| 1 | /* |
| 2 | * block queue tracing application |
| 3 | * |
| 4 | * Copyright (C) 2005 Jens Axboe <axboe@suse.de> |
| 5 | * Copyright (C) 2006 Jens Axboe <axboe@kernel.dk> |
| 6 | * |
| 7 | * Rewrite to have a single thread per CPU (managing all devices on that CPU) |
| 8 | * Alan D. Brunelle <alan.brunelle@hp.com> - January 2009 |
| 9 | * |
| 10 | * This program is free software; you can redistribute it and/or modify |
| 11 | * it under the terms of the GNU General Public License as published by |
| 12 | * the Free Software Foundation; either version 2 of the License, or |
| 13 | * (at your option) any later version. |
| 14 | * |
| 15 | * This program is distributed in the hope that it will be useful, |
| 16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 18 | * GNU General Public License for more details. |
| 19 | * |
| 20 | * You should have received a copy of the GNU General Public License |
| 21 | * along with this program; if not, write to the Free Software |
| 22 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
| 23 | * |
| 24 | */ |
| 25 | |
| 26 | #include <errno.h> |
| 27 | #include <stdarg.h> |
| 28 | #include <stdio.h> |
| 29 | #include <stdlib.h> |
| 30 | #include <string.h> |
| 31 | #include <fcntl.h> |
| 32 | #include <getopt.h> |
| 33 | #include <sched.h> |
| 34 | #include <unistd.h> |
| 35 | #include <poll.h> |
| 36 | #include <signal.h> |
| 37 | #include <pthread.h> |
| 38 | #include <locale.h> |
| 39 | #include <sys/ioctl.h> |
| 40 | #include <sys/types.h> |
| 41 | #include <sys/stat.h> |
| 42 | #include <sys/vfs.h> |
| 43 | #include <sys/mman.h> |
| 44 | #include <sys/param.h> |
| 45 | #include <sys/time.h> |
| 46 | #include <sys/resource.h> |
| 47 | #include <sys/socket.h> |
| 48 | #include <netinet/in.h> |
| 49 | #include <arpa/inet.h> |
| 50 | #include <netdb.h> |
| 51 | #include <sys/sendfile.h> |
| 52 | |
| 53 | #include "btt/list.h" |
| 54 | #include "blktrace.h" |
| 55 | |
| 56 | /* |
| 57 | * You may want to increase this even more, if you are logging at a high |
| 58 | * rate and see skipped/missed events |
| 59 | */ |
| 60 | #define BUF_SIZE (512 * 1024) |
| 61 | #define BUF_NR (4) |
| 62 | |
| 63 | #define FILE_VBUF_SIZE (128 * 1024) |
| 64 | |
| 65 | #define DEBUGFS_TYPE (0x64626720) |
| 66 | #define TRACE_NET_PORT (8462) |
| 67 | |
| 68 | enum { |
| 69 | Net_none = 0, |
| 70 | Net_server, |
| 71 | Net_client, |
| 72 | }; |
| 73 | |
| 74 | enum thread_status { |
| 75 | Th_running, |
| 76 | Th_leaving, |
| 77 | Th_error |
| 78 | }; |
| 79 | |
| 80 | /* |
| 81 | * Generic stats collected: nevents can be _roughly_ estimated by data_read |
| 82 | * (discounting pdu...) |
| 83 | * |
| 84 | * These fields are updated w/ pdc_dr_update & pdc_nev_update below. |
| 85 | */ |
| 86 | struct pdc_stats { |
| 87 | unsigned long long data_read; |
| 88 | unsigned long long nevents; |
| 89 | }; |
| 90 | |
| 91 | struct devpath { |
| 92 | struct list_head head; |
| 93 | char *path; /* path to device special file */ |
| 94 | char *buts_name; /* name returned from bt kernel code */ |
| 95 | struct pdc_stats *stats; |
| 96 | int fd, ncpus; |
| 97 | unsigned long long drops; |
| 98 | |
| 99 | /* |
| 100 | * For piped output only: |
| 101 | * |
| 102 | * Each tracer will have a tracer_devpath_head that it will add new |
| 103 | * data onto. It's list is protected above (tracer_devpath_head.mutex) |
| 104 | * and it will signal the processing thread using the dp_cond, |
| 105 | * dp_mutex & dp_entries variables above. |
| 106 | */ |
| 107 | struct tracer_devpath_head *heads; |
| 108 | |
| 109 | /* |
| 110 | * For network server mode only: |
| 111 | */ |
| 112 | struct cl_host *ch; |
| 113 | u32 cl_id; |
| 114 | time_t cl_connect_time; |
| 115 | int setup_done; /* ioctl BLKTRACESETUP done */ |
| 116 | struct io_info *ios; |
| 117 | }; |
| 118 | |
| 119 | /* |
| 120 | * For piped output to stdout we will have each tracer thread (one per dev) |
| 121 | * tack buffers read from the relay queues on a per-device list. |
| 122 | * |
| 123 | * The main thread will then collect trace buffers from each of lists in turn. |
| 124 | * |
| 125 | * We will use a mutex to guard each of the trace_buf list. The tracers |
| 126 | * can then signal the main thread using <dp_cond,dp_mutex> and |
| 127 | * dp_entries. (When dp_entries is 0, and a tracer adds an entry it will |
| 128 | * signal. When dp_entries is 0, the main thread will wait for that condition |
| 129 | * to be signalled.) |
| 130 | * |
| 131 | * adb: It may be better just to have a large buffer per tracer per dev, |
| 132 | * and then use it as a ring-buffer. This would certainly cut down a lot |
| 133 | * of malloc/free thrashing, at the cost of more memory movements (potentially). |
| 134 | */ |
| 135 | struct trace_buf { |
| 136 | struct list_head head; |
| 137 | struct devpath *dpp; |
| 138 | void *buf; |
| 139 | int cpu, len; |
| 140 | }; |
| 141 | |
| 142 | struct tracer_devpath_head { |
| 143 | pthread_mutex_t mutex; |
| 144 | struct list_head head; |
| 145 | struct trace_buf *prev; |
| 146 | }; |
| 147 | |
| 148 | /* |
| 149 | * Used to handle the mmap() interfaces for output file (containing traces) |
| 150 | */ |
| 151 | struct mmap_info { |
| 152 | void *fs_buf; |
| 153 | unsigned long long fs_size, fs_max_size, fs_off, fs_buf_len; |
| 154 | unsigned long buf_size, buf_nr; |
| 155 | int pagesize; |
| 156 | }; |
| 157 | |
| 158 | /* |
| 159 | * Each thread doing work on a (client) side of blktrace will have one |
| 160 | * of these. The ios array contains input/output information, pfds holds |
| 161 | * poll() data. The volatile's provide flags to/from the main executing |
| 162 | * thread. |
| 163 | */ |
| 164 | struct tracer { |
| 165 | struct list_head head; |
| 166 | struct io_info *ios; |
| 167 | struct pollfd *pfds; |
| 168 | pthread_t thread; |
| 169 | int cpu, nios; |
| 170 | volatile int status, is_done; |
| 171 | }; |
| 172 | |
| 173 | /* |
| 174 | * networking stuff follows. we include a magic number so we know whether |
| 175 | * to endianness convert or not. |
| 176 | * |
| 177 | * The len field is overloaded: |
| 178 | * 0 - Indicates an "open" - allowing the server to set up for a dev/cpu |
| 179 | * 1 - Indicates a "close" - Shut down connection orderly |
| 180 | * |
| 181 | * The cpu field is overloaded on close: it will contain the number of drops. |
| 182 | */ |
| 183 | struct blktrace_net_hdr { |
| 184 | u32 magic; /* same as trace magic */ |
| 185 | char buts_name[32]; /* trace name */ |
| 186 | u32 cpu; /* for which cpu */ |
| 187 | u32 max_cpus; |
| 188 | u32 len; /* length of following trace data */ |
| 189 | u32 cl_id; /* id for set of client per-cpu connections */ |
| 190 | u32 buf_size; /* client buf_size for this trace */ |
| 191 | u32 buf_nr; /* client buf_nr for this trace */ |
| 192 | u32 page_size; /* client page_size for this trace */ |
| 193 | }; |
| 194 | |
| 195 | /* |
| 196 | * Each host encountered has one of these. The head is used to link this |
| 197 | * on to the network server's ch_list. Connections associated with this |
| 198 | * host are linked on conn_list, and any devices traced on that host |
| 199 | * are connected on the devpaths list. |
| 200 | */ |
| 201 | struct cl_host { |
| 202 | struct list_head head; |
| 203 | struct list_head conn_list; |
| 204 | struct list_head devpaths; |
| 205 | struct net_server_s *ns; |
| 206 | char *hostname; |
| 207 | struct in_addr cl_in_addr; |
| 208 | int connects, ndevs, cl_opens; |
| 209 | }; |
| 210 | |
| 211 | /* |
| 212 | * Each connection (client to server socket ('fd')) has one of these. A |
| 213 | * back reference to the host ('ch'), and lists headers (for the host |
| 214 | * list, and the network server conn_list) are also included. |
| 215 | */ |
| 216 | struct cl_conn { |
| 217 | struct list_head ch_head, ns_head; |
| 218 | struct cl_host *ch; |
| 219 | int fd, ncpus; |
| 220 | time_t connect_time; |
| 221 | }; |
| 222 | |
| 223 | /* |
| 224 | * The network server requires some poll structures to be maintained - |
| 225 | * one per conection currently on conn_list. The nchs/ch_list values |
| 226 | * are for each host connected to this server. The addr field is used |
| 227 | * for scratch as new connections are established. |
| 228 | */ |
| 229 | struct net_server_s { |
| 230 | struct list_head conn_list; |
| 231 | struct list_head ch_list; |
| 232 | struct pollfd *pfds; |
| 233 | int listen_fd, connects, nchs; |
| 234 | struct sockaddr_in addr; |
| 235 | }; |
| 236 | |
| 237 | /* |
| 238 | * This structure is (generically) used to providide information |
| 239 | * for a read-to-write set of values. |
| 240 | * |
| 241 | * ifn & ifd represent input information |
| 242 | * |
| 243 | * ofn, ofd, ofp, obuf & mmap_info are used for output file (optionally). |
| 244 | */ |
| 245 | struct io_info { |
| 246 | struct devpath *dpp; |
| 247 | FILE *ofp; |
| 248 | char *obuf; |
| 249 | struct cl_conn *nc; /* Server network connection */ |
| 250 | |
| 251 | /* |
| 252 | * mmap controlled output files |
| 253 | */ |
| 254 | struct mmap_info mmap_info; |
| 255 | |
| 256 | /* |
| 257 | * Client network fields |
| 258 | */ |
| 259 | unsigned int ready; |
| 260 | unsigned long long data_queued; |
| 261 | |
| 262 | /* |
| 263 | * Input/output file descriptors & names |
| 264 | */ |
| 265 | int ifd, ofd; |
| 266 | char ifn[MAXPATHLEN + 64]; |
| 267 | char ofn[MAXPATHLEN + 64]; |
| 268 | }; |
| 269 | |
| 270 | static char blktrace_version[] = "2.0.0"; |
| 271 | |
| 272 | /* |
| 273 | * Linkage to blktrace helper routines (trace conversions) |
| 274 | */ |
| 275 | int data_is_native = -1; |
| 276 | |
| 277 | static int ndevs; |
| 278 | static int max_cpus; |
| 279 | static int ncpus; |
| 280 | static cpu_set_t *online_cpus; |
| 281 | static int pagesize; |
| 282 | static int act_mask = ~0U; |
| 283 | static int kill_running_trace; |
| 284 | static int stop_watch; |
| 285 | static int piped_output; |
| 286 | |
| 287 | static char *debugfs_path = "/sys/kernel/debug"; |
| 288 | static char *output_name; |
| 289 | static char *output_dir; |
| 290 | |
| 291 | static unsigned long buf_size = BUF_SIZE; |
| 292 | static unsigned long buf_nr = BUF_NR; |
| 293 | |
| 294 | static FILE *pfp; |
| 295 | |
| 296 | static LIST_HEAD(devpaths); |
| 297 | static LIST_HEAD(tracers); |
| 298 | |
| 299 | static volatile int done; |
| 300 | |
| 301 | /* |
| 302 | * tracer threads add entries, the main thread takes them off and processes |
| 303 | * them. These protect the dp_entries variable. |
| 304 | */ |
| 305 | static pthread_cond_t dp_cond = PTHREAD_COND_INITIALIZER; |
| 306 | static pthread_mutex_t dp_mutex = PTHREAD_MUTEX_INITIALIZER; |
| 307 | static volatile int dp_entries; |
| 308 | |
| 309 | /* |
| 310 | * These synchronize main / thread interactions. |
| 311 | */ |
| 312 | static pthread_cond_t mt_cond = PTHREAD_COND_INITIALIZER; |
| 313 | static pthread_mutex_t mt_mutex = PTHREAD_MUTEX_INITIALIZER; |
| 314 | static volatile int nthreads_running; |
| 315 | static volatile int nthreads_leaving; |
| 316 | static volatile int nthreads_error; |
| 317 | static volatile int tracers_run; |
| 318 | |
| 319 | /* |
| 320 | * network cmd line params |
| 321 | */ |
| 322 | static struct sockaddr_in hostname_addr; |
| 323 | static char hostname[MAXHOSTNAMELEN]; |
| 324 | static int net_port = TRACE_NET_PORT; |
| 325 | static int net_use_sendfile = 1; |
| 326 | static int net_mode; |
| 327 | static int *cl_fds; |
| 328 | |
| 329 | static int (*handle_pfds)(struct tracer *, int, int); |
| 330 | static int (*handle_list)(struct tracer_devpath_head *, struct list_head *); |
| 331 | |
| 332 | #define S_OPTS "d:a:A:r:o:kw:vVb:n:D:lh:p:sI:" |
| 333 | static struct option l_opts[] = { |
| 334 | { |
| 335 | .name = "dev", |
| 336 | .has_arg = required_argument, |
| 337 | .flag = NULL, |
| 338 | .val = 'd' |
| 339 | }, |
| 340 | { |
| 341 | .name = "input-devs", |
| 342 | .has_arg = required_argument, |
| 343 | .flag = NULL, |
| 344 | .val = 'I' |
| 345 | }, |
| 346 | { |
| 347 | .name = "act-mask", |
| 348 | .has_arg = required_argument, |
| 349 | .flag = NULL, |
| 350 | .val = 'a' |
| 351 | }, |
| 352 | { |
| 353 | .name = "set-mask", |
| 354 | .has_arg = required_argument, |
| 355 | .flag = NULL, |
| 356 | .val = 'A' |
| 357 | }, |
| 358 | { |
| 359 | .name = "relay", |
| 360 | .has_arg = required_argument, |
| 361 | .flag = NULL, |
| 362 | .val = 'r' |
| 363 | }, |
| 364 | { |
| 365 | .name = "output", |
| 366 | .has_arg = required_argument, |
| 367 | .flag = NULL, |
| 368 | .val = 'o' |
| 369 | }, |
| 370 | { |
| 371 | .name = "kill", |
| 372 | .has_arg = no_argument, |
| 373 | .flag = NULL, |
| 374 | .val = 'k' |
| 375 | }, |
| 376 | { |
| 377 | .name = "stopwatch", |
| 378 | .has_arg = required_argument, |
| 379 | .flag = NULL, |
| 380 | .val = 'w' |
| 381 | }, |
| 382 | { |
| 383 | .name = "version", |
| 384 | .has_arg = no_argument, |
| 385 | .flag = NULL, |
| 386 | .val = 'v' |
| 387 | }, |
| 388 | { |
| 389 | .name = "version", |
| 390 | .has_arg = no_argument, |
| 391 | .flag = NULL, |
| 392 | .val = 'V' |
| 393 | }, |
| 394 | { |
| 395 | .name = "buffer-size", |
| 396 | .has_arg = required_argument, |
| 397 | .flag = NULL, |
| 398 | .val = 'b' |
| 399 | }, |
| 400 | { |
| 401 | .name = "num-sub-buffers", |
| 402 | .has_arg = required_argument, |
| 403 | .flag = NULL, |
| 404 | .val = 'n' |
| 405 | }, |
| 406 | { |
| 407 | .name = "output-dir", |
| 408 | .has_arg = required_argument, |
| 409 | .flag = NULL, |
| 410 | .val = 'D' |
| 411 | }, |
| 412 | { |
| 413 | .name = "listen", |
| 414 | .has_arg = no_argument, |
| 415 | .flag = NULL, |
| 416 | .val = 'l' |
| 417 | }, |
| 418 | { |
| 419 | .name = "host", |
| 420 | .has_arg = required_argument, |
| 421 | .flag = NULL, |
| 422 | .val = 'h' |
| 423 | }, |
| 424 | { |
| 425 | .name = "port", |
| 426 | .has_arg = required_argument, |
| 427 | .flag = NULL, |
| 428 | .val = 'p' |
| 429 | }, |
| 430 | { |
| 431 | .name = "no-sendfile", |
| 432 | .has_arg = no_argument, |
| 433 | .flag = NULL, |
| 434 | .val = 's' |
| 435 | }, |
| 436 | { |
| 437 | .name = NULL, |
| 438 | } |
| 439 | }; |
| 440 | |
| 441 | static char usage_str[] = "\n\n" \ |
| 442 | "-d <dev> | --dev=<dev>\n" \ |
| 443 | "[ -r <debugfs path> | --relay=<debugfs path> ]\n" \ |
| 444 | "[ -o <file> | --output=<file>]\n" \ |
| 445 | "[ -D <dir> | --output-dir=<dir>\n" \ |
| 446 | "[ -w <time> | --stopwatch=<time>]\n" \ |
| 447 | "[ -a <action field> | --act-mask=<action field>]\n" \ |
| 448 | "[ -A <action mask> | --set-mask=<action mask>]\n" \ |
| 449 | "[ -b <size> | --buffer-size]\n" \ |
| 450 | "[ -n <number> | --num-sub-buffers=<number>]\n" \ |
| 451 | "[ -l | --listen]\n" \ |
| 452 | "[ -h <hostname> | --host=<hostname>]\n" \ |
| 453 | "[ -p <port number> | --port=<port number>]\n" \ |
| 454 | "[ -s | --no-sendfile]\n" \ |
| 455 | "[ -I <devs file> | --input-devs=<devs file>]\n" \ |
| 456 | "[ -v <version> | --version]\n" \ |
| 457 | "[ -V <version> | --version]\n" \ |
| 458 | |
| 459 | "\t-d Use specified device. May also be given last after options\n" \ |
| 460 | "\t-r Path to mounted debugfs, defaults to /sys/kernel/debug\n" \ |
| 461 | "\t-o File(s) to send output to\n" \ |
| 462 | "\t-D Directory to prepend to output file names\n" \ |
| 463 | "\t-w Stop after defined time, in seconds\n" \ |
| 464 | "\t-a Only trace specified actions. See documentation\n" \ |
| 465 | "\t-A Give trace mask as a single value. See documentation\n" \ |
| 466 | "\t-b Sub buffer size in KiB (default 512)\n" \ |
| 467 | "\t-n Number of sub buffers (default 4)\n" \ |
| 468 | "\t-l Run in network listen mode (blktrace server)\n" \ |
| 469 | "\t-h Run in network client mode, connecting to the given host\n" \ |
| 470 | "\t-p Network port to use (default 8462)\n" \ |
| 471 | "\t-s Make the network client NOT use sendfile() to transfer data\n" \ |
| 472 | "\t-I Add devices found in <devs file>\n" \ |
| 473 | "\t-v Print program version info\n" \ |
| 474 | "\t-V Print program version info\n\n"; |
| 475 | |
| 476 | static void clear_events(struct pollfd *pfd) |
| 477 | { |
| 478 | pfd->events = 0; |
| 479 | pfd->revents = 0; |
| 480 | } |
| 481 | |
| 482 | static inline int net_client_use_sendfile(void) |
| 483 | { |
| 484 | return net_mode == Net_client && net_use_sendfile; |
| 485 | } |
| 486 | |
| 487 | static inline int net_client_use_send(void) |
| 488 | { |
| 489 | return net_mode == Net_client && !net_use_sendfile; |
| 490 | } |
| 491 | |
| 492 | static inline int use_tracer_devpaths(void) |
| 493 | { |
| 494 | return piped_output || net_client_use_send(); |
| 495 | } |
| 496 | |
| 497 | static inline int in_addr_eq(struct in_addr a, struct in_addr b) |
| 498 | { |
| 499 | return a.s_addr == b.s_addr; |
| 500 | } |
| 501 | |
| 502 | static inline void pdc_dr_update(struct devpath *dpp, int cpu, int data_read) |
| 503 | { |
| 504 | dpp->stats[cpu].data_read += data_read; |
| 505 | } |
| 506 | |
| 507 | static inline void pdc_nev_update(struct devpath *dpp, int cpu, int nevents) |
| 508 | { |
| 509 | dpp->stats[cpu].nevents += nevents; |
| 510 | } |
| 511 | |
| 512 | static void show_usage(char *prog) |
| 513 | { |
| 514 | fprintf(stderr, "Usage: %s %s", prog, usage_str); |
| 515 | } |
| 516 | |
| 517 | /* |
| 518 | * Create a timespec 'msec' milliseconds into the future |
| 519 | */ |
| 520 | static inline void make_timespec(struct timespec *tsp, long delta_msec) |
| 521 | { |
| 522 | struct timeval now; |
| 523 | |
| 524 | gettimeofday(&now, NULL); |
| 525 | tsp->tv_sec = now.tv_sec; |
| 526 | tsp->tv_nsec = 1000L * now.tv_usec; |
| 527 | |
| 528 | tsp->tv_nsec += (delta_msec * 1000000L); |
| 529 | if (tsp->tv_nsec > 1000000000L) { |
| 530 | long secs = tsp->tv_nsec / 1000000000L; |
| 531 | |
| 532 | tsp->tv_sec += secs; |
| 533 | tsp->tv_nsec -= (secs * 1000000000L); |
| 534 | } |
| 535 | } |
| 536 | |
| 537 | /* |
| 538 | * Add a timer to ensure wait ends |
| 539 | */ |
| 540 | static void t_pthread_cond_wait(pthread_cond_t *cond, pthread_mutex_t *mutex) |
| 541 | { |
| 542 | struct timespec ts; |
| 543 | |
| 544 | make_timespec(&ts, 50); |
| 545 | pthread_cond_timedwait(cond, mutex, &ts); |
| 546 | } |
| 547 | |
| 548 | static void unblock_tracers(void) |
| 549 | { |
| 550 | pthread_mutex_lock(&mt_mutex); |
| 551 | tracers_run = 1; |
| 552 | pthread_cond_broadcast(&mt_cond); |
| 553 | pthread_mutex_unlock(&mt_mutex); |
| 554 | } |
| 555 | |
| 556 | static void tracer_wait_unblock(struct tracer *tp) |
| 557 | { |
| 558 | pthread_mutex_lock(&mt_mutex); |
| 559 | while (!tp->is_done && !tracers_run) |
| 560 | pthread_cond_wait(&mt_cond, &mt_mutex); |
| 561 | pthread_mutex_unlock(&mt_mutex); |
| 562 | } |
| 563 | |
| 564 | static void tracer_signal_ready(struct tracer *tp, |
| 565 | enum thread_status th_status, |
| 566 | int status) |
| 567 | { |
| 568 | pthread_mutex_lock(&mt_mutex); |
| 569 | tp->status = status; |
| 570 | |
| 571 | if (th_status == Th_running) |
| 572 | nthreads_running++; |
| 573 | else if (th_status == Th_error) |
| 574 | nthreads_error++; |
| 575 | else |
| 576 | nthreads_leaving++; |
| 577 | |
| 578 | pthread_cond_signal(&mt_cond); |
| 579 | pthread_mutex_unlock(&mt_mutex); |
| 580 | } |
| 581 | |
| 582 | static void wait_tracers_ready(int ncpus_started) |
| 583 | { |
| 584 | pthread_mutex_lock(&mt_mutex); |
| 585 | while ((nthreads_running + nthreads_error) < ncpus_started) |
| 586 | t_pthread_cond_wait(&mt_cond, &mt_mutex); |
| 587 | pthread_mutex_unlock(&mt_mutex); |
| 588 | } |
| 589 | |
| 590 | static void wait_tracers_leaving(void) |
| 591 | { |
| 592 | pthread_mutex_lock(&mt_mutex); |
| 593 | while (nthreads_leaving < nthreads_running) |
| 594 | t_pthread_cond_wait(&mt_cond, &mt_mutex); |
| 595 | pthread_mutex_unlock(&mt_mutex); |
| 596 | } |
| 597 | |
| 598 | static void init_mmap_info(struct mmap_info *mip) |
| 599 | { |
| 600 | mip->buf_size = buf_size; |
| 601 | mip->buf_nr = buf_nr; |
| 602 | mip->pagesize = pagesize; |
| 603 | } |
| 604 | |
| 605 | static void net_close_connection(int *fd) |
| 606 | { |
| 607 | shutdown(*fd, SHUT_RDWR); |
| 608 | close(*fd); |
| 609 | *fd = -1; |
| 610 | } |
| 611 | |
| 612 | static void dpp_free(struct devpath *dpp) |
| 613 | { |
| 614 | if (dpp->stats) |
| 615 | free(dpp->stats); |
| 616 | if (dpp->ios) |
| 617 | free(dpp->ios); |
| 618 | if (dpp->path) |
| 619 | free(dpp->path); |
| 620 | if (dpp->buts_name) |
| 621 | free(dpp->buts_name); |
| 622 | free(dpp); |
| 623 | } |
| 624 | |
| 625 | static int lock_on_cpu(int cpu) |
| 626 | { |
| 627 | cpu_set_t * cpu_mask; |
| 628 | size_t size; |
| 629 | |
| 630 | cpu_mask = CPU_ALLOC(max_cpus); |
| 631 | size = CPU_ALLOC_SIZE(max_cpus); |
| 632 | |
| 633 | CPU_ZERO_S(size, cpu_mask); |
| 634 | CPU_SET_S(cpu, size, cpu_mask); |
| 635 | if (sched_setaffinity(0, size, cpu_mask) < 0) { |
| 636 | CPU_FREE(cpu_mask); |
| 637 | return errno; |
| 638 | } |
| 639 | |
| 640 | CPU_FREE(cpu_mask); |
| 641 | return 0; |
| 642 | } |
| 643 | |
| 644 | static int increase_limit(int resource, rlim_t increase) |
| 645 | { |
| 646 | struct rlimit rlim; |
| 647 | int save_errno = errno; |
| 648 | |
| 649 | if (!getrlimit(resource, &rlim)) { |
| 650 | rlim.rlim_cur += increase; |
| 651 | if (rlim.rlim_cur >= rlim.rlim_max) |
| 652 | rlim.rlim_max = rlim.rlim_cur + increase; |
| 653 | |
| 654 | if (!setrlimit(resource, &rlim)) |
| 655 | return 1; |
| 656 | } |
| 657 | |
| 658 | errno = save_errno; |
| 659 | return 0; |
| 660 | } |
| 661 | |
| 662 | static int handle_open_failure(void) |
| 663 | { |
| 664 | if (errno == ENFILE || errno == EMFILE) |
| 665 | return increase_limit(RLIMIT_NOFILE, 16); |
| 666 | return 0; |
| 667 | } |
| 668 | |
| 669 | static int handle_mem_failure(size_t length) |
| 670 | { |
| 671 | if (errno == ENFILE) |
| 672 | return handle_open_failure(); |
| 673 | else if (errno == ENOMEM) |
| 674 | return increase_limit(RLIMIT_MEMLOCK, 2 * length); |
| 675 | return 0; |
| 676 | } |
| 677 | |
| 678 | static FILE *my_fopen(const char *path, const char *mode) |
| 679 | { |
| 680 | FILE *fp; |
| 681 | |
| 682 | do { |
| 683 | fp = fopen(path, mode); |
| 684 | } while (fp == NULL && handle_open_failure()); |
| 685 | |
| 686 | return fp; |
| 687 | } |
| 688 | |
| 689 | static int my_open(const char *path, int flags) |
| 690 | { |
| 691 | int fd; |
| 692 | |
| 693 | do { |
| 694 | fd = open(path, flags); |
| 695 | } while (fd < 0 && handle_open_failure()); |
| 696 | |
| 697 | return fd; |
| 698 | } |
| 699 | |
| 700 | static int my_socket(int domain, int type, int protocol) |
| 701 | { |
| 702 | int fd; |
| 703 | |
| 704 | do { |
| 705 | fd = socket(domain, type, protocol); |
| 706 | } while (fd < 0 && handle_open_failure()); |
| 707 | |
| 708 | return fd; |
| 709 | } |
| 710 | |
| 711 | static int my_accept(int sockfd, struct sockaddr *addr, socklen_t *addrlen) |
| 712 | { |
| 713 | int fd; |
| 714 | |
| 715 | do { |
| 716 | fd = accept(sockfd, addr, addrlen); |
| 717 | } while (fd < 0 && handle_open_failure()); |
| 718 | |
| 719 | return fd; |
| 720 | } |
| 721 | |
| 722 | static void *my_mmap(void *addr, size_t length, int prot, int flags, int fd, |
| 723 | off_t offset) |
| 724 | { |
| 725 | void *new; |
| 726 | |
| 727 | do { |
| 728 | new = mmap(addr, length, prot, flags, fd, offset); |
| 729 | } while (new == MAP_FAILED && handle_mem_failure(length)); |
| 730 | |
| 731 | return new; |
| 732 | } |
| 733 | |
| 734 | static int my_mlock(struct tracer *tp, |
| 735 | const void *addr, size_t len) |
| 736 | { |
| 737 | int ret, retry = 0; |
| 738 | |
| 739 | do { |
| 740 | ret = mlock(addr, len); |
| 741 | if ((retry >= 10) && tp && tp->is_done) |
| 742 | break; |
| 743 | retry++; |
| 744 | } while (ret < 0 && handle_mem_failure(len)); |
| 745 | |
| 746 | return ret; |
| 747 | } |
| 748 | |
| 749 | static int setup_mmap(int fd, unsigned int maxlen, |
| 750 | struct mmap_info *mip, |
| 751 | struct tracer *tp) |
| 752 | { |
| 753 | if (mip->fs_off + maxlen > mip->fs_buf_len) { |
| 754 | unsigned long nr = max(16, mip->buf_nr); |
| 755 | |
| 756 | if (mip->fs_buf) { |
| 757 | munlock(mip->fs_buf, mip->fs_buf_len); |
| 758 | munmap(mip->fs_buf, mip->fs_buf_len); |
| 759 | mip->fs_buf = NULL; |
| 760 | } |
| 761 | |
| 762 | mip->fs_off = mip->fs_size & (mip->pagesize - 1); |
| 763 | mip->fs_buf_len = (nr * mip->buf_size) - mip->fs_off; |
| 764 | mip->fs_max_size += mip->fs_buf_len; |
| 765 | |
| 766 | if (ftruncate(fd, mip->fs_max_size) < 0) { |
| 767 | perror("setup_mmap: ftruncate"); |
| 768 | return 1; |
| 769 | } |
| 770 | |
| 771 | mip->fs_buf = my_mmap(NULL, mip->fs_buf_len, PROT_WRITE, |
| 772 | MAP_SHARED, fd, |
| 773 | mip->fs_size - mip->fs_off); |
| 774 | if (mip->fs_buf == MAP_FAILED) { |
| 775 | perror("setup_mmap: mmap"); |
| 776 | return 1; |
| 777 | } |
| 778 | if (my_mlock(tp, mip->fs_buf, mip->fs_buf_len) < 0) { |
| 779 | perror("setup_mlock: mlock"); |
| 780 | return 1; |
| 781 | } |
| 782 | } |
| 783 | |
| 784 | return 0; |
| 785 | } |
| 786 | |
| 787 | static int __stop_trace(int fd) |
| 788 | { |
| 789 | /* |
| 790 | * Should be stopped, don't complain if it isn't |
| 791 | */ |
| 792 | ioctl(fd, BLKTRACESTOP); |
| 793 | return ioctl(fd, BLKTRACETEARDOWN); |
| 794 | } |
| 795 | |
| 796 | static int write_data(char *buf, int len) |
| 797 | { |
| 798 | int ret; |
| 799 | |
| 800 | rewrite: |
| 801 | ret = fwrite(buf, len, 1, pfp); |
| 802 | if (ferror(pfp) || ret != 1) { |
| 803 | if (errno == EINTR) { |
| 804 | clearerr(pfp); |
| 805 | goto rewrite; |
| 806 | } |
| 807 | |
| 808 | if (!piped_output || (errno != EPIPE && errno != EBADF)) { |
| 809 | fprintf(stderr, "write(%d) failed: %d/%s\n", |
| 810 | len, errno, strerror(errno)); |
| 811 | } |
| 812 | goto err; |
| 813 | } |
| 814 | |
| 815 | fflush(pfp); |
| 816 | return 0; |
| 817 | |
| 818 | err: |
| 819 | clearerr(pfp); |
| 820 | return 1; |
| 821 | } |
| 822 | |
| 823 | /* |
| 824 | * Returns the number of bytes read (successfully) |
| 825 | */ |
| 826 | static int __net_recv_data(int fd, void *buf, unsigned int len) |
| 827 | { |
| 828 | unsigned int bytes_left = len; |
| 829 | |
| 830 | while (bytes_left && !done) { |
| 831 | int ret = recv(fd, buf, bytes_left, MSG_WAITALL); |
| 832 | |
| 833 | if (ret == 0) |
| 834 | break; |
| 835 | else if (ret < 0) { |
| 836 | if (errno == EAGAIN) { |
| 837 | usleep(50); |
| 838 | continue; |
| 839 | } |
| 840 | perror("server: net_recv_data: recv failed"); |
| 841 | break; |
| 842 | } else { |
| 843 | buf += ret; |
| 844 | bytes_left -= ret; |
| 845 | } |
| 846 | } |
| 847 | |
| 848 | return len - bytes_left; |
| 849 | } |
| 850 | |
| 851 | static int net_recv_data(int fd, void *buf, unsigned int len) |
| 852 | { |
| 853 | return __net_recv_data(fd, buf, len); |
| 854 | } |
| 855 | |
| 856 | /* |
| 857 | * Returns number of bytes written |
| 858 | */ |
| 859 | static int net_send_data(int fd, void *buf, unsigned int buf_len) |
| 860 | { |
| 861 | int ret; |
| 862 | unsigned int bytes_left = buf_len; |
| 863 | |
| 864 | while (bytes_left) { |
| 865 | ret = send(fd, buf, bytes_left, 0); |
| 866 | if (ret < 0) { |
| 867 | perror("send"); |
| 868 | break; |
| 869 | } |
| 870 | |
| 871 | buf += ret; |
| 872 | bytes_left -= ret; |
| 873 | } |
| 874 | |
| 875 | return buf_len - bytes_left; |
| 876 | } |
| 877 | |
| 878 | static int net_send_header(int fd, int cpu, char *buts_name, int len) |
| 879 | { |
| 880 | struct blktrace_net_hdr hdr; |
| 881 | |
| 882 | memset(&hdr, 0, sizeof(hdr)); |
| 883 | |
| 884 | hdr.magic = BLK_IO_TRACE_MAGIC; |
| 885 | memset(hdr.buts_name, 0, sizeof(hdr.buts_name)); |
| 886 | strncpy(hdr.buts_name, buts_name, sizeof(hdr.buts_name)); |
| 887 | hdr.buts_name[sizeof(hdr.buts_name) - 1] = '\0'; |
| 888 | hdr.cpu = cpu; |
| 889 | hdr.max_cpus = max_cpus; |
| 890 | hdr.len = len; |
| 891 | hdr.cl_id = getpid(); |
| 892 | hdr.buf_size = buf_size; |
| 893 | hdr.buf_nr = buf_nr; |
| 894 | hdr.page_size = pagesize; |
| 895 | |
| 896 | return net_send_data(fd, &hdr, sizeof(hdr)) != sizeof(hdr); |
| 897 | } |
| 898 | |
| 899 | static void net_send_open_close(int fd, int cpu, char *buts_name, int len) |
| 900 | { |
| 901 | struct blktrace_net_hdr ret_hdr; |
| 902 | |
| 903 | net_send_header(fd, cpu, buts_name, len); |
| 904 | net_recv_data(fd, &ret_hdr, sizeof(ret_hdr)); |
| 905 | } |
| 906 | |
| 907 | static void net_send_open(int fd, int cpu, char *buts_name) |
| 908 | { |
| 909 | net_send_open_close(fd, cpu, buts_name, 0); |
| 910 | } |
| 911 | |
| 912 | static void net_send_close(int fd, char *buts_name, int drops) |
| 913 | { |
| 914 | /* |
| 915 | * Overload CPU w/ number of drops |
| 916 | * |
| 917 | * XXX: Need to clear/set done around call - done=1 (which |
| 918 | * is true here) stops reads from happening... :-( |
| 919 | */ |
| 920 | done = 0; |
| 921 | net_send_open_close(fd, drops, buts_name, 1); |
| 922 | done = 1; |
| 923 | } |
| 924 | |
| 925 | static void ack_open_close(int fd, char *buts_name) |
| 926 | { |
| 927 | net_send_header(fd, 0, buts_name, 2); |
| 928 | } |
| 929 | |
| 930 | static void net_send_drops(int fd) |
| 931 | { |
| 932 | struct list_head *p; |
| 933 | |
| 934 | __list_for_each(p, &devpaths) { |
| 935 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 936 | |
| 937 | net_send_close(fd, dpp->buts_name, dpp->drops); |
| 938 | } |
| 939 | } |
| 940 | |
| 941 | /* |
| 942 | * Returns: |
| 943 | * 0: "EOF" |
| 944 | * 1: OK |
| 945 | * -1: Error |
| 946 | */ |
| 947 | static int net_get_header(struct cl_conn *nc, struct blktrace_net_hdr *bnh) |
| 948 | { |
| 949 | int bytes_read; |
| 950 | int fl = fcntl(nc->fd, F_GETFL); |
| 951 | |
| 952 | fcntl(nc->fd, F_SETFL, fl | O_NONBLOCK); |
| 953 | bytes_read = __net_recv_data(nc->fd, bnh, sizeof(*bnh)); |
| 954 | fcntl(nc->fd, F_SETFL, fl & ~O_NONBLOCK); |
| 955 | |
| 956 | if (bytes_read == sizeof(*bnh)) |
| 957 | return 1; |
| 958 | else if (bytes_read == 0) |
| 959 | return 0; |
| 960 | else |
| 961 | return -1; |
| 962 | } |
| 963 | |
| 964 | static int net_setup_addr(void) |
| 965 | { |
| 966 | struct sockaddr_in *addr = &hostname_addr; |
| 967 | |
| 968 | memset(addr, 0, sizeof(*addr)); |
| 969 | addr->sin_family = AF_INET; |
| 970 | addr->sin_port = htons(net_port); |
| 971 | |
| 972 | if (inet_aton(hostname, &addr->sin_addr) != 1) { |
| 973 | struct hostent *hent; |
| 974 | retry: |
| 975 | hent = gethostbyname(hostname); |
| 976 | if (!hent) { |
| 977 | if (h_errno == TRY_AGAIN) { |
| 978 | usleep(100); |
| 979 | goto retry; |
| 980 | } else if (h_errno == NO_RECOVERY) { |
| 981 | fprintf(stderr, "gethostbyname(%s)" |
| 982 | "non-recoverable error encountered\n", |
| 983 | hostname); |
| 984 | } else { |
| 985 | /* |
| 986 | * HOST_NOT_FOUND, NO_ADDRESS or NO_DATA |
| 987 | */ |
| 988 | fprintf(stderr, "Host %s not found\n", |
| 989 | hostname); |
| 990 | } |
| 991 | return 1; |
| 992 | } |
| 993 | |
| 994 | memcpy(&addr->sin_addr, hent->h_addr, 4); |
| 995 | memset(hostname, 0, sizeof(hostname)); |
| 996 | strncpy(hostname, hent->h_name, sizeof(hostname)); |
| 997 | hostname[sizeof(hostname) - 1] = '\0'; |
| 998 | } |
| 999 | |
| 1000 | return 0; |
| 1001 | } |
| 1002 | |
| 1003 | static int net_setup_client(void) |
| 1004 | { |
| 1005 | int fd; |
| 1006 | struct sockaddr_in *addr = &hostname_addr; |
| 1007 | |
| 1008 | fd = my_socket(AF_INET, SOCK_STREAM, 0); |
| 1009 | if (fd < 0) { |
| 1010 | perror("client: socket"); |
| 1011 | return -1; |
| 1012 | } |
| 1013 | |
| 1014 | if (connect(fd, (struct sockaddr *)addr, sizeof(*addr)) < 0) { |
| 1015 | if (errno == ECONNREFUSED) |
| 1016 | fprintf(stderr, |
| 1017 | "\nclient: Connection to %s refused, " |
| 1018 | "perhaps the server is not started?\n\n", |
| 1019 | hostname); |
| 1020 | else |
| 1021 | perror("client: connect"); |
| 1022 | |
| 1023 | close(fd); |
| 1024 | return -1; |
| 1025 | } |
| 1026 | |
| 1027 | return fd; |
| 1028 | } |
| 1029 | |
| 1030 | static int open_client_connections(void) |
| 1031 | { |
| 1032 | int cpu; |
| 1033 | size_t alloc_size = CPU_ALLOC_SIZE(max_cpus); |
| 1034 | |
| 1035 | cl_fds = calloc(ncpus, sizeof(*cl_fds)); |
| 1036 | for (cpu = 0; cpu < max_cpus; cpu++) { |
| 1037 | if (!CPU_ISSET_S(cpu, alloc_size, online_cpus)) |
| 1038 | continue; |
| 1039 | cl_fds[cpu] = net_setup_client(); |
| 1040 | if (cl_fds[cpu] < 0) |
| 1041 | goto err; |
| 1042 | } |
| 1043 | return 0; |
| 1044 | |
| 1045 | err: |
| 1046 | while (cpu > 0) |
| 1047 | close(cl_fds[cpu--]); |
| 1048 | free(cl_fds); |
| 1049 | return 1; |
| 1050 | } |
| 1051 | |
| 1052 | static void close_client_connections(void) |
| 1053 | { |
| 1054 | if (cl_fds) { |
| 1055 | int cpu, *fdp; |
| 1056 | size_t alloc_size = CPU_ALLOC_SIZE(max_cpus); |
| 1057 | |
| 1058 | for (cpu = 0, fdp = cl_fds; cpu < max_cpus; cpu++, fdp++) { |
| 1059 | if (!CPU_ISSET_S(cpu, alloc_size, online_cpus)) |
| 1060 | continue; |
| 1061 | if (*fdp >= 0) { |
| 1062 | net_send_drops(*fdp); |
| 1063 | net_close_connection(fdp); |
| 1064 | } |
| 1065 | } |
| 1066 | free(cl_fds); |
| 1067 | } |
| 1068 | } |
| 1069 | |
| 1070 | static int setup_buts(void) |
| 1071 | { |
| 1072 | struct list_head *p; |
| 1073 | int ret = 0; |
| 1074 | |
| 1075 | __list_for_each(p, &devpaths) { |
| 1076 | struct blk_user_trace_setup buts; |
| 1077 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1078 | |
| 1079 | memset(&buts, 0, sizeof(buts)); |
| 1080 | buts.buf_size = buf_size; |
| 1081 | buts.buf_nr = buf_nr; |
| 1082 | buts.act_mask = act_mask; |
| 1083 | |
| 1084 | if (ioctl(dpp->fd, BLKTRACESETUP, &buts) >= 0) { |
| 1085 | dpp->ncpus = max_cpus; |
| 1086 | dpp->buts_name = strdup(buts.name); |
| 1087 | dpp->setup_done = 1; |
| 1088 | if (dpp->stats) |
| 1089 | free(dpp->stats); |
| 1090 | dpp->stats = calloc(dpp->ncpus, sizeof(*dpp->stats)); |
| 1091 | memset(dpp->stats, 0, dpp->ncpus * sizeof(*dpp->stats)); |
| 1092 | } else { |
| 1093 | fprintf(stderr, "BLKTRACESETUP(2) %s failed: %d/%s\n", |
| 1094 | dpp->path, errno, strerror(errno)); |
| 1095 | ret++; |
| 1096 | } |
| 1097 | } |
| 1098 | |
| 1099 | return ret; |
| 1100 | } |
| 1101 | |
| 1102 | static void start_buts(void) |
| 1103 | { |
| 1104 | struct list_head *p; |
| 1105 | |
| 1106 | __list_for_each(p, &devpaths) { |
| 1107 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1108 | |
| 1109 | if (ioctl(dpp->fd, BLKTRACESTART) < 0) { |
| 1110 | fprintf(stderr, "BLKTRACESTART %s failed: %d/%s\n", |
| 1111 | dpp->path, errno, strerror(errno)); |
| 1112 | } |
| 1113 | } |
| 1114 | } |
| 1115 | |
| 1116 | static int get_drops(struct devpath *dpp) |
| 1117 | { |
| 1118 | int fd, drops = 0; |
| 1119 | char fn[MAXPATHLEN + 64], tmp[256]; |
| 1120 | |
| 1121 | snprintf(fn, sizeof(fn), "%s/block/%s/dropped", debugfs_path, |
| 1122 | dpp->buts_name); |
| 1123 | |
| 1124 | fd = my_open(fn, O_RDONLY); |
| 1125 | if (fd < 0) { |
| 1126 | /* |
| 1127 | * This may be ok: the kernel may not support |
| 1128 | * dropped counts. |
| 1129 | */ |
| 1130 | if (errno != ENOENT) |
| 1131 | fprintf(stderr, "Could not open %s: %d/%s\n", |
| 1132 | fn, errno, strerror(errno)); |
| 1133 | return 0; |
| 1134 | } else if (read(fd, tmp, sizeof(tmp)) < 0) { |
| 1135 | fprintf(stderr, "Could not read %s: %d/%s\n", |
| 1136 | fn, errno, strerror(errno)); |
| 1137 | } else |
| 1138 | drops = atoi(tmp); |
| 1139 | close(fd); |
| 1140 | |
| 1141 | return drops; |
| 1142 | } |
| 1143 | |
| 1144 | static void get_all_drops(void) |
| 1145 | { |
| 1146 | struct list_head *p; |
| 1147 | |
| 1148 | __list_for_each(p, &devpaths) { |
| 1149 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1150 | |
| 1151 | dpp->drops = get_drops(dpp); |
| 1152 | } |
| 1153 | } |
| 1154 | |
| 1155 | static inline struct trace_buf *alloc_trace_buf(int cpu, int bufsize) |
| 1156 | { |
| 1157 | struct trace_buf *tbp; |
| 1158 | |
| 1159 | tbp = malloc(sizeof(*tbp) + bufsize); |
| 1160 | INIT_LIST_HEAD(&tbp->head); |
| 1161 | tbp->len = 0; |
| 1162 | tbp->buf = (void *)(tbp + 1); |
| 1163 | tbp->cpu = cpu; |
| 1164 | tbp->dpp = NULL; /* Will be set when tbp is added */ |
| 1165 | |
| 1166 | return tbp; |
| 1167 | } |
| 1168 | |
| 1169 | static void free_tracer_heads(struct devpath *dpp) |
| 1170 | { |
| 1171 | int cpu; |
| 1172 | struct tracer_devpath_head *hd; |
| 1173 | |
| 1174 | for (cpu = 0, hd = dpp->heads; cpu < max_cpus; cpu++, hd++) { |
| 1175 | if (hd->prev) |
| 1176 | free(hd->prev); |
| 1177 | |
| 1178 | pthread_mutex_destroy(&hd->mutex); |
| 1179 | } |
| 1180 | free(dpp->heads); |
| 1181 | } |
| 1182 | |
| 1183 | static int setup_tracer_devpaths(void) |
| 1184 | { |
| 1185 | struct list_head *p; |
| 1186 | |
| 1187 | if (net_client_use_send()) |
| 1188 | if (open_client_connections()) |
| 1189 | return 1; |
| 1190 | |
| 1191 | __list_for_each(p, &devpaths) { |
| 1192 | int cpu; |
| 1193 | struct tracer_devpath_head *hd; |
| 1194 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1195 | |
| 1196 | dpp->heads = calloc(max_cpus, sizeof(struct tracer_devpath_head)); |
| 1197 | for (cpu = 0, hd = dpp->heads; cpu < max_cpus; cpu++, hd++) { |
| 1198 | INIT_LIST_HEAD(&hd->head); |
| 1199 | pthread_mutex_init(&hd->mutex, NULL); |
| 1200 | hd->prev = NULL; |
| 1201 | } |
| 1202 | } |
| 1203 | |
| 1204 | return 0; |
| 1205 | } |
| 1206 | |
| 1207 | static inline void add_trace_buf(struct devpath *dpp, int cpu, |
| 1208 | struct trace_buf **tbpp) |
| 1209 | { |
| 1210 | struct trace_buf *tbp = *tbpp; |
| 1211 | struct tracer_devpath_head *hd = &dpp->heads[cpu]; |
| 1212 | |
| 1213 | tbp->dpp = dpp; |
| 1214 | |
| 1215 | pthread_mutex_lock(&hd->mutex); |
| 1216 | list_add_tail(&tbp->head, &hd->head); |
| 1217 | pthread_mutex_unlock(&hd->mutex); |
| 1218 | |
| 1219 | *tbpp = alloc_trace_buf(cpu, buf_size); |
| 1220 | } |
| 1221 | |
| 1222 | static inline void incr_entries(int entries_handled) |
| 1223 | { |
| 1224 | pthread_mutex_lock(&dp_mutex); |
| 1225 | if (dp_entries == 0) |
| 1226 | pthread_cond_signal(&dp_cond); |
| 1227 | dp_entries += entries_handled; |
| 1228 | pthread_mutex_unlock(&dp_mutex); |
| 1229 | } |
| 1230 | |
| 1231 | static void decr_entries(int handled) |
| 1232 | { |
| 1233 | pthread_mutex_lock(&dp_mutex); |
| 1234 | dp_entries -= handled; |
| 1235 | pthread_mutex_unlock(&dp_mutex); |
| 1236 | } |
| 1237 | |
| 1238 | static int wait_empty_entries(void) |
| 1239 | { |
| 1240 | pthread_mutex_lock(&dp_mutex); |
| 1241 | while (!done && dp_entries == 0) |
| 1242 | t_pthread_cond_wait(&dp_cond, &dp_mutex); |
| 1243 | pthread_mutex_unlock(&dp_mutex); |
| 1244 | |
| 1245 | return !done; |
| 1246 | } |
| 1247 | |
| 1248 | static int add_devpath(char *path) |
| 1249 | { |
| 1250 | int fd; |
| 1251 | struct devpath *dpp; |
| 1252 | struct list_head *p; |
| 1253 | |
| 1254 | /* |
| 1255 | * Verify device is not duplicated |
| 1256 | */ |
| 1257 | __list_for_each(p, &devpaths) { |
| 1258 | struct devpath *tmp = list_entry(p, struct devpath, head); |
| 1259 | if (!strcmp(tmp->path, path)) |
| 1260 | return 0; |
| 1261 | } |
| 1262 | /* |
| 1263 | * Verify device is valid before going too far |
| 1264 | */ |
| 1265 | fd = my_open(path, O_RDONLY | O_NONBLOCK); |
| 1266 | if (fd < 0) { |
| 1267 | fprintf(stderr, "Invalid path %s specified: %d/%s\n", |
| 1268 | path, errno, strerror(errno)); |
| 1269 | return 1; |
| 1270 | } |
| 1271 | |
| 1272 | dpp = malloc(sizeof(*dpp)); |
| 1273 | memset(dpp, 0, sizeof(*dpp)); |
| 1274 | dpp->path = strdup(path); |
| 1275 | dpp->fd = fd; |
| 1276 | ndevs++; |
| 1277 | list_add_tail(&dpp->head, &devpaths); |
| 1278 | |
| 1279 | return 0; |
| 1280 | } |
| 1281 | |
| 1282 | static void rel_devpaths(void) |
| 1283 | { |
| 1284 | struct list_head *p, *q; |
| 1285 | |
| 1286 | list_for_each_safe(p, q, &devpaths) { |
| 1287 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1288 | |
| 1289 | list_del(&dpp->head); |
| 1290 | if (dpp->setup_done) |
| 1291 | __stop_trace(dpp->fd); |
| 1292 | close(dpp->fd); |
| 1293 | |
| 1294 | if (dpp->heads) |
| 1295 | free_tracer_heads(dpp); |
| 1296 | |
| 1297 | dpp_free(dpp); |
| 1298 | ndevs--; |
| 1299 | } |
| 1300 | } |
| 1301 | |
| 1302 | static int flush_subbuf_net(struct trace_buf *tbp) |
| 1303 | { |
| 1304 | int fd = cl_fds[tbp->cpu]; |
| 1305 | struct devpath *dpp = tbp->dpp; |
| 1306 | |
| 1307 | if (net_send_header(fd, tbp->cpu, dpp->buts_name, tbp->len)) |
| 1308 | return 1; |
| 1309 | else if (net_send_data(fd, tbp->buf, tbp->len) != tbp->len) |
| 1310 | return 1; |
| 1311 | |
| 1312 | return 0; |
| 1313 | } |
| 1314 | |
| 1315 | static int |
| 1316 | handle_list_net(__attribute__((__unused__))struct tracer_devpath_head *hd, |
| 1317 | struct list_head *list) |
| 1318 | { |
| 1319 | struct trace_buf *tbp; |
| 1320 | struct list_head *p, *q; |
| 1321 | int entries_handled = 0; |
| 1322 | |
| 1323 | list_for_each_safe(p, q, list) { |
| 1324 | tbp = list_entry(p, struct trace_buf, head); |
| 1325 | |
| 1326 | list_del(&tbp->head); |
| 1327 | entries_handled++; |
| 1328 | |
| 1329 | if (cl_fds[tbp->cpu] >= 0) { |
| 1330 | if (flush_subbuf_net(tbp)) { |
| 1331 | close(cl_fds[tbp->cpu]); |
| 1332 | cl_fds[tbp->cpu] = -1; |
| 1333 | } |
| 1334 | } |
| 1335 | |
| 1336 | free(tbp); |
| 1337 | } |
| 1338 | |
| 1339 | return entries_handled; |
| 1340 | } |
| 1341 | |
| 1342 | /* |
| 1343 | * Tack 'tbp's buf onto the tail of 'prev's buf |
| 1344 | */ |
| 1345 | static struct trace_buf *tb_combine(struct trace_buf *prev, |
| 1346 | struct trace_buf *tbp) |
| 1347 | { |
| 1348 | unsigned long tot_len; |
| 1349 | |
| 1350 | tot_len = prev->len + tbp->len; |
| 1351 | if (tot_len > buf_size) { |
| 1352 | /* |
| 1353 | * tbp->head isn't connected (it was 'prev' |
| 1354 | * so it had been taken off of the list |
| 1355 | * before). Therefore, we can realloc |
| 1356 | * the whole structures, as the other fields |
| 1357 | * are "static". |
| 1358 | */ |
| 1359 | prev = realloc(prev, sizeof(*prev) + tot_len); |
| 1360 | prev->buf = (void *)(prev + 1); |
| 1361 | } |
| 1362 | |
| 1363 | memcpy(prev->buf + prev->len, tbp->buf, tbp->len); |
| 1364 | prev->len = tot_len; |
| 1365 | |
| 1366 | free(tbp); |
| 1367 | return prev; |
| 1368 | } |
| 1369 | |
| 1370 | static int handle_list_file(struct tracer_devpath_head *hd, |
| 1371 | struct list_head *list) |
| 1372 | { |
| 1373 | int off, t_len, nevents; |
| 1374 | struct blk_io_trace *t; |
| 1375 | struct list_head *p, *q; |
| 1376 | int entries_handled = 0; |
| 1377 | struct trace_buf *tbp, *prev; |
| 1378 | |
| 1379 | prev = hd->prev; |
| 1380 | list_for_each_safe(p, q, list) { |
| 1381 | tbp = list_entry(p, struct trace_buf, head); |
| 1382 | list_del(&tbp->head); |
| 1383 | entries_handled++; |
| 1384 | |
| 1385 | /* |
| 1386 | * If there was some leftover before, tack this new |
| 1387 | * entry onto the tail of the previous one. |
| 1388 | */ |
| 1389 | if (prev) |
| 1390 | tbp = tb_combine(prev, tbp); |
| 1391 | |
| 1392 | /* |
| 1393 | * See how many whole traces there are - send them |
| 1394 | * all out in one go. |
| 1395 | */ |
| 1396 | off = 0; |
| 1397 | nevents = 0; |
| 1398 | while (off + (int)sizeof(*t) <= tbp->len) { |
| 1399 | t = (struct blk_io_trace *)(tbp->buf + off); |
| 1400 | t_len = sizeof(*t) + t->pdu_len; |
| 1401 | if (off + t_len > tbp->len) |
| 1402 | break; |
| 1403 | |
| 1404 | off += t_len; |
| 1405 | nevents++; |
| 1406 | } |
| 1407 | if (nevents) |
| 1408 | pdc_nev_update(tbp->dpp, tbp->cpu, nevents); |
| 1409 | |
| 1410 | /* |
| 1411 | * Write any full set of traces, any remaining data is kept |
| 1412 | * for the next pass. |
| 1413 | */ |
| 1414 | if (off) { |
| 1415 | if (write_data(tbp->buf, off) || off == tbp->len) { |
| 1416 | free(tbp); |
| 1417 | prev = NULL; |
| 1418 | } |
| 1419 | else { |
| 1420 | /* |
| 1421 | * Move valid data to beginning of buffer |
| 1422 | */ |
| 1423 | tbp->len -= off; |
| 1424 | memmove(tbp->buf, tbp->buf + off, tbp->len); |
| 1425 | prev = tbp; |
| 1426 | } |
| 1427 | } else |
| 1428 | prev = tbp; |
| 1429 | } |
| 1430 | hd->prev = prev; |
| 1431 | |
| 1432 | return entries_handled; |
| 1433 | } |
| 1434 | |
| 1435 | static void __process_trace_bufs(void) |
| 1436 | { |
| 1437 | int cpu; |
| 1438 | struct list_head *p; |
| 1439 | struct list_head list; |
| 1440 | int handled = 0; |
| 1441 | |
| 1442 | __list_for_each(p, &devpaths) { |
| 1443 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1444 | struct tracer_devpath_head *hd = dpp->heads; |
| 1445 | |
| 1446 | for (cpu = 0; cpu < max_cpus; cpu++, hd++) { |
| 1447 | pthread_mutex_lock(&hd->mutex); |
| 1448 | if (list_empty(&hd->head)) { |
| 1449 | pthread_mutex_unlock(&hd->mutex); |
| 1450 | continue; |
| 1451 | } |
| 1452 | |
| 1453 | list_replace_init(&hd->head, &list); |
| 1454 | pthread_mutex_unlock(&hd->mutex); |
| 1455 | |
| 1456 | handled += handle_list(hd, &list); |
| 1457 | } |
| 1458 | } |
| 1459 | |
| 1460 | if (handled) |
| 1461 | decr_entries(handled); |
| 1462 | } |
| 1463 | |
| 1464 | static void process_trace_bufs(void) |
| 1465 | { |
| 1466 | while (wait_empty_entries()) |
| 1467 | __process_trace_bufs(); |
| 1468 | } |
| 1469 | |
| 1470 | static void clean_trace_bufs(void) |
| 1471 | { |
| 1472 | /* |
| 1473 | * No mutex needed here: we're only reading from the lists, |
| 1474 | * tracers are done |
| 1475 | */ |
| 1476 | while (dp_entries) |
| 1477 | __process_trace_bufs(); |
| 1478 | } |
| 1479 | |
| 1480 | static inline void read_err(int cpu, char *ifn) |
| 1481 | { |
| 1482 | if (errno != EAGAIN) |
| 1483 | fprintf(stderr, "Thread %d failed read of %s: %d/%s\n", |
| 1484 | cpu, ifn, errno, strerror(errno)); |
| 1485 | } |
| 1486 | |
| 1487 | static int net_sendfile(struct io_info *iop) |
| 1488 | { |
| 1489 | int ret; |
| 1490 | |
| 1491 | ret = sendfile(iop->ofd, iop->ifd, NULL, iop->ready); |
| 1492 | if (ret < 0) { |
| 1493 | perror("sendfile"); |
| 1494 | return 1; |
| 1495 | } else if (ret < (int)iop->ready) { |
| 1496 | fprintf(stderr, "short sendfile send (%d of %d)\n", |
| 1497 | ret, iop->ready); |
| 1498 | return 1; |
| 1499 | } |
| 1500 | |
| 1501 | return 0; |
| 1502 | } |
| 1503 | |
| 1504 | static inline int net_sendfile_data(struct tracer *tp, struct io_info *iop) |
| 1505 | { |
| 1506 | struct devpath *dpp = iop->dpp; |
| 1507 | |
| 1508 | if (net_send_header(iop->ofd, tp->cpu, dpp->buts_name, iop->ready)) |
| 1509 | return 1; |
| 1510 | return net_sendfile(iop); |
| 1511 | } |
| 1512 | |
| 1513 | static int fill_ofname(char *dst, int dstlen, char *subdir, char *buts_name, |
| 1514 | int cpu) |
| 1515 | { |
| 1516 | int len; |
| 1517 | struct stat sb; |
| 1518 | |
| 1519 | if (output_dir) |
| 1520 | len = snprintf(dst, dstlen, "%s/", output_dir); |
| 1521 | else |
| 1522 | len = snprintf(dst, dstlen, "./"); |
| 1523 | |
| 1524 | if (subdir) |
| 1525 | len += snprintf(dst + len, dstlen - len, "%s", subdir); |
| 1526 | |
| 1527 | if (stat(dst, &sb) < 0) { |
| 1528 | if (errno != ENOENT) { |
| 1529 | fprintf(stderr, |
| 1530 | "Destination dir %s stat failed: %d/%s\n", |
| 1531 | dst, errno, strerror(errno)); |
| 1532 | return 1; |
| 1533 | } |
| 1534 | /* |
| 1535 | * There is no synchronization between multiple threads |
| 1536 | * trying to create the directory at once. It's harmless |
| 1537 | * to let them try, so just detect the problem and move on. |
| 1538 | */ |
| 1539 | if (mkdir(dst, 0755) < 0 && errno != EEXIST) { |
| 1540 | fprintf(stderr, |
| 1541 | "Destination dir %s can't be made: %d/%s\n", |
| 1542 | dst, errno, strerror(errno)); |
| 1543 | return 1; |
| 1544 | } |
| 1545 | } |
| 1546 | |
| 1547 | if (output_name) |
| 1548 | snprintf(dst + len, dstlen - len, "%s.blktrace.%d", |
| 1549 | output_name, cpu); |
| 1550 | else |
| 1551 | snprintf(dst + len, dstlen - len, "%s.blktrace.%d", |
| 1552 | buts_name, cpu); |
| 1553 | |
| 1554 | return 0; |
| 1555 | } |
| 1556 | |
| 1557 | static int set_vbuf(struct io_info *iop, int mode, size_t size) |
| 1558 | { |
| 1559 | iop->obuf = malloc(size); |
| 1560 | if (setvbuf(iop->ofp, iop->obuf, mode, size) < 0) { |
| 1561 | fprintf(stderr, "setvbuf(%s, %d) failed: %d/%s\n", |
| 1562 | iop->dpp->path, (int)size, errno, |
| 1563 | strerror(errno)); |
| 1564 | free(iop->obuf); |
| 1565 | return 1; |
| 1566 | } |
| 1567 | |
| 1568 | return 0; |
| 1569 | } |
| 1570 | |
| 1571 | static int iop_open(struct io_info *iop, int cpu) |
| 1572 | { |
| 1573 | char hostdir[MAXPATHLEN + 64]; |
| 1574 | |
| 1575 | iop->ofd = -1; |
| 1576 | if (net_mode == Net_server) { |
| 1577 | struct cl_conn *nc = iop->nc; |
| 1578 | int len; |
| 1579 | |
| 1580 | len = snprintf(hostdir, sizeof(hostdir), "%s-", |
| 1581 | nc->ch->hostname); |
| 1582 | len += strftime(hostdir + len, sizeof(hostdir) - len, "%F-%T/", |
| 1583 | gmtime(&iop->dpp->cl_connect_time)); |
| 1584 | } else { |
| 1585 | hostdir[0] = 0; |
| 1586 | } |
| 1587 | |
| 1588 | if (fill_ofname(iop->ofn, sizeof(iop->ofn), hostdir, |
| 1589 | iop->dpp->buts_name, cpu)) |
| 1590 | return 1; |
| 1591 | |
| 1592 | iop->ofp = my_fopen(iop->ofn, "w+"); |
| 1593 | if (iop->ofp == NULL) { |
| 1594 | fprintf(stderr, "Open output file %s failed: %d/%s\n", |
| 1595 | iop->ofn, errno, strerror(errno)); |
| 1596 | return 1; |
| 1597 | } |
| 1598 | |
| 1599 | if (set_vbuf(iop, _IOLBF, FILE_VBUF_SIZE)) { |
| 1600 | fprintf(stderr, "set_vbuf for file %s failed: %d/%s\n", |
| 1601 | iop->ofn, errno, strerror(errno)); |
| 1602 | fclose(iop->ofp); |
| 1603 | return 1; |
| 1604 | } |
| 1605 | |
| 1606 | iop->ofd = fileno(iop->ofp); |
| 1607 | return 0; |
| 1608 | } |
| 1609 | |
| 1610 | static void close_iop(struct io_info *iop) |
| 1611 | { |
| 1612 | struct mmap_info *mip = &iop->mmap_info; |
| 1613 | |
| 1614 | if (mip->fs_buf) |
| 1615 | munmap(mip->fs_buf, mip->fs_buf_len); |
| 1616 | |
| 1617 | if (!piped_output) { |
| 1618 | if (ftruncate(fileno(iop->ofp), mip->fs_size) < 0) { |
| 1619 | fprintf(stderr, |
| 1620 | "Ignoring err: ftruncate(%s): %d/%s\n", |
| 1621 | iop->ofn, errno, strerror(errno)); |
| 1622 | } |
| 1623 | } |
| 1624 | |
| 1625 | if (iop->ofp) |
| 1626 | fclose(iop->ofp); |
| 1627 | if (iop->obuf) |
| 1628 | free(iop->obuf); |
| 1629 | } |
| 1630 | |
| 1631 | static void close_ios(struct tracer *tp) |
| 1632 | { |
| 1633 | while (tp->nios > 0) { |
| 1634 | struct io_info *iop = &tp->ios[--tp->nios]; |
| 1635 | |
| 1636 | iop->dpp->drops = get_drops(iop->dpp); |
| 1637 | if (iop->ifd >= 0) |
| 1638 | close(iop->ifd); |
| 1639 | |
| 1640 | if (iop->ofp) |
| 1641 | close_iop(iop); |
| 1642 | else if (iop->ofd >= 0) { |
| 1643 | struct devpath *dpp = iop->dpp; |
| 1644 | |
| 1645 | net_send_close(iop->ofd, dpp->buts_name, dpp->drops); |
| 1646 | net_close_connection(&iop->ofd); |
| 1647 | } |
| 1648 | } |
| 1649 | |
| 1650 | free(tp->ios); |
| 1651 | free(tp->pfds); |
| 1652 | } |
| 1653 | |
| 1654 | static int open_ios(struct tracer *tp) |
| 1655 | { |
| 1656 | struct pollfd *pfd; |
| 1657 | struct io_info *iop; |
| 1658 | struct list_head *p; |
| 1659 | |
| 1660 | tp->ios = calloc(ndevs, sizeof(struct io_info)); |
| 1661 | memset(tp->ios, 0, ndevs * sizeof(struct io_info)); |
| 1662 | |
| 1663 | tp->pfds = calloc(ndevs, sizeof(struct pollfd)); |
| 1664 | memset(tp->pfds, 0, ndevs * sizeof(struct pollfd)); |
| 1665 | |
| 1666 | tp->nios = 0; |
| 1667 | iop = tp->ios; |
| 1668 | pfd = tp->pfds; |
| 1669 | __list_for_each(p, &devpaths) { |
| 1670 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1671 | |
| 1672 | iop->dpp = dpp; |
| 1673 | iop->ofd = -1; |
| 1674 | snprintf(iop->ifn, sizeof(iop->ifn), "%s/block/%s/trace%d", |
| 1675 | debugfs_path, dpp->buts_name, tp->cpu); |
| 1676 | |
| 1677 | iop->ifd = my_open(iop->ifn, O_RDONLY | O_NONBLOCK); |
| 1678 | if (iop->ifd < 0) { |
| 1679 | fprintf(stderr, "Thread %d failed open %s: %d/%s\n", |
| 1680 | tp->cpu, iop->ifn, errno, strerror(errno)); |
| 1681 | return 1; |
| 1682 | } |
| 1683 | |
| 1684 | init_mmap_info(&iop->mmap_info); |
| 1685 | |
| 1686 | pfd->fd = iop->ifd; |
| 1687 | pfd->events = POLLIN; |
| 1688 | |
| 1689 | if (piped_output) |
| 1690 | ; |
| 1691 | else if (net_client_use_sendfile()) { |
| 1692 | iop->ofd = net_setup_client(); |
| 1693 | if (iop->ofd < 0) |
| 1694 | goto err; |
| 1695 | net_send_open(iop->ofd, tp->cpu, dpp->buts_name); |
| 1696 | } else if (net_mode == Net_none) { |
| 1697 | if (iop_open(iop, tp->cpu)) |
| 1698 | goto err; |
| 1699 | } else { |
| 1700 | /* |
| 1701 | * This ensures that the server knows about all |
| 1702 | * connections & devices before _any_ closes |
| 1703 | */ |
| 1704 | net_send_open(cl_fds[tp->cpu], tp->cpu, dpp->buts_name); |
| 1705 | } |
| 1706 | |
| 1707 | pfd++; |
| 1708 | iop++; |
| 1709 | tp->nios++; |
| 1710 | } |
| 1711 | |
| 1712 | return 0; |
| 1713 | |
| 1714 | err: |
| 1715 | close(iop->ifd); /* tp->nios _not_ bumped */ |
| 1716 | close_ios(tp); |
| 1717 | return 1; |
| 1718 | } |
| 1719 | |
| 1720 | static int handle_pfds_file(struct tracer *tp, int nevs, int force_read) |
| 1721 | { |
| 1722 | struct mmap_info *mip; |
| 1723 | int i, ret, nentries = 0; |
| 1724 | struct pollfd *pfd = tp->pfds; |
| 1725 | struct io_info *iop = tp->ios; |
| 1726 | |
| 1727 | for (i = 0; nevs > 0 && i < ndevs; i++, pfd++, iop++) { |
| 1728 | if (pfd->revents & POLLIN || force_read) { |
| 1729 | mip = &iop->mmap_info; |
| 1730 | |
| 1731 | ret = setup_mmap(iop->ofd, buf_size, mip, tp); |
| 1732 | if (ret < 0) { |
| 1733 | pfd->events = 0; |
| 1734 | break; |
| 1735 | } |
| 1736 | |
| 1737 | ret = read(iop->ifd, mip->fs_buf + mip->fs_off, |
| 1738 | buf_size); |
| 1739 | if (ret > 0) { |
| 1740 | pdc_dr_update(iop->dpp, tp->cpu, ret); |
| 1741 | mip->fs_size += ret; |
| 1742 | mip->fs_off += ret; |
| 1743 | nentries++; |
| 1744 | } else if (ret == 0) { |
| 1745 | /* |
| 1746 | * Short reads after we're done stop us |
| 1747 | * from trying reads. |
| 1748 | */ |
| 1749 | if (tp->is_done) |
| 1750 | clear_events(pfd); |
| 1751 | } else { |
| 1752 | read_err(tp->cpu, iop->ifn); |
| 1753 | if (errno != EAGAIN || tp->is_done) |
| 1754 | clear_events(pfd); |
| 1755 | } |
| 1756 | nevs--; |
| 1757 | } |
| 1758 | } |
| 1759 | |
| 1760 | return nentries; |
| 1761 | } |
| 1762 | |
| 1763 | static int handle_pfds_netclient(struct tracer *tp, int nevs, int force_read) |
| 1764 | { |
| 1765 | struct stat sb; |
| 1766 | int i, nentries = 0; |
| 1767 | struct pollfd *pfd = tp->pfds; |
| 1768 | struct io_info *iop = tp->ios; |
| 1769 | |
| 1770 | for (i = 0; i < ndevs; i++, pfd++, iop++) { |
| 1771 | if (pfd->revents & POLLIN || force_read) { |
| 1772 | if (fstat(iop->ifd, &sb) < 0) { |
| 1773 | perror(iop->ifn); |
| 1774 | pfd->events = 0; |
| 1775 | } else if (sb.st_size > (off_t)iop->data_queued) { |
| 1776 | iop->ready = sb.st_size - iop->data_queued; |
| 1777 | iop->data_queued = sb.st_size; |
| 1778 | |
| 1779 | if (!net_sendfile_data(tp, iop)) { |
| 1780 | pdc_dr_update(iop->dpp, tp->cpu, |
| 1781 | iop->ready); |
| 1782 | nentries++; |
| 1783 | } else |
| 1784 | clear_events(pfd); |
| 1785 | } |
| 1786 | if (--nevs == 0) |
| 1787 | break; |
| 1788 | } |
| 1789 | } |
| 1790 | |
| 1791 | if (nentries) |
| 1792 | incr_entries(nentries); |
| 1793 | |
| 1794 | return nentries; |
| 1795 | } |
| 1796 | |
| 1797 | static int handle_pfds_entries(struct tracer *tp, int nevs, int force_read) |
| 1798 | { |
| 1799 | int i, nentries = 0; |
| 1800 | struct trace_buf *tbp; |
| 1801 | struct pollfd *pfd = tp->pfds; |
| 1802 | struct io_info *iop = tp->ios; |
| 1803 | |
| 1804 | tbp = alloc_trace_buf(tp->cpu, buf_size); |
| 1805 | for (i = 0; i < ndevs; i++, pfd++, iop++) { |
| 1806 | if (pfd->revents & POLLIN || force_read) { |
| 1807 | tbp->len = read(iop->ifd, tbp->buf, buf_size); |
| 1808 | if (tbp->len > 0) { |
| 1809 | pdc_dr_update(iop->dpp, tp->cpu, tbp->len); |
| 1810 | add_trace_buf(iop->dpp, tp->cpu, &tbp); |
| 1811 | nentries++; |
| 1812 | } else if (tbp->len == 0) { |
| 1813 | /* |
| 1814 | * Short reads after we're done stop us |
| 1815 | * from trying reads. |
| 1816 | */ |
| 1817 | if (tp->is_done) |
| 1818 | clear_events(pfd); |
| 1819 | } else { |
| 1820 | read_err(tp->cpu, iop->ifn); |
| 1821 | if (errno != EAGAIN || tp->is_done) |
| 1822 | clear_events(pfd); |
| 1823 | } |
| 1824 | if (!piped_output && --nevs == 0) |
| 1825 | break; |
| 1826 | } |
| 1827 | } |
| 1828 | free(tbp); |
| 1829 | |
| 1830 | if (nentries) |
| 1831 | incr_entries(nentries); |
| 1832 | |
| 1833 | return nentries; |
| 1834 | } |
| 1835 | |
| 1836 | static void *thread_main(void *arg) |
| 1837 | { |
| 1838 | int ret, ndone, to_val; |
| 1839 | struct tracer *tp = arg; |
| 1840 | |
| 1841 | ret = lock_on_cpu(tp->cpu); |
| 1842 | if (ret) |
| 1843 | goto err; |
| 1844 | |
| 1845 | ret = open_ios(tp); |
| 1846 | if (ret) |
| 1847 | goto err; |
| 1848 | |
| 1849 | if (piped_output) |
| 1850 | to_val = 50; /* Frequent partial handles */ |
| 1851 | else |
| 1852 | to_val = 500; /* 1/2 second intervals */ |
| 1853 | |
| 1854 | |
| 1855 | tracer_signal_ready(tp, Th_running, 0); |
| 1856 | tracer_wait_unblock(tp); |
| 1857 | |
| 1858 | while (!tp->is_done) { |
| 1859 | ndone = poll(tp->pfds, ndevs, to_val); |
| 1860 | if (ndone || piped_output) |
| 1861 | (void)handle_pfds(tp, ndone, piped_output); |
| 1862 | else if (ndone < 0 && errno != EINTR) |
| 1863 | fprintf(stderr, "Thread %d poll failed: %d/%s\n", |
| 1864 | tp->cpu, errno, strerror(errno)); |
| 1865 | } |
| 1866 | |
| 1867 | /* |
| 1868 | * Trace is stopped, pull data until we get a short read |
| 1869 | */ |
| 1870 | while (handle_pfds(tp, ndevs, 1) > 0) |
| 1871 | ; |
| 1872 | |
| 1873 | close_ios(tp); |
| 1874 | tracer_signal_ready(tp, Th_leaving, 0); |
| 1875 | return NULL; |
| 1876 | |
| 1877 | err: |
| 1878 | tracer_signal_ready(tp, Th_error, ret); |
| 1879 | return NULL; |
| 1880 | } |
| 1881 | |
| 1882 | static int start_tracer(int cpu) |
| 1883 | { |
| 1884 | struct tracer *tp; |
| 1885 | |
| 1886 | tp = malloc(sizeof(*tp)); |
| 1887 | memset(tp, 0, sizeof(*tp)); |
| 1888 | |
| 1889 | INIT_LIST_HEAD(&tp->head); |
| 1890 | tp->status = 0; |
| 1891 | tp->cpu = cpu; |
| 1892 | |
| 1893 | if (pthread_create(&tp->thread, NULL, thread_main, tp)) { |
| 1894 | fprintf(stderr, "FAILED to start thread on CPU %d: %d/%s\n", |
| 1895 | cpu, errno, strerror(errno)); |
| 1896 | free(tp); |
| 1897 | return 1; |
| 1898 | } |
| 1899 | |
| 1900 | list_add_tail(&tp->head, &tracers); |
| 1901 | return 0; |
| 1902 | } |
| 1903 | |
| 1904 | static int create_output_files(int cpu) |
| 1905 | { |
| 1906 | char fname[MAXPATHLEN + 64]; |
| 1907 | struct list_head *p; |
| 1908 | FILE *f; |
| 1909 | |
| 1910 | __list_for_each(p, &devpaths) { |
| 1911 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1912 | |
| 1913 | if (fill_ofname(fname, sizeof(fname), NULL, dpp->buts_name, |
| 1914 | cpu)) |
| 1915 | return 1; |
| 1916 | f = my_fopen(fname, "w+"); |
| 1917 | if (!f) |
| 1918 | return 1; |
| 1919 | fclose(f); |
| 1920 | } |
| 1921 | return 0; |
| 1922 | } |
| 1923 | |
| 1924 | static void start_tracers(void) |
| 1925 | { |
| 1926 | int cpu, started = 0; |
| 1927 | struct list_head *p; |
| 1928 | size_t alloc_size = CPU_ALLOC_SIZE(max_cpus); |
| 1929 | |
| 1930 | for (cpu = 0; cpu < max_cpus; cpu++) { |
| 1931 | if (!CPU_ISSET_S(cpu, alloc_size, online_cpus)) { |
| 1932 | /* |
| 1933 | * Create fake empty output files so that other tools |
| 1934 | * like blkparse don't have to bother with sparse CPU |
| 1935 | * number space. |
| 1936 | */ |
| 1937 | if (create_output_files(cpu)) |
| 1938 | break; |
| 1939 | continue; |
| 1940 | } |
| 1941 | if (start_tracer(cpu)) |
| 1942 | break; |
| 1943 | started++; |
| 1944 | } |
| 1945 | |
| 1946 | wait_tracers_ready(started); |
| 1947 | |
| 1948 | __list_for_each(p, &tracers) { |
| 1949 | struct tracer *tp = list_entry(p, struct tracer, head); |
| 1950 | if (tp->status) |
| 1951 | fprintf(stderr, |
| 1952 | "FAILED to start thread on CPU %d: %d/%s\n", |
| 1953 | tp->cpu, tp->status, strerror(tp->status)); |
| 1954 | } |
| 1955 | } |
| 1956 | |
| 1957 | static void stop_tracers(void) |
| 1958 | { |
| 1959 | struct list_head *p; |
| 1960 | |
| 1961 | /* |
| 1962 | * Stop the tracing - makes the tracer threads clean up quicker. |
| 1963 | */ |
| 1964 | __list_for_each(p, &devpaths) { |
| 1965 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 1966 | (void)ioctl(dpp->fd, BLKTRACESTOP); |
| 1967 | } |
| 1968 | |
| 1969 | /* |
| 1970 | * Tell each tracer to quit |
| 1971 | */ |
| 1972 | __list_for_each(p, &tracers) { |
| 1973 | struct tracer *tp = list_entry(p, struct tracer, head); |
| 1974 | tp->is_done = 1; |
| 1975 | } |
| 1976 | pthread_cond_broadcast(&mt_cond); |
| 1977 | } |
| 1978 | |
| 1979 | static void del_tracers(void) |
| 1980 | { |
| 1981 | struct list_head *p, *q; |
| 1982 | |
| 1983 | list_for_each_safe(p, q, &tracers) { |
| 1984 | struct tracer *tp = list_entry(p, struct tracer, head); |
| 1985 | |
| 1986 | list_del(&tp->head); |
| 1987 | free(tp); |
| 1988 | } |
| 1989 | } |
| 1990 | |
| 1991 | static void wait_tracers(void) |
| 1992 | { |
| 1993 | struct list_head *p; |
| 1994 | |
| 1995 | if (use_tracer_devpaths()) |
| 1996 | process_trace_bufs(); |
| 1997 | |
| 1998 | wait_tracers_leaving(); |
| 1999 | |
| 2000 | __list_for_each(p, &tracers) { |
| 2001 | int ret; |
| 2002 | struct tracer *tp = list_entry(p, struct tracer, head); |
| 2003 | |
| 2004 | ret = pthread_join(tp->thread, NULL); |
| 2005 | if (ret) |
| 2006 | fprintf(stderr, "Thread join %d failed %d\n", |
| 2007 | tp->cpu, ret); |
| 2008 | } |
| 2009 | |
| 2010 | if (use_tracer_devpaths()) |
| 2011 | clean_trace_bufs(); |
| 2012 | |
| 2013 | get_all_drops(); |
| 2014 | } |
| 2015 | |
| 2016 | static void exit_tracing(void) |
| 2017 | { |
| 2018 | signal(SIGINT, SIG_IGN); |
| 2019 | signal(SIGHUP, SIG_IGN); |
| 2020 | signal(SIGTERM, SIG_IGN); |
| 2021 | signal(SIGALRM, SIG_IGN); |
| 2022 | |
| 2023 | stop_tracers(); |
| 2024 | wait_tracers(); |
| 2025 | del_tracers(); |
| 2026 | rel_devpaths(); |
| 2027 | } |
| 2028 | |
| 2029 | static void handle_sigint(__attribute__((__unused__)) int sig) |
| 2030 | { |
| 2031 | done = 1; |
| 2032 | stop_tracers(); |
| 2033 | } |
| 2034 | |
| 2035 | static void show_stats(struct list_head *devpaths) |
| 2036 | { |
| 2037 | FILE *ofp; |
| 2038 | struct list_head *p; |
| 2039 | unsigned long long nevents, data_read; |
| 2040 | unsigned long long total_drops = 0; |
| 2041 | unsigned long long total_events = 0; |
| 2042 | |
| 2043 | if (piped_output) |
| 2044 | ofp = my_fopen("/dev/null", "w"); |
| 2045 | else |
| 2046 | ofp = stdout; |
| 2047 | |
| 2048 | __list_for_each(p, devpaths) { |
| 2049 | int cpu; |
| 2050 | struct pdc_stats *sp; |
| 2051 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 2052 | |
| 2053 | if (net_mode == Net_server) |
| 2054 | printf("server: end of run for %s:%s\n", |
| 2055 | dpp->ch->hostname, dpp->buts_name); |
| 2056 | |
| 2057 | data_read = 0; |
| 2058 | nevents = 0; |
| 2059 | |
| 2060 | fprintf(ofp, "=== %s ===\n", dpp->buts_name); |
| 2061 | for (cpu = 0, sp = dpp->stats; cpu < dpp->ncpus; cpu++, sp++) { |
| 2062 | /* |
| 2063 | * Estimate events if not known... |
| 2064 | */ |
| 2065 | if (sp->nevents == 0) { |
| 2066 | sp->nevents = sp->data_read / |
| 2067 | sizeof(struct blk_io_trace); |
| 2068 | } |
| 2069 | |
| 2070 | fprintf(ofp, |
| 2071 | " CPU%3d: %20llu events, %8llu KiB data\n", |
| 2072 | cpu, sp->nevents, (sp->data_read + 1023) >> 10); |
| 2073 | |
| 2074 | data_read += sp->data_read; |
| 2075 | nevents += sp->nevents; |
| 2076 | } |
| 2077 | |
| 2078 | fprintf(ofp, " Total: %20llu events (dropped %llu)," |
| 2079 | " %8llu KiB data\n", nevents, |
| 2080 | dpp->drops, (data_read + 1024) >> 10); |
| 2081 | |
| 2082 | total_drops += dpp->drops; |
| 2083 | total_events += (nevents + dpp->drops); |
| 2084 | } |
| 2085 | |
| 2086 | fflush(ofp); |
| 2087 | if (piped_output) |
| 2088 | fclose(ofp); |
| 2089 | |
| 2090 | if (total_drops) { |
| 2091 | double drops_ratio = 1.0; |
| 2092 | |
| 2093 | if (total_events) |
| 2094 | drops_ratio = (double)total_drops/(double)total_events; |
| 2095 | |
| 2096 | fprintf(stderr, "\nYou have %llu (%5.1lf%%) dropped events\n" |
| 2097 | "Consider using a larger buffer size (-b) " |
| 2098 | "and/or more buffers (-n)\n", |
| 2099 | total_drops, 100.0 * drops_ratio); |
| 2100 | } |
| 2101 | } |
| 2102 | |
| 2103 | static int handle_args(int argc, char *argv[]) |
| 2104 | { |
| 2105 | int c, i; |
| 2106 | struct statfs st; |
| 2107 | int act_mask_tmp = 0; |
| 2108 | |
| 2109 | while ((c = getopt_long(argc, argv, S_OPTS, l_opts, NULL)) >= 0) { |
| 2110 | switch (c) { |
| 2111 | case 'a': |
| 2112 | i = find_mask_map(optarg); |
| 2113 | if (i < 0) { |
| 2114 | fprintf(stderr, "Invalid action mask %s\n", |
| 2115 | optarg); |
| 2116 | return 1; |
| 2117 | } |
| 2118 | act_mask_tmp |= i; |
| 2119 | break; |
| 2120 | |
| 2121 | case 'A': |
| 2122 | if ((sscanf(optarg, "%x", &i) != 1) || |
| 2123 | !valid_act_opt(i)) { |
| 2124 | fprintf(stderr, |
| 2125 | "Invalid set action mask %s/0x%x\n", |
| 2126 | optarg, i); |
| 2127 | return 1; |
| 2128 | } |
| 2129 | act_mask_tmp = i; |
| 2130 | break; |
| 2131 | |
| 2132 | case 'd': |
| 2133 | if (add_devpath(optarg) != 0) |
| 2134 | return 1; |
| 2135 | break; |
| 2136 | |
| 2137 | case 'I': { |
| 2138 | char dev_line[256]; |
| 2139 | FILE *ifp = my_fopen(optarg, "r"); |
| 2140 | |
| 2141 | if (!ifp) { |
| 2142 | fprintf(stderr, |
| 2143 | "Invalid file for devices %s\n", |
| 2144 | optarg); |
| 2145 | return 1; |
| 2146 | } |
| 2147 | |
| 2148 | while (fscanf(ifp, "%s\n", dev_line) == 1) { |
| 2149 | if (add_devpath(dev_line) != 0) { |
| 2150 | fclose(ifp); |
| 2151 | return 1; |
| 2152 | } |
| 2153 | } |
| 2154 | fclose(ifp); |
| 2155 | break; |
| 2156 | } |
| 2157 | |
| 2158 | case 'r': |
| 2159 | debugfs_path = optarg; |
| 2160 | break; |
| 2161 | |
| 2162 | case 'o': |
| 2163 | output_name = optarg; |
| 2164 | break; |
| 2165 | case 'k': |
| 2166 | kill_running_trace = 1; |
| 2167 | break; |
| 2168 | case 'w': |
| 2169 | stop_watch = atoi(optarg); |
| 2170 | if (stop_watch <= 0) { |
| 2171 | fprintf(stderr, |
| 2172 | "Invalid stopwatch value (%d secs)\n", |
| 2173 | stop_watch); |
| 2174 | return 1; |
| 2175 | } |
| 2176 | break; |
| 2177 | case 'V': |
| 2178 | case 'v': |
| 2179 | printf("%s version %s\n", argv[0], blktrace_version); |
| 2180 | exit(0); |
| 2181 | /*NOTREACHED*/ |
| 2182 | case 'b': |
| 2183 | buf_size = strtoul(optarg, NULL, 10); |
| 2184 | if (buf_size <= 0 || buf_size > 16*1024) { |
| 2185 | fprintf(stderr, "Invalid buffer size (%lu)\n", |
| 2186 | buf_size); |
| 2187 | return 1; |
| 2188 | } |
| 2189 | buf_size <<= 10; |
| 2190 | break; |
| 2191 | case 'n': |
| 2192 | buf_nr = strtoul(optarg, NULL, 10); |
| 2193 | if (buf_nr <= 0) { |
| 2194 | fprintf(stderr, |
| 2195 | "Invalid buffer nr (%lu)\n", buf_nr); |
| 2196 | return 1; |
| 2197 | } |
| 2198 | break; |
| 2199 | case 'D': |
| 2200 | output_dir = optarg; |
| 2201 | break; |
| 2202 | case 'h': |
| 2203 | net_mode = Net_client; |
| 2204 | memset(hostname, 0, sizeof(hostname)); |
| 2205 | strncpy(hostname, optarg, sizeof(hostname)); |
| 2206 | hostname[sizeof(hostname) - 1] = '\0'; |
| 2207 | break; |
| 2208 | case 'l': |
| 2209 | net_mode = Net_server; |
| 2210 | break; |
| 2211 | case 'p': |
| 2212 | net_port = atoi(optarg); |
| 2213 | break; |
| 2214 | case 's': |
| 2215 | net_use_sendfile = 0; |
| 2216 | break; |
| 2217 | default: |
| 2218 | show_usage(argv[0]); |
| 2219 | exit(1); |
| 2220 | /*NOTREACHED*/ |
| 2221 | } |
| 2222 | } |
| 2223 | |
| 2224 | while (optind < argc) |
| 2225 | if (add_devpath(argv[optind++]) != 0) |
| 2226 | return 1; |
| 2227 | |
| 2228 | if (net_mode != Net_server && ndevs == 0) { |
| 2229 | show_usage(argv[0]); |
| 2230 | return 1; |
| 2231 | } |
| 2232 | |
| 2233 | if (statfs(debugfs_path, &st) < 0) { |
| 2234 | fprintf(stderr, "Invalid debug path %s: %d/%s\n", |
| 2235 | debugfs_path, errno, strerror(errno)); |
| 2236 | return 1; |
| 2237 | } |
| 2238 | |
| 2239 | if (st.f_type != (long)DEBUGFS_TYPE) { |
| 2240 | fprintf(stderr, "Debugfs is not mounted at %s\n", debugfs_path); |
| 2241 | return 1; |
| 2242 | } |
| 2243 | |
| 2244 | if (act_mask_tmp != 0) |
| 2245 | act_mask = act_mask_tmp; |
| 2246 | |
| 2247 | if (net_mode == Net_client && net_setup_addr()) |
| 2248 | return 1; |
| 2249 | |
| 2250 | /* |
| 2251 | * Set up for appropriate PFD handler based upon output name. |
| 2252 | */ |
| 2253 | if (net_client_use_sendfile()) |
| 2254 | handle_pfds = handle_pfds_netclient; |
| 2255 | else if (net_client_use_send()) |
| 2256 | handle_pfds = handle_pfds_entries; |
| 2257 | else if (output_name && (strcmp(output_name, "-") == 0)) { |
| 2258 | piped_output = 1; |
| 2259 | handle_pfds = handle_pfds_entries; |
| 2260 | pfp = stdout; |
| 2261 | if (setvbuf(pfp, NULL, _IONBF, 0)) { |
| 2262 | perror("setvbuf stdout"); |
| 2263 | return 1; |
| 2264 | } |
| 2265 | } else |
| 2266 | handle_pfds = handle_pfds_file; |
| 2267 | return 0; |
| 2268 | } |
| 2269 | |
| 2270 | static void ch_add_connection(struct net_server_s *ns, struct cl_host *ch, |
| 2271 | int fd) |
| 2272 | { |
| 2273 | struct cl_conn *nc; |
| 2274 | |
| 2275 | nc = malloc(sizeof(*nc)); |
| 2276 | memset(nc, 0, sizeof(*nc)); |
| 2277 | |
| 2278 | time(&nc->connect_time); |
| 2279 | nc->ch = ch; |
| 2280 | nc->fd = fd; |
| 2281 | nc->ncpus = -1; |
| 2282 | |
| 2283 | list_add_tail(&nc->ch_head, &ch->conn_list); |
| 2284 | ch->connects++; |
| 2285 | |
| 2286 | list_add_tail(&nc->ns_head, &ns->conn_list); |
| 2287 | ns->connects++; |
| 2288 | ns->pfds = realloc(ns->pfds, (ns->connects+1) * sizeof(struct pollfd)); |
| 2289 | } |
| 2290 | |
| 2291 | static void ch_rem_connection(struct net_server_s *ns, struct cl_host *ch, |
| 2292 | struct cl_conn *nc) |
| 2293 | { |
| 2294 | net_close_connection(&nc->fd); |
| 2295 | |
| 2296 | list_del(&nc->ch_head); |
| 2297 | ch->connects--; |
| 2298 | |
| 2299 | list_del(&nc->ns_head); |
| 2300 | ns->connects--; |
| 2301 | ns->pfds = realloc(ns->pfds, (ns->connects+1) * sizeof(struct pollfd)); |
| 2302 | |
| 2303 | free(nc); |
| 2304 | } |
| 2305 | |
| 2306 | static struct cl_host *net_find_client_host(struct net_server_s *ns, |
| 2307 | struct in_addr cl_in_addr) |
| 2308 | { |
| 2309 | struct list_head *p; |
| 2310 | |
| 2311 | __list_for_each(p, &ns->ch_list) { |
| 2312 | struct cl_host *ch = list_entry(p, struct cl_host, head); |
| 2313 | |
| 2314 | if (in_addr_eq(ch->cl_in_addr, cl_in_addr)) |
| 2315 | return ch; |
| 2316 | } |
| 2317 | |
| 2318 | return NULL; |
| 2319 | } |
| 2320 | |
| 2321 | static struct cl_host *net_add_client_host(struct net_server_s *ns, |
| 2322 | struct sockaddr_in *addr) |
| 2323 | { |
| 2324 | struct cl_host *ch; |
| 2325 | |
| 2326 | ch = malloc(sizeof(*ch)); |
| 2327 | memset(ch, 0, sizeof(*ch)); |
| 2328 | |
| 2329 | ch->ns = ns; |
| 2330 | ch->cl_in_addr = addr->sin_addr; |
| 2331 | list_add_tail(&ch->head, &ns->ch_list); |
| 2332 | ns->nchs++; |
| 2333 | |
| 2334 | ch->hostname = strdup(inet_ntoa(addr->sin_addr)); |
| 2335 | printf("server: connection from %s\n", ch->hostname); |
| 2336 | |
| 2337 | INIT_LIST_HEAD(&ch->conn_list); |
| 2338 | INIT_LIST_HEAD(&ch->devpaths); |
| 2339 | |
| 2340 | return ch; |
| 2341 | } |
| 2342 | |
| 2343 | static void device_done(struct devpath *dpp, int ncpus) |
| 2344 | { |
| 2345 | int cpu; |
| 2346 | struct io_info *iop; |
| 2347 | |
| 2348 | for (cpu = 0, iop = dpp->ios; cpu < ncpus; cpu++, iop++) |
| 2349 | close_iop(iop); |
| 2350 | |
| 2351 | list_del(&dpp->head); |
| 2352 | dpp_free(dpp); |
| 2353 | } |
| 2354 | |
| 2355 | static void net_ch_remove(struct cl_host *ch, int ncpus) |
| 2356 | { |
| 2357 | struct list_head *p, *q; |
| 2358 | struct net_server_s *ns = ch->ns; |
| 2359 | |
| 2360 | list_for_each_safe(p, q, &ch->devpaths) { |
| 2361 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 2362 | device_done(dpp, ncpus); |
| 2363 | } |
| 2364 | |
| 2365 | list_for_each_safe(p, q, &ch->conn_list) { |
| 2366 | struct cl_conn *nc = list_entry(p, struct cl_conn, ch_head); |
| 2367 | |
| 2368 | ch_rem_connection(ns, ch, nc); |
| 2369 | } |
| 2370 | |
| 2371 | list_del(&ch->head); |
| 2372 | ns->nchs--; |
| 2373 | |
| 2374 | if (ch->hostname) |
| 2375 | free(ch->hostname); |
| 2376 | free(ch); |
| 2377 | } |
| 2378 | |
| 2379 | static void net_add_connection(struct net_server_s *ns) |
| 2380 | { |
| 2381 | int fd; |
| 2382 | struct cl_host *ch; |
| 2383 | socklen_t socklen = sizeof(ns->addr); |
| 2384 | |
| 2385 | fd = my_accept(ns->listen_fd, (struct sockaddr *)&ns->addr, &socklen); |
| 2386 | if (fd < 0) { |
| 2387 | /* |
| 2388 | * This is OK: we just won't accept this connection, |
| 2389 | * nothing fatal. |
| 2390 | */ |
| 2391 | perror("accept"); |
| 2392 | } else { |
| 2393 | ch = net_find_client_host(ns, ns->addr.sin_addr); |
| 2394 | if (!ch) |
| 2395 | ch = net_add_client_host(ns, &ns->addr); |
| 2396 | |
| 2397 | ch_add_connection(ns, ch, fd); |
| 2398 | } |
| 2399 | } |
| 2400 | |
| 2401 | static struct devpath *nc_add_dpp(struct cl_conn *nc, |
| 2402 | struct blktrace_net_hdr *bnh, |
| 2403 | time_t connect_time) |
| 2404 | { |
| 2405 | int cpu; |
| 2406 | struct io_info *iop; |
| 2407 | struct devpath *dpp; |
| 2408 | |
| 2409 | dpp = malloc(sizeof(*dpp)); |
| 2410 | memset(dpp, 0, sizeof(*dpp)); |
| 2411 | |
| 2412 | dpp->buts_name = strdup(bnh->buts_name); |
| 2413 | dpp->path = strdup(bnh->buts_name); |
| 2414 | dpp->fd = -1; |
| 2415 | dpp->ch = nc->ch; |
| 2416 | dpp->cl_id = bnh->cl_id; |
| 2417 | dpp->cl_connect_time = connect_time; |
| 2418 | dpp->ncpus = nc->ncpus; |
| 2419 | dpp->stats = calloc(dpp->ncpus, sizeof(*dpp->stats)); |
| 2420 | memset(dpp->stats, 0, dpp->ncpus * sizeof(*dpp->stats)); |
| 2421 | |
| 2422 | list_add_tail(&dpp->head, &nc->ch->devpaths); |
| 2423 | nc->ch->ndevs++; |
| 2424 | |
| 2425 | dpp->ios = calloc(nc->ncpus, sizeof(*iop)); |
| 2426 | memset(dpp->ios, 0, ndevs * sizeof(*iop)); |
| 2427 | |
| 2428 | for (cpu = 0, iop = dpp->ios; cpu < nc->ncpus; cpu++, iop++) { |
| 2429 | iop->dpp = dpp; |
| 2430 | iop->nc = nc; |
| 2431 | init_mmap_info(&iop->mmap_info); |
| 2432 | |
| 2433 | if (iop_open(iop, cpu)) |
| 2434 | goto err; |
| 2435 | } |
| 2436 | |
| 2437 | return dpp; |
| 2438 | |
| 2439 | err: |
| 2440 | /* |
| 2441 | * Need to unravel what's been done... |
| 2442 | */ |
| 2443 | while (cpu >= 0) |
| 2444 | close_iop(&dpp->ios[cpu--]); |
| 2445 | dpp_free(dpp); |
| 2446 | |
| 2447 | return NULL; |
| 2448 | } |
| 2449 | |
| 2450 | static struct devpath *nc_find_dpp(struct cl_conn *nc, |
| 2451 | struct blktrace_net_hdr *bnh) |
| 2452 | { |
| 2453 | struct list_head *p; |
| 2454 | time_t connect_time = nc->connect_time; |
| 2455 | |
| 2456 | __list_for_each(p, &nc->ch->devpaths) { |
| 2457 | struct devpath *dpp = list_entry(p, struct devpath, head); |
| 2458 | |
| 2459 | if (!strcmp(dpp->buts_name, bnh->buts_name)) |
| 2460 | return dpp; |
| 2461 | |
| 2462 | if (dpp->cl_id == bnh->cl_id) |
| 2463 | connect_time = dpp->cl_connect_time; |
| 2464 | } |
| 2465 | |
| 2466 | return nc_add_dpp(nc, bnh, connect_time); |
| 2467 | } |
| 2468 | |
| 2469 | static void net_client_read_data(struct cl_conn *nc, struct devpath *dpp, |
| 2470 | struct blktrace_net_hdr *bnh) |
| 2471 | { |
| 2472 | int ret; |
| 2473 | struct io_info *iop = &dpp->ios[bnh->cpu]; |
| 2474 | struct mmap_info *mip = &iop->mmap_info; |
| 2475 | |
| 2476 | if (setup_mmap(iop->ofd, bnh->len, &iop->mmap_info, NULL)) { |
| 2477 | fprintf(stderr, "ncd(%s:%d): mmap failed\n", |
| 2478 | nc->ch->hostname, nc->fd); |
| 2479 | exit(1); |
| 2480 | } |
| 2481 | |
| 2482 | ret = net_recv_data(nc->fd, mip->fs_buf + mip->fs_off, bnh->len); |
| 2483 | if (ret > 0) { |
| 2484 | pdc_dr_update(dpp, bnh->cpu, ret); |
| 2485 | mip->fs_size += ret; |
| 2486 | mip->fs_off += ret; |
| 2487 | } else if (ret < 0) |
| 2488 | exit(1); |
| 2489 | } |
| 2490 | |
| 2491 | /* |
| 2492 | * Returns 1 if we closed a host - invalidates other polling information |
| 2493 | * that may be present. |
| 2494 | */ |
| 2495 | static int net_client_data(struct cl_conn *nc) |
| 2496 | { |
| 2497 | int ret; |
| 2498 | struct devpath *dpp; |
| 2499 | struct blktrace_net_hdr bnh; |
| 2500 | |
| 2501 | ret = net_get_header(nc, &bnh); |
| 2502 | if (ret == 0) |
| 2503 | return 0; |
| 2504 | |
| 2505 | if (ret < 0) { |
| 2506 | fprintf(stderr, "ncd(%d): header read failed\n", nc->fd); |
| 2507 | exit(1); |
| 2508 | } |
| 2509 | |
| 2510 | if (data_is_native == -1 && check_data_endianness(bnh.magic)) { |
| 2511 | fprintf(stderr, "ncd(%d): received data is bad\n", nc->fd); |
| 2512 | exit(1); |
| 2513 | } |
| 2514 | |
| 2515 | if (!data_is_native) { |
| 2516 | bnh.magic = be32_to_cpu(bnh.magic); |
| 2517 | bnh.cpu = be32_to_cpu(bnh.cpu); |
| 2518 | bnh.max_cpus = be32_to_cpu(bnh.max_cpus); |
| 2519 | bnh.len = be32_to_cpu(bnh.len); |
| 2520 | bnh.cl_id = be32_to_cpu(bnh.cl_id); |
| 2521 | bnh.buf_size = be32_to_cpu(bnh.buf_size); |
| 2522 | bnh.buf_nr = be32_to_cpu(bnh.buf_nr); |
| 2523 | bnh.page_size = be32_to_cpu(bnh.page_size); |
| 2524 | } |
| 2525 | |
| 2526 | if ((bnh.magic & 0xffffff00) != BLK_IO_TRACE_MAGIC) { |
| 2527 | fprintf(stderr, "ncd(%s:%d): bad data magic\n", |
| 2528 | nc->ch->hostname, nc->fd); |
| 2529 | exit(1); |
| 2530 | } |
| 2531 | |
| 2532 | if (nc->ncpus == -1) |
| 2533 | nc->ncpus = bnh.max_cpus; |
| 2534 | |
| 2535 | /* |
| 2536 | * len == 0 means the other end is sending us a new connection/dpp |
| 2537 | * len == 1 means that the other end signalled end-of-run |
| 2538 | */ |
| 2539 | dpp = nc_find_dpp(nc, &bnh); |
| 2540 | if (bnh.len == 0) { |
| 2541 | /* |
| 2542 | * Just adding in the dpp above is enough |
| 2543 | */ |
| 2544 | ack_open_close(nc->fd, dpp->buts_name); |
| 2545 | nc->ch->cl_opens++; |
| 2546 | } else if (bnh.len == 1) { |
| 2547 | /* |
| 2548 | * overload cpu count with dropped events |
| 2549 | */ |
| 2550 | dpp->drops = bnh.cpu; |
| 2551 | |
| 2552 | ack_open_close(nc->fd, dpp->buts_name); |
| 2553 | if (--nc->ch->cl_opens == 0) { |
| 2554 | show_stats(&nc->ch->devpaths); |
| 2555 | net_ch_remove(nc->ch, nc->ncpus); |
| 2556 | return 1; |
| 2557 | } |
| 2558 | } else |
| 2559 | net_client_read_data(nc, dpp, &bnh); |
| 2560 | |
| 2561 | return 0; |
| 2562 | } |
| 2563 | |
| 2564 | static void handle_client_data(struct net_server_s *ns, int events) |
| 2565 | { |
| 2566 | struct cl_conn *nc; |
| 2567 | struct pollfd *pfd; |
| 2568 | struct list_head *p, *q; |
| 2569 | |
| 2570 | pfd = &ns->pfds[1]; |
| 2571 | list_for_each_safe(p, q, &ns->conn_list) { |
| 2572 | if (pfd->revents & POLLIN) { |
| 2573 | nc = list_entry(p, struct cl_conn, ns_head); |
| 2574 | |
| 2575 | if (net_client_data(nc) || --events == 0) |
| 2576 | break; |
| 2577 | } |
| 2578 | pfd++; |
| 2579 | } |
| 2580 | } |
| 2581 | |
| 2582 | static void net_setup_pfds(struct net_server_s *ns) |
| 2583 | { |
| 2584 | struct pollfd *pfd; |
| 2585 | struct list_head *p; |
| 2586 | |
| 2587 | ns->pfds[0].fd = ns->listen_fd; |
| 2588 | ns->pfds[0].events = POLLIN; |
| 2589 | |
| 2590 | pfd = &ns->pfds[1]; |
| 2591 | __list_for_each(p, &ns->conn_list) { |
| 2592 | struct cl_conn *nc = list_entry(p, struct cl_conn, ns_head); |
| 2593 | |
| 2594 | pfd->fd = nc->fd; |
| 2595 | pfd->events = POLLIN; |
| 2596 | pfd++; |
| 2597 | } |
| 2598 | } |
| 2599 | |
| 2600 | static int net_server_handle_connections(struct net_server_s *ns) |
| 2601 | { |
| 2602 | int events; |
| 2603 | |
| 2604 | printf("server: waiting for connections...\n"); |
| 2605 | |
| 2606 | while (!done) { |
| 2607 | net_setup_pfds(ns); |
| 2608 | events = poll(ns->pfds, ns->connects + 1, -1); |
| 2609 | if (events < 0) { |
| 2610 | if (errno != EINTR) { |
| 2611 | perror("FATAL: poll error"); |
| 2612 | return 1; |
| 2613 | } |
| 2614 | } else if (events > 0) { |
| 2615 | if (ns->pfds[0].revents & POLLIN) { |
| 2616 | net_add_connection(ns); |
| 2617 | events--; |
| 2618 | } |
| 2619 | |
| 2620 | if (events) |
| 2621 | handle_client_data(ns, events); |
| 2622 | } |
| 2623 | } |
| 2624 | |
| 2625 | return 0; |
| 2626 | } |
| 2627 | |
| 2628 | static int net_server(void) |
| 2629 | { |
| 2630 | int fd, opt; |
| 2631 | int ret = 1; |
| 2632 | struct net_server_s net_server; |
| 2633 | struct net_server_s *ns = &net_server; |
| 2634 | |
| 2635 | memset(ns, 0, sizeof(*ns)); |
| 2636 | INIT_LIST_HEAD(&ns->ch_list); |
| 2637 | INIT_LIST_HEAD(&ns->conn_list); |
| 2638 | ns->pfds = malloc(sizeof(struct pollfd)); |
| 2639 | |
| 2640 | fd = my_socket(AF_INET, SOCK_STREAM, 0); |
| 2641 | if (fd < 0) { |
| 2642 | perror("server: socket"); |
| 2643 | goto out; |
| 2644 | } |
| 2645 | |
| 2646 | opt = 1; |
| 2647 | if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt)) < 0) { |
| 2648 | perror("setsockopt"); |
| 2649 | goto out; |
| 2650 | } |
| 2651 | |
| 2652 | memset(&ns->addr, 0, sizeof(ns->addr)); |
| 2653 | ns->addr.sin_family = AF_INET; |
| 2654 | ns->addr.sin_addr.s_addr = htonl(INADDR_ANY); |
| 2655 | ns->addr.sin_port = htons(net_port); |
| 2656 | |
| 2657 | if (bind(fd, (struct sockaddr *) &ns->addr, sizeof(ns->addr)) < 0) { |
| 2658 | perror("bind"); |
| 2659 | goto out; |
| 2660 | } |
| 2661 | |
| 2662 | if (listen(fd, 1) < 0) { |
| 2663 | perror("listen"); |
| 2664 | goto out; |
| 2665 | } |
| 2666 | |
| 2667 | /* |
| 2668 | * The actual server looping is done here: |
| 2669 | */ |
| 2670 | ns->listen_fd = fd; |
| 2671 | ret = net_server_handle_connections(ns); |
| 2672 | |
| 2673 | /* |
| 2674 | * Clean up and return... |
| 2675 | */ |
| 2676 | out: |
| 2677 | free(ns->pfds); |
| 2678 | return ret; |
| 2679 | } |
| 2680 | |
| 2681 | static int run_tracers(void) |
| 2682 | { |
| 2683 | atexit(exit_tracing); |
| 2684 | if (net_mode == Net_client) |
| 2685 | printf("blktrace: connecting to %s\n", hostname); |
| 2686 | |
| 2687 | if (setup_buts()) |
| 2688 | return 1; |
| 2689 | |
| 2690 | if (use_tracer_devpaths()) { |
| 2691 | if (setup_tracer_devpaths()) |
| 2692 | return 1; |
| 2693 | |
| 2694 | if (piped_output) |
| 2695 | handle_list = handle_list_file; |
| 2696 | else |
| 2697 | handle_list = handle_list_net; |
| 2698 | } |
| 2699 | |
| 2700 | start_tracers(); |
| 2701 | if (nthreads_running == ncpus) { |
| 2702 | unblock_tracers(); |
| 2703 | start_buts(); |
| 2704 | if (net_mode == Net_client) |
| 2705 | printf("blktrace: connected!\n"); |
| 2706 | if (stop_watch) |
| 2707 | alarm(stop_watch); |
| 2708 | } else |
| 2709 | stop_tracers(); |
| 2710 | |
| 2711 | wait_tracers(); |
| 2712 | if (nthreads_running == ncpus) |
| 2713 | show_stats(&devpaths); |
| 2714 | if (net_client_use_send()) |
| 2715 | close_client_connections(); |
| 2716 | del_tracers(); |
| 2717 | |
| 2718 | return 0; |
| 2719 | } |
| 2720 | |
| 2721 | static cpu_set_t *get_online_cpus(void) |
| 2722 | { |
| 2723 | FILE *cpus; |
| 2724 | cpu_set_t *set; |
| 2725 | size_t alloc_size; |
| 2726 | int cpuid, prevcpuid = -1; |
| 2727 | char nextch; |
| 2728 | int n, ncpu, curcpu = 0; |
| 2729 | int *cpu_nums; |
| 2730 | |
| 2731 | ncpu = sysconf(_SC_NPROCESSORS_CONF); |
| 2732 | if (ncpu < 0) |
| 2733 | return NULL; |
| 2734 | |
| 2735 | cpu_nums = malloc(sizeof(int)*ncpu); |
| 2736 | if (!cpu_nums) { |
| 2737 | errno = ENOMEM; |
| 2738 | return NULL; |
| 2739 | } |
| 2740 | |
| 2741 | /* |
| 2742 | * There is no way to easily get maximum CPU number. So we have to |
| 2743 | * parse the file first to find it out and then create appropriate |
| 2744 | * cpuset |
| 2745 | */ |
| 2746 | cpus = my_fopen("/sys/devices/system/cpu/online", "r"); |
| 2747 | for (;;) { |
| 2748 | n = fscanf(cpus, "%d%c", &cpuid, &nextch); |
| 2749 | if (n <= 0) |
| 2750 | break; |
| 2751 | if (n == 2 && nextch == '-') { |
| 2752 | prevcpuid = cpuid; |
| 2753 | continue; |
| 2754 | } |
| 2755 | if (prevcpuid == -1) |
| 2756 | prevcpuid = cpuid; |
| 2757 | while (prevcpuid <= cpuid) { |
| 2758 | /* More CPUs listed than configured? */ |
| 2759 | if (curcpu >= ncpu) { |
| 2760 | errno = EINVAL; |
| 2761 | return NULL; |
| 2762 | } |
| 2763 | cpu_nums[curcpu++] = prevcpuid++; |
| 2764 | } |
| 2765 | prevcpuid = -1; |
| 2766 | } |
| 2767 | fclose(cpus); |
| 2768 | |
| 2769 | ncpu = curcpu; |
| 2770 | max_cpus = cpu_nums[ncpu - 1] + 1; |
| 2771 | |
| 2772 | /* Now that we have maximum cpu number, create a cpuset */ |
| 2773 | set = CPU_ALLOC(max_cpus); |
| 2774 | if (!set) { |
| 2775 | errno = ENOMEM; |
| 2776 | return NULL; |
| 2777 | } |
| 2778 | alloc_size = CPU_ALLOC_SIZE(max_cpus); |
| 2779 | CPU_ZERO_S(alloc_size, set); |
| 2780 | |
| 2781 | for (curcpu = 0; curcpu < ncpu; curcpu++) |
| 2782 | CPU_SET_S(cpu_nums[curcpu], alloc_size, set); |
| 2783 | |
| 2784 | free(cpu_nums); |
| 2785 | |
| 2786 | return set; |
| 2787 | } |
| 2788 | |
| 2789 | int main(int argc, char *argv[]) |
| 2790 | { |
| 2791 | int ret = 0; |
| 2792 | |
| 2793 | setlocale(LC_NUMERIC, "en_US"); |
| 2794 | pagesize = getpagesize(); |
| 2795 | online_cpus = get_online_cpus(); |
| 2796 | if (!online_cpus) { |
| 2797 | fprintf(stderr, "cannot get online cpus %d/%s\n", |
| 2798 | errno, strerror(errno)); |
| 2799 | ret = 1; |
| 2800 | goto out; |
| 2801 | } else if (handle_args(argc, argv)) { |
| 2802 | ret = 1; |
| 2803 | goto out; |
| 2804 | } |
| 2805 | |
| 2806 | ncpus = CPU_COUNT_S(CPU_ALLOC_SIZE(max_cpus), online_cpus); |
| 2807 | if (ndevs > 1 && output_name && strcmp(output_name, "-") != 0) { |
| 2808 | fprintf(stderr, "-o not supported with multiple devices\n"); |
| 2809 | ret = 1; |
| 2810 | goto out; |
| 2811 | } |
| 2812 | |
| 2813 | signal(SIGINT, handle_sigint); |
| 2814 | signal(SIGHUP, handle_sigint); |
| 2815 | signal(SIGTERM, handle_sigint); |
| 2816 | signal(SIGALRM, handle_sigint); |
| 2817 | signal(SIGPIPE, SIG_IGN); |
| 2818 | |
| 2819 | if (kill_running_trace) { |
| 2820 | struct devpath *dpp; |
| 2821 | struct list_head *p; |
| 2822 | |
| 2823 | __list_for_each(p, &devpaths) { |
| 2824 | dpp = list_entry(p, struct devpath, head); |
| 2825 | if (__stop_trace(dpp->fd)) { |
| 2826 | fprintf(stderr, |
| 2827 | "BLKTRACETEARDOWN %s failed: %d/%s\n", |
| 2828 | dpp->path, errno, strerror(errno)); |
| 2829 | } |
| 2830 | } |
| 2831 | } else if (net_mode == Net_server) { |
| 2832 | if (output_name) { |
| 2833 | fprintf(stderr, "-o ignored in server mode\n"); |
| 2834 | output_name = NULL; |
| 2835 | } |
| 2836 | ret = net_server(); |
| 2837 | } else |
| 2838 | ret = run_tracers(); |
| 2839 | |
| 2840 | out: |
| 2841 | if (pfp) |
| 2842 | fclose(pfp); |
| 2843 | rel_devpaths(); |
| 2844 | return ret; |
| 2845 | } |