cpu time)");
}
switch ( outputtimetype ) {
case WALL_TIME_TYPE:
write_meta("time-used","wall-time");
timelimit_reached = walllimit_reached;
break;
case CPU_TIME_TYPE:
write_meta("time-used","cpu-time");
timelimit_reached = cpulimit_reached;
break;
default:
error(0,"cannot write unknown time type `%d' to file",outputtimetype);
}
/* Hard limitlimit reached always has precedence. */
if ( (walllimit_reached | cpulimit_reached) & hard_timelimit ) {
timelimit_reached |= hard_timelimit;
}
write_meta("time-result","%s",output_timelimit_str[timelimit_reached]);
}
/* Return whether we need to use cgroups. This is checked in the
* cgroup_* functions below. If not used they return without
* performing any action.
*/
int use_cgroup()
{
return use_cputime || memsize!=RLIM_INFINITY ||
( cpuset!=NULL && strlen(cpuset)>0 );
}
void output_cgroup_stats(double *cputime)
{
int ret;
int64_t max_usage, cpu_time_int;
struct cgroup *cg;
struct cgroup_controller *cg_controller;
if ( !use_cgroup() ) return;
if ( (cg = cgroup_new_cgroup(cgroupname))==NULL ) error(0,"cgroup_new_cgroup");
if ((ret = cgroup_get_cgroup(cg)) != 0) error(ret,"get cgroup information");
cg_controller = cgroup_get_controller(cg, "memory");
ret = cgroup_get_value_int64(cg_controller, "memory.memsw.max_usage_in_bytes", &max_usage);
if ( ret!=0 ) error(ret,"get cgroup value memory.memsw.max_usage_in_bytes");
verbose("total memory used: %" PRId64 " kB", max_usage/1024);
write_meta("memory-bytes","%" PRId64, max_usage);
cg_controller = cgroup_get_controller(cg, "cpuacct");
ret = cgroup_get_value_int64(cg_controller, "cpuacct.usage", &cpu_time_int);
if ( ret!=0 ) error(ret,"get cgroup value cpuacct.usage");
*cputime = (double) cpu_time_int / 1.e9;
cgroup_free(&cg);
}
/* Temporary shorthand define for error handling. */
#define cgroup_add_value(type,name,value) \
ret = cgroup_add_value_ ## type(cg_controller, name, value); \
if ( ret!=0 ) error(ret,"set cgroup value " #name);
void cgroup_create()
{
int ret;
struct cgroup *cg;
struct cgroup_controller *cg_controller;
if ( !use_cgroup() ) return;
cg = cgroup_new_cgroup(cgroupname);
if (!cg) error(0,"cgroup_new_cgroup");
/* Set up the memory restrictions; these two options limit ram use
and ram+swap use. They are the same so no swapping can occur */
if ( (cg_controller = cgroup_add_controller(cg, "memory"))==NULL ) {
error(0,"cgroup_add_controller memory");
}
cgroup_add_value(int64, "memory.limit_in_bytes", memsize);
cgroup_add_value(int64, "memory.memsw.limit_in_bytes", memsize);
/* Set up cpu restrictions; we pin the task to a specific set of
cpus. We also give it exclusive access to those cores, and set
no limits on memory nodes */
if ( cpuset!=NULL && strlen(cpuset)>0 ) {
if ( (cg_controller = cgroup_add_controller(cg, "cpuset"))==NULL ) {
error(0,"cgroup_add_controller cpuset");
}
/* To make a cpuset exclusive, some additional setup outside of domjudge is
required, so for now, we will leave this commented out. */
/* cgroup_add_value_int64(cg_controller, "cpuset.cpu_exclusive", 1); */
cgroup_add_value(string, "cpuset.mems", "0");
cgroup_add_value(string, "cpuset.cpus", cpuset);
} else {
verbose("cpuset undefined");
}
if ( (cg_controller = cgroup_add_controller(cg, "cpuacct"))==NULL ) {
error(0,"cgroup_add_controller cpuacct");
}
/* Perform the actual creation of the cgroup */
if ( (ret = cgroup_create_cgroup(cg, 1))!=0 ) error(ret,"creating cgroup");
cgroup_free(&cg);
verbose("created cgroup '%s'",cgroupname);
}
#undef cgroup_setval
void cgroup_attach()
{
int ret;
struct cgroup *cg;
if ( !use_cgroup() ) return;
cg = cgroup_new_cgroup(cgroupname);
if (!cg) error(0,"cgroup_new_cgroup");
if ( (ret = cgroup_get_cgroup(cg))!=0 ) error(ret,"get cgroup information");
/* Attach task to the cgroup */
if ( (ret = cgroup_attach_task(cg))!=0 ) error(ret,"attach task to cgroup");
cgroup_free(&cg);
}
void cgroup_kill()
{
int ret;
void *handle = NULL;
pid_t pid;
if ( !use_cgroup() ) return;
/* kill any remaining tasks, and wait for them to be gone */
while(1) {
ret = cgroup_get_task_begin(cgroupname, "memory", &handle, &pid);
cgroup_get_task_end(&handle);
if (ret == ECGEOF) break;
kill(pid, SIGKILL);
}
}
void cgroup_delete()
{
int ret;
struct cgroup *cg;
if ( !use_cgroup() ) return;
cg = cgroup_new_cgroup(cgroupname);
if (!cg) error(0,"cgroup_new_cgroup");
if ( cgroup_add_controller(cg, "cpuacct")==NULL ) error(0,"cgroup_add_controller cpuacct");
if ( cgroup_add_controller(cg, "memory")==NULL ) error(0,"cgroup_add_controller memory");
if ( cpuset!=NULL && strlen(cpuset)>0 ) {
if ( cgroup_add_controller(cg, "cpuset")==NULL ) error(0,"cgroup_add_controller cpuset");
}
/* Clean up our cgroup */
nanosleep(&cg_delete_delay,NULL);
ret = cgroup_delete_cgroup_ext(cg, CGFLAG_DELETE_IGNORE_MIGRATION | CGFLAG_DELETE_RECURSIVE);
if ( ret!=0 ) error(ret,"deleting cgroup");
cgroup_free(&cg);
verbose("deleted cgroup '%s'",cgroupname);
}
void terminate(int sig)
{
struct sigaction sigact;
/* Reset signal handlers to default */
sigact.sa_handler = SIG_DFL;
sigact.sa_flags = 0;
if ( sigemptyset(&sigact.sa_mask)!=0 ) {
warning("could not initialize signal mask");
}
if ( sigaction(SIGTERM,&sigact,NULL)!=0 ) {
warning("could not restore signal handler");
}
if ( sigaction(SIGALRM,&sigact,NULL)!=0 ) {
warning("could not restore signal handler");
}
if ( sig==SIGALRM ) {
walllimit_reached |= hard_timelimit;
warning("timelimit exceeded (hard wall time): aborting command");
} else {
warning("received signal %d: aborting command",sig);
}
received_signal = sig;
/* First try to kill graciously, then hard.
Don't report an already exited process as error. */
verbose("sending SIGTERM");
if ( kill(-child_pid,SIGTERM)!=0 && errno!=ESRCH ) {
error(errno,"sending SIGTERM to command");
}
/* Prefer nanosleep over sleep because of higher resolution and
it does not interfere with signals. */
nanosleep(&killdelay,NULL);
verbose("sending SIGKILL");
if ( kill(-child_pid,SIGKILL)!=0 && errno!=ESRCH ) {
error(errno,"sending SIGKILL to command");
}
/* Wait another while to make sure the process is killed by now. */
nanosleep(&killdelay,NULL);
}
static void child_handler(int sig)
{
received_SIGCHLD = 1;
}
int userid(char *name)
{
struct passwd *pwd;
errno = 0; /* per the linux GETPWNAM(3) man-page */
pwd = getpwnam(name);
if ( pwd==NULL || errno ) return -1;
return (int) pwd->pw_uid;
}
int groupid(char *name)
{
struct group *grp;
errno = 0; /* per the linux GETGRNAM(3) man-page */
grp = getgrnam(name);
if ( grp==NULL || errno ) return -1;
return (int) grp->gr_gid;
}
long read_optarg_int(const char *desc, long minval, long maxval)
{
long arg;
char *ptr;
arg = strtol(optarg,&ptr,10);
if ( errno || *ptr!='\0' || arg<minval || arg>maxval ) {
error(errno,"invalid %s specified: `%s'",desc,optarg);
}
return arg;
}
void read_optarg_time(const char *desc, double *times)
{
char *optcopy, *ptr, *sep;
if ( (optcopy=strdup(optarg))==NULL ) error(0,"strdup() failed");
/* Check for soft:hard limit separator and cut string. */
if ( (sep=strchr(optcopy,':'))!=NULL ) *sep = 0;
times[0] = strtod(optcopy,&ptr);
if ( errno || *ptr!='\0' || !finite(times[0]) || times[0]<=0 ) {
error(errno,"invalid %s specified: `%s'",desc,optarg);
}
/* And repeat for hard limit if we found the ':' separator. */
if ( sep!=NULL ) {
times[1] = strtod(sep+1,&ptr);
if ( errno || *ptr!='\0' || !finite(times[1]) || times[1]<=0 ) {
error(errno,"invalid %s specified: `%s'",desc,optarg);
}
if ( times[1]<times[0] ) {
error(0,"invalid %s specified: hard limit is lower than soft limit",desc);
}
} else {
/* Set soft and hard limits equal. */
times[1] = times[0];
}
free(optcopy);
}
void setrestrictions()
{
char *path;
char cwd[PATH_MAX+1];
gid_t aux_groups[10];
struct rlimit lim;
/* Clear environment to prevent all kinds of security holes, save PATH */
if ( !preserve_environment ) {
path = getenv("PATH");
environ[0] = NULL;
/* FIXME: Clean path before setting it again? */
if ( path!=NULL ) setenv("PATH",path,1);
}
/* Set additional environment variables. */
if (environment_variables != NULL) {
char *token = strtok(environment_variables, ";");
while (token != NULL) {
verbose("setting environment variable: %s", token);
putenv(token);
token = strtok(NULL, ";");
}
}
/* Set resource limits: must be root to raise hard limits.
Note that limits can thus be raised from the systems defaults! */
/* First define shorthand macro function */
#define setlim(type) \
if ( setrlimit(RLIMIT_ ## type, &lim)!=0 ) { \
if ( errno==EPERM ) { \
warning("no permission to set resource RLIMIT_" #type); \
} else { \
error(errno,"setting resource RLIMIT_" #type); \
} \
}
if ( use_cputime ) {
/* The CPU-time resource limit can only be specified in
seconds, so round up: we can measure actual CPU time used
more accurately. Also set the real hard limit one second
higher: at the soft limit the kernel will send SIGXCPU at
the hard limit a SIGKILL. The SIGXCPU can be caught, but is
not by default and gives us a reliable way to detect if the
CPU-time limit was reached. */
rlim_t cputime_limit = (rlim_t)ceil(cputimelimit[1]);
verbose("setting hard CPU-time limit to %d(+1) seconds",(int)cputime_limit);
lim.rlim_cur = cputime_limit;
lim.rlim_max = cputime_limit+1;
setlim(CPU);
}
/* Memory limits should be unlimited, since we use cgroups. */
lim.rlim_cur = lim.rlim_max = RLIM_INFINITY;
setlim(AS);
setlim(DATA);
/* Always set the stack size to be unlimited. */
lim.rlim_cur = lim.rlim_max = RLIM_INFINITY;
setlim(STACK);
if ( filesize!=RLIM_INFINITY ) {
verbose("setting filesize limit to %d bytes",(int)filesize);
lim.rlim_cur = lim.rlim_max = filesize;
setlim(FSIZE);
}
if ( nproc!=RLIM_INFINITY ) {
verbose("setting process limit to %d",(int)nproc);
lim.rlim_cur = lim.rlim_max = nproc;
setlim(NPROC);
}
#undef setlim
if ( no_coredump ) {
verbose("disabling core dumps");
lim.rlim_cur = lim.rlim_max = 0;
if ( setrlimit(RLIMIT_CORE,&lim)!=0 ) error(errno,"disabling core dumps");
}
/* Put the child process in the cgroup */
cgroup_attach();
/* Run the command in a separate process group so that the command
and all its children can be killed off with one signal. */
if ( setsid()==-1 ) error(errno,"setsid failed");
/* Set root-directory and change directory to there. */
if ( use_root ) {
/* Small security issue: when running setuid-root, people can find
out which directories exist from error message. */
if ( chdir(rootdir)!=0 ) error(errno,"cannot chdir to `%s'",rootdir);
/* Get absolute pathname of rootdir, by reading it. */
if ( getcwd(cwd,PATH_MAX)==NULL ) error(errno,"cannot get directory");
if ( cwd[strlen(cwd)-1]!='/' ) strcat(cwd,"/");
/* Canonicalize CHROOT_PREFIX. */
if ( (path = (char *) malloc(PATH_MAX+1))==NULL ) {
error(errno,"allocating memory");
}
if ( realpath(CHROOT_PREFIX,path)==NULL ) {
error(errno,"cannot canonicalize path '%s'",CHROOT_PREFIX);
}
/* Check that we are within prescribed path. */
if ( strncmp(cwd,path,strlen(path))!=0 ) {
error(0,"invalid root: must be within `%s'",path);
}
free(path);
if ( chroot(".")!=0 ) error(errno,"cannot change root to `%s'",cwd);
/* Just to make sure and satisfy Coverity scan: */
if ( chdir("/")!=0 ) error(errno,"cannot chdir to `/' in chroot");
verbose("using root-directory `%s'",cwd);
}
/* Set group-id (must be root for this, so before setting user). */
if ( use_group ) {
if ( setgid(rungid) ) error(errno,"cannot set group ID to `%d'",rungid);
aux_groups[0] = rungid;
if ( setgroups(1, aux_groups) ) error(errno,"cannot clear auxiliary groups");
verbose("using group ID `%d'",rungid);
}
/* Set user-id (must be root for this). */
if ( use_user ) {
if ( setuid(runuid) ) error(errno,"cannot set user ID to `%d'",runuid);
verbose("using user ID `%d' for command",runuid);
} else {
/* Permanently reset effective uid to real uid, to prevent
child command from having root privileges.
Note that this means that the child runs as the same user
as the watchdog process and can thus manipulate it, e.g. by
sending SIGSTOP/SIGCONT! */
if ( setuid(getuid()) ) error(errno,"cannot reset real user ID");
verbose("reset user ID to `%d' for command",getuid());
}
if ( geteuid()==0 || getuid()==0 ) {
error(0,"root privileges not dropped. Do not run judgedaemon as root.");
}
}
void pump_pipes(fd_set* readfds, size_t data_read[], size_t data_passed[])
{
char buf[BUF_SIZE];
ssize_t nread, nwritten;
size_t to_read, to_write;
int i;
/* Check to see if data is available and pass it on */
for(i=1; i<=2; i++) {
if ( child_pipefd[i][PIPE_OUT] != -1 &&
FD_ISSET(child_pipefd[i][PIPE_OUT], readfds) ) {
if (limit_streamsize && data_passed[i] == streamsize) {
/* Throw away data if we're at the output limit, but
still count how much data we consumed */
nread = read(child_pipefd[i][PIPE_OUT], buf, BUF_SIZE);
} else {
/* Otherwise copy the output to a file */
to_read = BUF_SIZE;
if (limit_streamsize) {
to_read = min(BUF_SIZE, streamsize-data_passed[i]);
}
if ( use_splice ) {
nread = splice(child_pipefd[i][PIPE_OUT], NULL,
child_redirfd[i], NULL,
to_read, SPLICE_F_MOVE | SPLICE_F_NONBLOCK);
if ( nread==-1 && errno==EINVAL ) {
use_splice = 0;
verbose("splice failed, switching to read/write");
/* Setting errno here to repeat the copy. */
errno = EAGAIN;
}
} else {
nread = read(child_pipefd[i][PIPE_OUT], buf, to_read);
if ( nread>0 ) {
to_write = nread;
while ( to_write>0 ) {
nwritten = write(child_redirfd[i], buf, to_write);
if ( nwritten==-1 ) {
nread = -1;
break;
}
to_write -= nwritten;
}
}
}
if ( nread>0 ) data_passed[i] += nread;
/* print message if we're at the streamsize limit */
if (limit_streamsize && data_passed[i] == streamsize) {
verbose("child fd %i limit reached",i);
}
}
if ( nread==-1 ) {
if (errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK) continue;
error(errno,"copying data fd %d",i);
}
if ( nread==0 ) {
/* EOF detected: close fd and indicate this with -1 */
if ( close(child_pipefd[i][PIPE_OUT])!=0 ) {
error(errno,"closing pipe for fd %d",i);
}
child_pipefd[i][PIPE_OUT] = -1;
continue;
}
data_read[i] += nread;
}
}
}
int main(int argc, char **argv)
{
sigset_t sigmask, emptymask;
fd_set readfds;
pid_t pid;
int i, r, nfds;
int ret;
FILE *fp;
char *oom_path;
int status;
int exitcode;
char *valid_users;
char *ptr;
regex_t userregex;
int opt;
double tmpd;
size_t data_read[3];
size_t data_passed[3];
size_t total_data;
char str[256];
struct itimerval itimer;
struct sigaction sigact;
progname = argv[0];
if ( gettimeofday(&progstarttime,NULL) ) error(errno,"getting time");
/* Parse command-line options */
use_root = use_walltime = use_cputime = use_user = no_coredump = 0;
outputmeta = walllimit_reached = cpulimit_reached = 0;
outputtimetype = CPU_TIME_TYPE;
preserve_environment = 0;
memsize = filesize = nproc = RLIM_INFINITY;
redir_stdout = redir_stderr = limit_streamsize = 0;
be_verbose = be_quiet = 0;
show_help = show_version = 0;
opterr = 0;
while ( (opt = getopt_long(argc,argv,"+r:u:g:t:C:m:f:p:P:co:e:s:EV:M:vq",long_opts,(int *) 0))!=-1 ) {
switch ( opt ) {
case 0: /* long-only option */
break;
case 'r': /* rootdir option */
use_root = 1;
rootdir = (char *) malloc(strlen(optarg)+2);
if ( rootdir==NULL ) error(errno,"allocating memory");
strcpy(rootdir,optarg);
break;
case 'u': /* user option: uid or string */
use_user = 1;
runuid = strtol(optarg,&ptr,10);
runuser = strdup(optarg);
if ( errno || *ptr!='\0' ) {
runuid = userid(optarg);
if ( regcomp(&userregex,"^[A-Za-z][A-Za-z0-9\\._-]*$", REG_NOSUB)!=0 ) {
error(0,"could not create username regex");
}
if ( regexec(&userregex, runuser, 0, NULL, 0)!=0 ) {
error(0,"username `%s' does not match POSIX pattern", runuser);
}
}
if ( runuid<0 ) error(0,"invalid username or ID specified: `%s'",optarg);
break;
case 'g': /* group option: gid or string */
use_group = 1;