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
|
/*
* Oracle Linux DTrace.
* Copyright (c) 2016, 2023, Oracle and/or its affiliates. All rights reserved.
* Licensed under the Universal Permissive License v 1.0 as shown at
* http://oss.oracle.com/licenses/upl.
*/
/* We define this here to stop pcap from including a conflicting bpf.h. */
#define _BPF_H_
#include <errno.h>
#include <fcntl.h>
#include <grp.h>
#include <pcap.h>
#include <poll.h>
#include <port.h>
#include <pthread.h>
#include <signal.h>
#include <stddef.h>
#include <stdio.h>
#include <string.h>
#include <sys/sysinfo.h>
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
#include <dt_pcap.h>
#include <dt_impl.h>
typedef struct dt_pcap {
dt_list_t dpc_list;
pthread_mutex_t dpc_lock;
dtrace_hdl_t *dpc_hdl;
char *dpc_filename;
FILE *dpc_outpipe;
uint64_t dpc_linktype;
uint32_t dpc_maxlen;
uint64_t dpc_boottime; /* boottime in seconds since epoch */
pcap_t *dpc_pcap;
pcap_dumper_t *dpc_pcap_dump;
} dt_pcap_t;
dt_pcap_t *
dt_pcap_create(dtrace_hdl_t *dtp, const char *filename, uint32_t maxlen)
{
dt_pcap_t *dpc;
struct sysinfo info;
dpc = dt_zalloc(dtp, sizeof(dt_pcap_t));
if (dpc == NULL) {
dt_set_errno(dtp, ENOMEM);
return NULL;
}
dpc->dpc_filename = strdup(filename);
dpc->dpc_maxlen = maxlen;
dpc->dpc_outpipe = NULL;
dpc->dpc_pcap = NULL;
dpc->dpc_pcap_dump = NULL;
/*
* Times collected in-kernel are relative to boot-time; for wireshark
* we need to adjust times relative to the epoch. We calculate boot
* time in seconds since the epoch here; we can then add it to
* the capture times we record.
*/
if (sysinfo(&info) == 0)
dpc->dpc_boottime = time(NULL) - info.uptime;
pthread_mutex_init(&dpc->dpc_lock, NULL);
dt_list_append(&dtp->dt_pcap.dt_pcaps, dpc);
return dpc;
}
void
dt_pcap_destroy(dtrace_hdl_t *dtp)
{
dt_pcap_t *dpc, *npc;
/* flush all captures; happens after tracing is stopped. */
for (dpc = dt_list_next(&dtp->dt_pcap.dt_pcaps); dpc != NULL; dpc = npc) {
npc = dt_list_next(dpc);
dt_list_delete(&dtp->dt_pcap.dt_pcaps, dpc);
if (dpc->dpc_pcap_dump != NULL) {
pcap_dump_flush(dpc->dpc_pcap_dump);
pcap_dump_close(dpc->dpc_pcap_dump);
}
if (dpc->dpc_pcap != NULL)
pcap_close(dpc->dpc_pcap);
pthread_mutex_destroy(&dpc->dpc_lock);
free(dpc->dpc_filename);
dt_free(dtp, dpc);
}
if (dtp->dt_pcap.dt_pcap_pid > 0) {
/*
* tshark will now print any remaining output and die. Wait for
* it to do so.
*/
close(dtp->dt_pcap.dt_pcap_pipe[1]);
pthread_join(dtp->dt_pcap.dt_pcap_output, NULL);
while (waitpid(dtp->dt_pcap.dt_pcap_pid, NULL, 0) < 0 &&
errno == EINTR);
}
}
/*
* The thread that emits output from tshark does nothing but printf()s in a
* tight loop: stdio ensures that output is nonoverlapping.
*/
static void *
dt_pcap_print(void *d)
{
dtrace_hdl_t *dtp = d;
FILE *inpipe;
char *line = NULL;
size_t len = 0;
inpipe = fdopen(dtp->dt_pcap.dt_pcap_pipe[0], "r");
if (inpipe == NULL) {
fprintf(stderr, "Cannot open tshark output pipe: %s\n",
strerror(errno));
return NULL;
}
while (getline(&line, &len, inpipe) >= 0) {
dt_printf(dtp, dtp->dt_pcap.dt_pcap_out_fp,
"pcap \t%s", line);
}
free(line);
fclose(inpipe);
dtp->dt_pcap.dt_pcap_pid = -1;
return NULL;
}
/*
* The file used to write packet capture data to is either specified via
* dt_freopen_filename (the result of calling freopen()); or the file is an
* open fd constituting the input end of a pipe. If fifo creation or
* invoking tshark fail, we set the dt_pcap_pid value to -1; this prevents us
* from retrying these failing operations. If no file is specified, and a fifo
* cannot be created for tshark, we return a NULL filename which signifies the
* caller should fall back to tracemem()-like display of packet data.
*
* We indicate to the caller that an open fd is in use by passing back an empty
* string. (This is disgusting, but it's simpler than most alternatives, and
* means that the relevant knowledge can be localized to this file.)
*/
const char *
dt_pcap_filename(dtrace_hdl_t *dtp, FILE *fp)
{
int status;
pid_t pid;
int pipe_in[2];
int pipe_out[2];
int pipe_err[2];
if (dtp->dt_freopen_filename != NULL &&
strcmp(dtp->dt_freopen_filename, DT_FREOPEN_RESTORE) != 0)
return dtp->dt_freopen_filename;
if (dtp->dt_pcap.dt_pcap_pid < 0) {
/*
* dt_pcap_pid == -1 means we could not create the pipe or run
* tshark on a previous attempt; we return NULL which results in
* the caller falling back to using tracemem() output of
* captured data.
*/
return NULL;
} else if (dtp->dt_pcap.dt_pcap_pid > 0) {
/* tshark is running, return the magic we-are-using-a-pipe
* value. */
return "";
}
/*
* We have not previously attempted to set up a named pipe and invoke
* tshark; let's try it now.
*/
dtp->dt_pcap.dt_pcap_pid = -1;
/*
* First, is tshark available? Saves us making assumptions about path.
*/
status = system("tshark -v >/dev/null 2>&1");
if (status < 0 || WEXITSTATUS(status) != 0)
return NULL;
if (pipe(pipe_in) < 0)
return NULL;
if (pipe(pipe_out) < 0)
goto fail_pipe_in;
if (pipe(pipe_err) < 0)
goto fail_pipe_out;
/*
* Finally we invoke tshark to use the pipe as input; the "-i -"
* argument will accomplish this. "-l" forces tshark to flush the pipe
* after each packet, rather than allowing data to build up in it.
*
* We could block and mask SIGCHLD around all this, to avoid trouble if
* the process dies, but this is a long-running process, so it will
* probably die much later. Applications intercepting SIGCHLD will
* simply have to live with SIGCHLDs from tshark. (Anyone intercepting
* SIGCHLD is probably used to SIGCHLDs from all sorts of things.)
*
* Sanitize the environment: we are forced to point HOME somewhere
* readable so that wireshark can look for various config files which it
* won't find, so point it at /run/initramfs, which will always be
* present and almost certainly does not contain any files that may be
* confused with startup files.
*
* Forcibly switch uid to (uid_t)-3 to try to get out from all
* privilege, since tshark is a nest of security vulnerabilities. (We
* use -3 because -1 is nobody, which can have files "owned" by it on
* NFS clients, and avoid -2 because other people have thought of this
* as well (indeed, on some systems nobody has uid -2). (TODO: unshare
* network namespaces too: figure out something to do in user
* namespaces, where setgid() and setuid() to random uids and gids will
* fail. Switch gid to "wireshark" if possible, since otherwise it may
* not be able to run dumpcap: TODO this needs customization for other
* distros.
*
* We synchronize with tshark by waiting for something to emerge on
* stderr (since tshark emits at least one line on stderr even when it
* starts normally), then waitpid()ing for it: if it died, we know we
* should fall back to tracemem().
*/
pid = fork();
switch (pid) {
case 0: {
/*
* Mask SIGINT. DTrace will close the pipes to tshark when it is
* interrupted. (We must do this to avoid a race at termination
* time: if tshark dies first, then any more data flowing
* into the pcap action will be printed as tracemem output,
* which users will not expect.)
*/
sigset_t mask;
sigemptyset(&mask);
sigaddset(&mask, SIGINT);
pthread_sigmask(SIG_BLOCK, &mask, NULL);
/*
* Set up our pipes.
*/
close(pipe_in[1]);
if (pipe_in[0] != STDIN_FILENO) {
if(dup2(pipe_in[0], STDIN_FILENO) < 0)
exit(1);
close(pipe_in[0]);
}
close(pipe_out[0]);
if (pipe_out[1] != STDOUT_FILENO) {
if(dup2(pipe_out[1], STDOUT_FILENO) < 0)
exit(1);
close(pipe_out[1]);
}
/*
* No need for a stderr pipe from tshark, but we need it before
* tshark starts, or if it fails to start, for synchronization.
*/
close(pipe_err[0]);
if (pipe_err[1] != STDERR_FILENO) {
if(dup2(pipe_err[1], STDERR_FILENO) < 0)
exit(1);
close(pipe_err[1]);
}
fcntl(STDERR_FILENO, F_SETFD, FD_CLOEXEC);
/*
* Remove privilege, if needed.
*/
if (getuid() == 0) {
struct group wsg;
struct group *dummy;
char groups[1024];
gid_t wireshark_group = (gid_t)UNPRIV_UID;
if (getgrnam_r(DUMPCAP_GROUP, &wsg, groups, 1024, &dummy) >= 0) {
wireshark_group = wsg.gr_gid;
}
if (chdir("/") < 0)
goto nopriv_die;
if (setgid(wireshark_group) < 0)
goto nopriv_die;
if (setgroups(1, &wireshark_group) < 0)
goto nopriv_die;
if (setuid((uid_t)UNPRIV_UID) < 0)
goto nopriv_die;
}
/*
* Sanitize the environment.
*/
clearenv();
putenv("PATH=/usr/sbin:/usr/bin:/sbin:/bin");
putenv("HOME=" UNPRIV_HOME);
putenv("SHELL=/bin/sh");
putenv("USER=nobody");
execlp("tshark", "tshark", "-l", "-i", "-", NULL);
nopriv_die:
fprintf(stderr, "Cannot drop privileges or exec tshark: %s\n",
strerror(errno));
exit(1);
}
case -1:
/* fork failed; clean up. */
goto fail_pipe_err;
default: {
struct pollfd err;
/*
* In parent. Synchronize with the stderr pipe to ensure the
* process has started or died (and waitpid for it and give up
* if it died), then record pid and the appropriate ends
* of the pipe fds, kick off a pthread to printf() the output
* from the tshark, and return an empty string to signify that a
* pipe is in use.
*/
close(pipe_in[0]);
close(pipe_out[1]);
close(pipe_err[1]);
dt_dprintf("Waiting for tshark startup.\n");
err.fd = pipe_err[0];
err.events = POLLIN;
err.revents = 0;
while (errno = EINTR,
poll(&err, 1, -1) <= 0 && errno == EINTR)
continue;
/*
* Output on the stderr pipe means failure.
*/
if (err.revents & POLLIN) {
char err[1024];
read(pipe_err[0], err, 1023);
dt_dprintf("%s", err);
close(pipe_err[0]);
goto fail_pipe_close_wait;
}
close(pipe_err[0]);
dtp->dt_pcap.dt_pcap_pipe[0] = pipe_out[0];
dtp->dt_pcap.dt_pcap_pipe[1] = pipe_in[1];
dtp->dt_pcap.dt_pcap_pid = pid;
dtp->dt_pcap.dt_pcap_out_fp = fp;
if (pthread_create(&dtp->dt_pcap.dt_pcap_output, NULL,
dt_pcap_print, dtp) < 0)
goto fail_pipe_close_wait;
return "";
fail_pipe_close_wait:
/*
* tshark failed to start or we can't create its
* printing thread. Close the pipes and give up.
* (tshark will terminate immediately.)
*/
close(pipe_in[1]);
close(pipe_out[0]);
while (waitpid(pid, NULL, 0) < 0 && errno == EINTR);
dtp->dt_pcap.dt_pcap_pid = -1;
return NULL;
}
}
fail_pipe_err:
close(pipe_err[0]);
close(pipe_err[1]);
fail_pipe_out:
close(pipe_out[0]);
close(pipe_out[1]);
fail_pipe_in:
close(pipe_in[0]);
close(pipe_in[1]);
return NULL;
}
void
dt_pcap_dump(dtrace_hdl_t *dtp, const char *filename, uint64_t linktype,
uint64_t time, void *data, uint32_t datalen, uint32_t maxlen)
{
struct pcap_pkthdr hdr;
dt_pcap_t *dpc, *npc;
struct sigaction act, oact;
for (dpc = dt_list_next(&dtp->dt_pcap.dt_pcaps); dpc != NULL; dpc = npc) {
npc = dt_list_next(dpc);
if (strcmp(dpc->dpc_filename, filename) == 0)
break;
}
if (dpc == NULL) {
dpc = dt_pcap_create(dtp, filename, maxlen);
if (dpc == NULL)
return;
}
if (dpc->dpc_pcap == NULL) {
dpc->dpc_pcap = pcap_open_dead((int)linktype,
(int)dpc->dpc_maxlen);
if (dpc->dpc_pcap == NULL) {
dt_set_errno(dtp, EINVAL);
return;
}
if (dpc->dpc_filename[0] != '\0') {
dpc->dpc_pcap_dump = pcap_dump_open(dpc->dpc_pcap,
dpc->dpc_filename);
} else {
int fd;
fd = fcntl(dtp->dt_pcap.dt_pcap_pipe[1],
F_DUPFD_CLOEXEC, 3);
if (fd >= 0)
dpc->dpc_outpipe = fdopen(fd, "a");
if (fd < 0 || dpc->dpc_outpipe == NULL) {
if (fd >= 0)
close(fd);
dt_dprintf("Cannot connect pipe: "
"%s\n", strerror(errno));
dt_set_errno(dtp, errno);
pcap_close(dpc->dpc_pcap);
dpc->dpc_pcap = NULL;
return;
}
dpc->dpc_pcap_dump = pcap_dump_fopen(dpc->dpc_pcap,
dpc->dpc_outpipe);
}
dpc->dpc_linktype = linktype;
} else if (linktype != dpc->dpc_linktype) {
/* Handle linktype mismatch here... */
dt_dprintf("pcap() expected linktype %lu, got %lu.\n",
dpc->dpc_linktype, linktype);
dt_set_errno(dtp, EINVAL);
return;
}
hdr.len = datalen;
hdr.caplen = datalen;
if (datalen > maxlen)
hdr.caplen = maxlen;
hdr.ts.tv_sec = dpc->dpc_boottime + (time/NANOSEC);
hdr.ts.tv_usec = (time % NANOSEC) / 1000;
/*
* Reset SIGPIPE here, to avoid SIGPIPEs if tshark dies before we do.
*/
memset(&act, 0, sizeof(act));
act.sa_handler = SIG_IGN;
sigaction(SIGPIPE, &act, &oact);
pthread_mutex_lock(&dpc->dpc_lock);
pcap_dump((uchar_t *)dpc->dpc_pcap_dump, &hdr, data);
/*
* We flush when using a pipe to avoid a backlog building up in the pipe
* buffers.
*/
if (dtp->dt_pcap.dt_pcap_pid > 0)
pcap_dump_flush(dpc->dpc_pcap_dump);
pthread_mutex_unlock(&dpc->dpc_lock);
sigaction(SIGCHLD, &oact, NULL);
}
|