source3/modules/vfs_aio_fork.c

   1 /*
   2  * Simulate the Posix AIO using mmap/fork
   3  *
   4  * Copyright (C) Volker Lendecke 2008
   5  * Copyright (C) Jeremy Allison 2010
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License
  18  * along with this program; if not, write to the Free Software
  19  * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  20  */
  21
  22 #include "includes.h"
  23 #include "system/filesys.h"
  24 #include "system/shmem.h"
  25 #include "smbd/smbd.h"
  26 #include "smbd/globals.h"
  27 #include "lib/async_req/async_sock.h"
  28 #include "lib/util/tevent_unix.h"
  29 #include "lib/sys_rw.h"
  30 #include "lib/sys_rw_data.h"
  31 #include "lib/msghdr.h"
  32
  33 #if !defined(HAVE_STRUCT_MSGHDR_MSG_CONTROL) && !defined(HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS)
  34 # error Can not pass file descriptors
  35 #endif
  36
  37 #undef recvmsg
  38
  39 #ifndef MAP_FILE
  40 #define MAP_FILE 0
  41 #endif
  42
  43 struct aio_fork_config {
  44         bool erratic_testing_mode;
  45 };
  46
  47 struct mmap_area {
  48         size_t size;
  49         volatile void *ptr;
  50 };
  51
  52 static int mmap_area_destructor(struct mmap_area *area)
  53 {
  54         munmap((void *)area->ptr, area->size);
  55         return 0;
  56 }
  57
  58 static struct mmap_area *mmap_area_init(TALLOC_CTX *mem_ctx, size_t size)
  59 {
  60         struct mmap_area *result;
  61         int fd;
  62
  63         result = talloc(mem_ctx, struct mmap_area);
  64         if (result == NULL) {
  65                 DEBUG(0, ("talloc failed\n"));
  66                 goto fail;
  67         }
  68
  69         fd = open("/dev/zero", O_RDWR);
  70         if (fd == -1) {
  71                 DEBUG(3, ("open(\"/dev/zero\") failed: %s\n",
  72                           strerror(errno)));
  73                 goto fail;
  74         }
  75
  76         result->ptr = mmap(NULL, size, PROT_READ|PROT_WRITE,
  77                            MAP_SHARED|MAP_FILE, fd, 0);
  78         close(fd);
  79         if (result->ptr == MAP_FAILED) {
  80                 DEBUG(1, ("mmap failed: %s\n", strerror(errno)));
  81                 goto fail;
  82         }
  83
  84         result->size = size;
  85         talloc_set_destructor(result, mmap_area_destructor);
  86
  87         return result;
  88
  89 fail:
  90         TALLOC_FREE(result);
  91         return NULL;
  92 }
  93
  94 enum cmd_type {
  95         READ_CMD,
  96         WRITE_CMD,
  97         FSYNC_CMD
  98 };
  99
 100 static const char *cmd_type_str(enum cmd_type cmd)
 101 {
 102         const char *result;
 103
 104         switch (cmd) {
 105         case READ_CMD:
 106                 result = "READ";
 107                 break;
 108         case WRITE_CMD:
 109                 result = "WRITE";
 110                 break;
 111         case FSYNC_CMD:
 112                 result = "FSYNC";
 113                 break;
 114         default:
 115                 result = "<UNKNOWN>";
 116                 break;
 117         }
 118         return result;
 119 }
 120
 121 struct rw_cmd {
 122         size_t n;
 123         off_t offset;
 124         enum cmd_type cmd;
 125         bool erratic_testing_mode;
 126 };
 127
 128 struct rw_ret {
 129         ssize_t size;
 130         int ret_errno;
 131 };
 132
 133 struct aio_child_list;
 134
 135 struct aio_child {
 136         struct aio_child *prev, *next;
 137         struct aio_child_list *list;
 138         pid_t pid;
 139         int sockfd;
 140         struct mmap_area *map;
 141         bool dont_delete;       /* Marked as in use since last cleanup */
 142         bool busy;
 143 };
 144
 145 struct aio_child_list {
 146         struct aio_child *children;
 147         struct tevent_timer *cleanup_event;
 148 };
 149
 150 static void free_aio_children(void **p)
 151 {
 152         TALLOC_FREE(*p);
 153 }
 154
 155 static ssize_t read_fd(int fd, void *ptr, size_t nbytes, int *recvfd)
 156 {
 157         struct msghdr msg;
 158         struct iovec iov[1];
 159         ssize_t n;
 160         size_t bufsize = msghdr_prep_recv_fds(NULL, NULL, 0, 1);
 161         uint8_t buf[bufsize];
 162
 163         msghdr_prep_recv_fds(&msg, buf, bufsize, 1);
 164
 165         msg.msg_name = NULL;
 166         msg.msg_namelen = 0;
 167
 168         iov[0].iov_base = (void *)ptr;
 169         iov[0].iov_len = nbytes;
 170         msg.msg_iov = iov;
 171         msg.msg_iovlen = 1;
 172
 173         do {
 174                 n = recvmsg(fd, &msg, 0);
 175         } while ((n == -1) && (errno == EINTR));
 176
 177         if (n <= 0) {
 178                 return n;
 179         }
 180
 181         {
 182                 size_t num_fds = msghdr_extract_fds(&msg, NULL, 0);
 183                 int fds[num_fds];
 184
 185                 msghdr_extract_fds(&msg, fds, num_fds);
 186
 187                 if (num_fds != 1) {
 188                         size_t i;
 189
 190                         for (i=0; i<num_fds; i++) {
 191                                 close(fds[i]);
 192                         }
 193
 194                         *recvfd = -1;
 195                         return n;
 196                 }
 197
 198                 *recvfd = fds[0];
 199         }
 200
 201         return(n);
 202 }
 203
 204 static ssize_t write_fd(int fd, void *ptr, size_t nbytes, int sendfd)
 205 {
 206         struct msghdr msg = {0};
 207         size_t bufsize = msghdr_prep_fds(NULL, NULL, 0, &sendfd, 1);
 208         uint8_t buf[bufsize];
 209         struct iovec iov;
 210         ssize_t sent;
 211
 212         msghdr_prep_fds(&msg, buf, bufsize, &sendfd, 1);
 213
 214         iov.iov_base = (void *)ptr;
 215         iov.iov_len = nbytes;
 216         msg.msg_iov = &iov;
 217         msg.msg_iovlen = 1;
 218
 219         do {
 220                 sent = sendmsg(fd, &msg, 0);
 221         } while ((sent == -1) && (errno == EINTR));
 222
 223         return sent;
 224 }
 225
 226 static void aio_child_cleanup(struct tevent_context *event_ctx,
 227                               struct tevent_timer *te,
 228                               struct timeval now,
 229                               void *private_data)
 230 {
 231         struct aio_child_list *list = talloc_get_type_abort(
 232                 private_data, struct aio_child_list);
 233         struct aio_child *child, *next;
 234
 235         TALLOC_FREE(list->cleanup_event);
 236
 237         for (child = list->children; child != NULL; child = next) {
 238                 next = child->next;
 239
 240                 if (child->busy) {
 241                         DEBUG(10, ("child %d currently active\n",
 242                                    (int)child->pid));
 243                         continue;
 244                 }
 245
 246                 if (child->dont_delete) {
 247                         DEBUG(10, ("Child %d was active since last cleanup\n",
 248                                    (int)child->pid));
 249                         child->dont_delete = false;
 250                         continue;
 251                 }
 252
 253                 DEBUG(10, ("Child %d idle for more than 30 seconds, "
 254                            "deleting\n", (int)child->pid));
 255
 256                 TALLOC_FREE(child);
 257                 child = next;
 258         }
 259
 260         if (list->children != NULL) {
 261                 /*
 262                  * Re-schedule the next cleanup round
 263                  */
 264                 list->cleanup_event = tevent_add_timer(server_event_context(), list,
 265                                                       timeval_add(&now, 30, 0),
 266                                                       aio_child_cleanup, list);
 267
 268         }
 269 }
 270
 271 static struct aio_child_list *init_aio_children(struct vfs_handle_struct *handle)
 272 {
 273         struct aio_child_list *data = NULL;
 274
 275         if (SMB_VFS_HANDLE_TEST_DATA(handle)) {
 276                 SMB_VFS_HANDLE_GET_DATA(handle, data, struct aio_child_list,
 277                                         return NULL);
 278         }
 279
 280         if (data == NULL) {
 281                 data = talloc_zero(NULL, struct aio_child_list);
 282                 if (data == NULL) {
 283                         return NULL;
 284                 }
 285         }
 286
 287         /*
 288          * Regardless of whether the child_list had been around or not, make
 289          * sure that we have a cleanup timed event. This timed event will
 290          * delete itself when it finds that no children are around anymore.
 291          */
 292
 293         if (data->cleanup_event == NULL) {
 294                 data->cleanup_event = tevent_add_timer(server_event_context(), data,
 295                                                       timeval_current_ofs(30, 0),
 296                                                       aio_child_cleanup, data);
 297                 if (data->cleanup_event == NULL) {
 298                         TALLOC_FREE(data);
 299                         return NULL;
 300                 }
 301         }
 302
 303         if (!SMB_VFS_HANDLE_TEST_DATA(handle)) {
 304                 SMB_VFS_HANDLE_SET_DATA(handle, data, free_aio_children,
 305                                         struct aio_child_list, return False);
 306         }
 307
 308         return data;
 309 }
 310
 311 static void aio_child_loop(int sockfd, struct mmap_area *map)
 312 {
 313         while (true) {
 314                 int fd = -1;
 315                 ssize_t ret;
 316                 struct rw_cmd cmd_struct;
 317                 struct rw_ret ret_struct;
 318
 319                 ret = read_fd(sockfd, &cmd_struct, sizeof(cmd_struct), &fd);
 320                 if (ret != sizeof(cmd_struct)) {
 321                         DEBUG(10, ("read_fd returned %d: %s\n", (int)ret,
 322                                    strerror(errno)));
 323                         exit(1);
 324                 }
 325
 326                 DEBUG(10, ("aio_child_loop: %s %d bytes at %d from fd %d\n",
 327                            cmd_type_str(cmd_struct.cmd),
 328                            (int)cmd_struct.n, (int)cmd_struct.offset, fd));
 329
 330                 if (cmd_struct.erratic_testing_mode) {
 331                         /*
 332                          * For developer testing, we want erratic behaviour for
 333                          * async I/O times
 334                          */
 335                         uint8_t randval;
 336                         unsigned msecs;
 337                         /*
 338                          * use generate_random_buffer, we just forked from a
 339                          * common parent state
 340                          */
 341                         generate_random_buffer(&randval, sizeof(randval));
 342                         msecs = randval + 20;
 343                         DEBUG(10, ("delaying for %u msecs\n", msecs));
 344                         smb_msleep(msecs);
 345                 }
 346
 347                 ZERO_STRUCT(ret_struct);
 348
 349                 switch (cmd_struct.cmd) {
 350                 case READ_CMD:
 351                         ret_struct.size = sys_pread(
 352                                 fd, (void *)map->ptr, cmd_struct.n,
 353                                 cmd_struct.offset);
 354 #if 0
 355 /* This breaks "make test" when run with aio_fork module. */
 356 #ifdef DEVELOPER
 357                         ret_struct.size = MAX(1, ret_struct.size * 0.9);
 358 #endif
 359 #endif
 360                         break;
 361                 case WRITE_CMD:
 362                         ret_struct.size = sys_pwrite(
 363                                 fd, (void *)map->ptr, cmd_struct.n,
 364                                 cmd_struct.offset);
 365                         break;
 366                 case FSYNC_CMD:
 367                         ret_struct.size = fsync(fd);
 368                         break;
 369                 default:
 370                         ret_struct.size = -1;
 371                         errno = EINVAL;
 372                 }
 373
 374                 DEBUG(10, ("aio_child_loop: syscall returned %d\n",
 375                            (int)ret_struct.size));
 376
 377                 if (ret_struct.size == -1) {
 378                         ret_struct.ret_errno = errno;
 379                 }
 380
 381                 /*
 382                  * Close the fd before telling our parent we're done. The
 383                  * parent might close and re-open the file very quickly, and
 384                  * with system-level share modes (GPFS) we would get an
 385                  * unjustified SHARING_VIOLATION.
 386                  */
 387                 close(fd);
 388
 389                 ret = write_data(sockfd, (char *)&ret_struct,
 390                                  sizeof(ret_struct));
 391                 if (ret != sizeof(ret_struct)) {
 392                         DEBUG(10, ("could not write ret_struct: %s\n",
 393                                    strerror(errno)));
 394                         exit(2);
 395                 }
 396         }
 397 }
 398
 399 static int aio_child_destructor(struct aio_child *child)
 400 {
 401         char c=0;
 402
 403         SMB_ASSERT(!child->busy);
 404
 405         DEBUG(10, ("aio_child_destructor: removing child %d on fd %d\n",
 406                         child->pid, child->sockfd));
 407
 408         /*
 409          * closing the sockfd makes the child not return from recvmsg() on RHEL
 410          * 5.5 so instead force the child to exit by writing bad data to it
 411          */
 412         write(child->sockfd, &c, sizeof(c));
 413         close(child->sockfd);
 414         DLIST_REMOVE(child->list->children, child);
 415         return 0;
 416 }
 417
 418 /*
 419  * We have to close all fd's in open files, we might incorrectly hold a system
 420  * level share mode on a file.
 421  */
 422
 423 static struct files_struct *close_fsp_fd(struct files_struct *fsp,
 424                                          void *private_data)
 425 {
 426         if ((fsp->fh != NULL) && (fsp->fh->fd != -1)) {
 427                 close(fsp->fh->fd);
 428                 fsp->fh->fd = -1;
 429         }
 430         return NULL;
 431 }
 432
 433 static int create_aio_child(struct smbd_server_connection *sconn,
 434                             struct aio_child_list *children,
 435                             size_t map_size,
 436                             struct aio_child **presult)
 437 {
 438         struct aio_child *result;
 439         int fdpair[2];
 440         int ret;
 441
 442         fdpair[0] = fdpair[1] = -1;
 443
 444         result = talloc_zero(children, struct aio_child);
 445         if (result == NULL) {
 446                 return ENOMEM;
 447         }
 448
 449         if (socketpair(AF_UNIX, SOCK_STREAM, 0, fdpair) == -1) {
 450                 ret = errno;
 451                 DEBUG(10, ("socketpair() failed: %s\n", strerror(errno)));
 452                 goto fail;
 453         }
 454
 455         DEBUG(10, ("fdpair = %d/%d\n", fdpair[0], fdpair[1]));
 456
 457         result->map = mmap_area_init(result, map_size);
 458         if (result->map == NULL) {
 459                 ret = errno;
 460                 DEBUG(0, ("Could not create mmap area\n"));
 461                 goto fail;
 462         }
 463
 464         result->pid = fork();
 465         if (result->pid == -1) {
 466                 ret = errno;
 467                 DEBUG(0, ("fork failed: %s\n", strerror(errno)));
 468                 goto fail;
 469         }
 470
 471         if (result->pid == 0) {
 472                 close(fdpair[0]);
 473                 result->sockfd = fdpair[1];
 474                 files_forall(sconn, close_fsp_fd, NULL);
 475                 aio_child_loop(result->sockfd, result->map);
 476         }
 477
 478         DEBUG(10, ("Child %d created with sockfd %d\n",
 479                         result->pid, fdpair[0]));
 480
 481         result->sockfd = fdpair[0];
 482         close(fdpair[1]);
 483
 484         result->list = children;
 485         DLIST_ADD(children->children, result);
 486
 487         talloc_set_destructor(result, aio_child_destructor);
 488
 489         *presult = result;
 490
 491         return 0;
 492
 493  fail:
 494         if (fdpair[0] != -1) close(fdpair[0]);
 495         if (fdpair[1] != -1) close(fdpair[1]);
 496         TALLOC_FREE(result);
 497
 498         return ret;
 499 }
 500
 501 static int get_idle_child(struct vfs_handle_struct *handle,
 502                           struct aio_child **pchild)
 503 {
 504         struct aio_child_list *children;
 505         struct aio_child *child;
 506
 507         children = init_aio_children(handle);
 508         if (children == NULL) {
 509                 return ENOMEM;
 510         }
 511
 512         for (child = children->children; child != NULL; child = child->next) {
 513                 if (!child->busy) {
 514                         break;
 515                 }
 516         }
 517
 518         if (child == NULL) {
 519                 int ret;
 520
 521                 DEBUG(10, ("no idle child found, creating new one\n"));
 522
 523                 ret = create_aio_child(handle->conn->sconn, children,
 524                                           128*1024, &child);
 525                 if (ret != 0) {
 526                         DEBUG(10, ("create_aio_child failed: %s\n",
 527                                    strerror(errno)));
 528                         return ret;
 529                 }
 530         }
 531
 532         child->dont_delete = true;
 533         child->busy = true;
 534
 535         *pchild = child;
 536         return 0;
 537 }
 538
 539 struct aio_fork_pread_state {
 540         struct aio_child *child;
 541         ssize_t ret;
 542         int err;
 543 };
 544
 545 static void aio_fork_pread_done(struct tevent_req *subreq);
 546
 547 static struct tevent_req *aio_fork_pread_send(struct vfs_handle_struct *handle,
 548                                               TALLOC_CTX *mem_ctx,
 549                                               struct tevent_context *ev,
 550                                               struct files_struct *fsp,
 551                                               void *data,
 552                                               size_t n, off_t offset)
 553 {
 554         struct tevent_req *req, *subreq;
 555         struct aio_fork_pread_state *state;
 556         struct rw_cmd cmd;
 557         ssize_t written;
 558         int err;
 559         struct aio_fork_config *config;
 560
 561         SMB_VFS_HANDLE_GET_DATA(handle, config,
 562                                 struct aio_fork_config,
 563                                 return NULL);
 564
 565         req = tevent_req_create(mem_ctx, &state, struct aio_fork_pread_state);
 566         if (req == NULL) {
 567                 return NULL;
 568         }
 569
 570         if (n > 128*1024) {
 571                 /* TODO: support variable buffers */
 572                 tevent_req_error(req, EINVAL);
 573                 return tevent_req_post(req, ev);
 574         }
 575
 576         err = get_idle_child(handle, &state->child);
 577         if (err != 0) {
 578                 tevent_req_error(req, err);
 579                 return tevent_req_post(req, ev);
 580         }
 581
 582         ZERO_STRUCT(cmd);
 583         cmd.n = n;
 584         cmd.offset = offset;
 585         cmd.cmd = READ_CMD;
 586         cmd.erratic_testing_mode = config->erratic_testing_mode;
 587
 588         DEBUG(10, ("sending fd %d to child %d\n", fsp->fh->fd,
 589                    (int)state->child->pid));
 590
 591         /*
 592          * Not making this async. We're writing into an empty unix
 593          * domain socket. This should never block.
 594          */
 595         written = write_fd(state->child->sockfd, &cmd, sizeof(cmd),
 596                            fsp->fh->fd);
 597         if (written == -1) {
 598                 err = errno;
 599
 600                 TALLOC_FREE(state->child);
 601
 602                 DEBUG(10, ("write_fd failed: %s\n", strerror(err)));
 603                 tevent_req_error(req, err);
 604                 return tevent_req_post(req, ev);
 605         }
 606
 607         subreq = read_packet_send(state, ev, state->child->sockfd,
 608                                   sizeof(struct rw_ret), NULL, NULL);
 609         if (tevent_req_nomem(subreq, req)) {
 610                 TALLOC_FREE(state->child); /* we sent sth down */
 611                 return tevent_req_post(req, ev);
 612         }
 613         tevent_req_set_callback(subreq, aio_fork_pread_done, req);
 614         return req;
 615 }
 616
 617 static void aio_fork_pread_done(struct tevent_req *subreq)
 618 {
 619         struct tevent_req *req = tevent_req_callback_data(
 620                 subreq, struct tevent_req);
 621         struct aio_fork_pread_state *state = tevent_req_data(
 622                 req, struct aio_fork_pread_state);
 623         ssize_t nread;
 624         uint8_t *buf;
 625         int err;
 626         struct rw_ret *retbuf;
 627
 628         nread = read_packet_recv(subreq, talloc_tos(), &buf, &err);
 629         TALLOC_FREE(subreq);
 630         if (nread == -1) {
 631                 TALLOC_FREE(state->child);
 632                 tevent_req_error(req, err);
 633                 return;
 634         }
 635
 636         state->child->busy = false;
 637
 638         retbuf = (struct rw_ret *)buf;
 639         state->ret = retbuf->size;
 640         state->err = retbuf->ret_errno;
 641         tevent_req_done(req);
 642 }
 643
 644 static ssize_t aio_fork_pread_recv(struct tevent_req *req, int *err)
 645 {
 646         struct aio_fork_pread_state *state = tevent_req_data(
 647                 req, struct aio_fork_pread_state);
 648
 649         if (tevent_req_is_unix_error(req, err)) {
 650                 return -1;
 651         }
 652         if (state->ret == -1) {
 653                 *err = state->err;
 654         }
 655         return state->ret;
 656 }
 657
 658 struct aio_fork_pwrite_state {
 659         struct aio_child *child;
 660         ssize_t ret;
 661         int err;
 662 };
 663
 664 static void aio_fork_pwrite_done(struct tevent_req *subreq);
 665
 666 static struct tevent_req *aio_fork_pwrite_send(
 667         struct vfs_handle_struct *handle, TALLOC_CTX *mem_ctx,
 668         struct tevent_context *ev, struct files_struct *fsp,
 669         const void *data, size_t n, off_t offset)
 670 {
 671         struct tevent_req *req, *subreq;
 672         struct aio_fork_pwrite_state *state;
 673         struct rw_cmd cmd;
 674         ssize_t written;
 675         int err;
 676         struct aio_fork_config *config;
 677         SMB_VFS_HANDLE_GET_DATA(handle, config,
 678                                 struct aio_fork_config,
 679                                 return NULL);
 680
 681         req = tevent_req_create(mem_ctx, &state, struct aio_fork_pwrite_state);
 682         if (req == NULL) {
 683                 return NULL;
 684         }
 685
 686         if (n > 128*1024) {
 687                 /* TODO: support variable buffers */
 688                 tevent_req_error(req, EINVAL);
 689                 return tevent_req_post(req, ev);
 690         }
 691
 692         err = get_idle_child(handle, &state->child);
 693         if (err != 0) {
 694                 tevent_req_error(req, err);
 695                 return tevent_req_post(req, ev);
 696         }
 697
 698         ZERO_STRUCT(cmd);
 699         cmd.n = n;
 700         cmd.offset = offset;
 701         cmd.cmd = WRITE_CMD;
 702         cmd.erratic_testing_mode = config->erratic_testing_mode;
 703
 704         DEBUG(10, ("sending fd %d to child %d\n", fsp->fh->fd,
 705                    (int)state->child->pid));
 706
 707         /*
 708          * Not making this async. We're writing into an empty unix
 709          * domain socket. This should never block.
 710          */
 711         written = write_fd(state->child->sockfd, &cmd, sizeof(cmd),
 712                            fsp->fh->fd);
 713         if (written == -1) {
 714                 err = errno;
 715
 716                 TALLOC_FREE(state->child);
 717
 718                 DEBUG(10, ("write_fd failed: %s\n", strerror(err)));
 719                 tevent_req_error(req, err);
 720                 return tevent_req_post(req, ev);
 721         }
 722
 723         subreq = read_packet_send(state, ev, state->child->sockfd,
 724                                   sizeof(struct rw_ret), NULL, NULL);
 725         if (tevent_req_nomem(subreq, req)) {
 726                 TALLOC_FREE(state->child); /* we sent sth down */
 727                 return tevent_req_post(req, ev);
 728         }
 729         tevent_req_set_callback(subreq, aio_fork_pwrite_done, req);
 730         return req;
 731 }
 732
 733 static void aio_fork_pwrite_done(struct tevent_req *subreq)
 734 {
 735         struct tevent_req *req = tevent_req_callback_data(
 736                 subreq, struct tevent_req);
 737         struct aio_fork_pwrite_state *state = tevent_req_data(
 738                 req, struct aio_fork_pwrite_state);
 739         ssize_t nread;
 740         uint8_t *buf;
 741         int err;
 742         struct rw_ret *retbuf;
 743
 744         nread = read_packet_recv(subreq, talloc_tos(), &buf, &err);
 745         TALLOC_FREE(subreq);
 746         if (nread == -1) {
 747                 TALLOC_FREE(state->child);
 748                 tevent_req_error(req, err);
 749                 return;
 750         }
 751
 752         state->child->busy = false;
 753
 754         retbuf = (struct rw_ret *)buf;
 755         state->ret = retbuf->size;
 756         state->err = retbuf->ret_errno;
 757         tevent_req_done(req);
 758 }
 759
 760 static ssize_t aio_fork_pwrite_recv(struct tevent_req *req, int *err)
 761 {
 762         struct aio_fork_pwrite_state *state = tevent_req_data(
 763                 req, struct aio_fork_pwrite_state);
 764
 765         if (tevent_req_is_unix_error(req, err)) {
 766                 return -1;
 767         }
 768         if (state->ret == -1) {
 769                 *err = state->err;
 770         }
 771         return state->ret;
 772 }
 773
 774 struct aio_fork_fsync_state {
 775         struct aio_child *child;
 776         ssize_t ret;
 777         int err;
 778 };
 779
 780 static void aio_fork_fsync_done(struct tevent_req *subreq);
 781
 782 static struct tevent_req *aio_fork_fsync_send(
 783         struct vfs_handle_struct *handle, TALLOC_CTX *mem_ctx,
 784         struct tevent_context *ev, struct files_struct *fsp)
 785 {
 786         struct tevent_req *req, *subreq;
 787         struct aio_fork_fsync_state *state;
 788         struct rw_cmd cmd;
 789         ssize_t written;
 790         int err;
 791         struct aio_fork_config *config;
 792
 793         SMB_VFS_HANDLE_GET_DATA(handle, config,
 794                                 struct aio_fork_config,
 795                                 return NULL);
 796
 797         req = tevent_req_create(mem_ctx, &state, struct aio_fork_fsync_state);
 798         if (req == NULL) {
 799                 return NULL;
 800         }
 801
 802         err = get_idle_child(handle, &state->child);
 803         if (err != 0) {
 804                 tevent_req_error(req, err);
 805                 return tevent_req_post(req, ev);
 806         }
 807
 808         ZERO_STRUCT(cmd);
 809         cmd.cmd = FSYNC_CMD;
 810         cmd.erratic_testing_mode = config->erratic_testing_mode;
 811
 812         DEBUG(10, ("sending fd %d to child %d\n", fsp->fh->fd,
 813                    (int)state->child->pid));
 814
 815         /*
 816          * Not making this async. We're writing into an empty unix
 817          * domain socket. This should never block.
 818          */
 819         written = write_fd(state->child->sockfd, &cmd, sizeof(cmd),
 820                            fsp->fh->fd);
 821         if (written == -1) {
 822                 err = errno;
 823
 824                 TALLOC_FREE(state->child);
 825
 826                 DEBUG(10, ("write_fd failed: %s\n", strerror(err)));
 827                 tevent_req_error(req, err);
 828                 return tevent_req_post(req, ev);
 829         }
 830
 831         subreq = read_packet_send(state, ev, state->child->sockfd,
 832                                   sizeof(struct rw_ret), NULL, NULL);
 833         if (tevent_req_nomem(subreq, req)) {
 834                 TALLOC_FREE(state->child); /* we sent sth down */
 835                 return tevent_req_post(req, ev);
 836         }
 837         tevent_req_set_callback(subreq, aio_fork_fsync_done, req);
 838         return req;
 839 }
 840
 841 static void aio_fork_fsync_done(struct tevent_req *subreq)
 842 {
 843         struct tevent_req *req = tevent_req_callback_data(
 844                 subreq, struct tevent_req);
 845         struct aio_fork_fsync_state *state = tevent_req_data(
 846                 req, struct aio_fork_fsync_state);
 847         ssize_t nread;
 848         uint8_t *buf;
 849         int err;
 850         struct rw_ret *retbuf;
 851
 852         nread = read_packet_recv(subreq, talloc_tos(), &buf, &err);
 853         TALLOC_FREE(subreq);
 854         if (nread == -1) {
 855                 TALLOC_FREE(state->child);
 856                 tevent_req_error(req, err);
 857                 return;
 858         }
 859
 860         state->child->busy = false;
 861
 862         retbuf = (struct rw_ret *)buf;
 863         state->ret = retbuf->size;
 864         state->err = retbuf->ret_errno;
 865         tevent_req_done(req);
 866 }
 867
 868 static int aio_fork_fsync_recv(struct tevent_req *req, int *err)
 869 {
 870         struct aio_fork_fsync_state *state = tevent_req_data(
 871                 req, struct aio_fork_fsync_state);
 872
 873         if (tevent_req_is_unix_error(req, err)) {
 874                 return -1;
 875         }
 876         if (state->ret == -1) {
 877                 *err = state->err;
 878         }
 879         return state->ret;
 880 }
 881
 882 static int aio_fork_connect(vfs_handle_struct *handle, const char *service,
 883                             const char *user)
 884 {
 885         int ret;
 886         struct aio_fork_config *config;
 887         ret = SMB_VFS_NEXT_CONNECT(handle, service, user);
 888
 889         if (ret < 0) {
 890                 return ret;
 891         }
 892
 893         config = talloc_zero(handle->conn, struct aio_fork_config);
 894         if (!config) {
 895                 SMB_VFS_NEXT_DISCONNECT(handle);
 896                 DEBUG(0, ("talloc_zero() failed\n"));
 897                 return -1;
 898         }
 899
 900         config->erratic_testing_mode = lp_parm_bool(SNUM(handle->conn), "vfs_aio_fork",
 901                                                     "erratic_testing_mode", false);
 902
 903         SMB_VFS_HANDLE_SET_DATA(handle, config,
 904                                 NULL, struct aio_fork_config,
 905                                 return -1);
 906
 907         /*********************************************************************
 908          * How many threads to initialize ?
 909          * 100 per process seems insane as a default until you realize that
 910          * (a) Threads terminate after 1 second when idle.
 911          * (b) Throttling is done in SMB2 via the crediting algorithm.
 912          * (c) SMB1 clients are limited to max_mux (50) outstanding
 913          *     requests and Windows clients don't use this anyway.
 914          * Essentially we want this to be unlimited unless smb.conf
 915          * says different.
 916          *********************************************************************/
 917         aio_pending_size = 100;
 918         return 0;
 919 }
 920
 921 static struct vfs_fn_pointers vfs_aio_fork_fns = {
 922         .connect_fn = aio_fork_connect,
 923         .pread_send_fn = aio_fork_pread_send,
 924         .pread_recv_fn = aio_fork_pread_recv,
 925         .pwrite_send_fn = aio_fork_pwrite_send,
 926         .pwrite_recv_fn = aio_fork_pwrite_recv,
 927         .fsync_send_fn = aio_fork_fsync_send,
 928         .fsync_recv_fn = aio_fork_fsync_recv,
 929 };
 930
 931 NTSTATUS vfs_aio_fork_init(void);
 932 NTSTATUS vfs_aio_fork_init(void)
 933 {
 934         return smb_register_vfs(SMB_VFS_INTERFACE_VERSION,
 935                                 "aio_fork", &vfs_aio_fork_fns);
 936 }