summaryrefslogtreecommitdiffstats
path: root/xlators/performance/io-cache/src
diff options
context:
space:
mode:
Diffstat (limited to 'xlators/performance/io-cache/src')
-rw-r--r--xlators/performance/io-cache/src/io-cache.c2
-rw-r--r--xlators/performance/io-cache/src/io-cache.h2
-rw-r--r--xlators/performance/io-cache/src/ioc-inode.c2
-rw-r--r--xlators/performance/io-cache/src/page.c4
4 files changed, 5 insertions, 5 deletions
diff --git a/xlators/performance/io-cache/src/io-cache.c b/xlators/performance/io-cache/src/io-cache.c
index 5e93d164aba..bc7451c2654 100644
--- a/xlators/performance/io-cache/src/io-cache.c
+++ b/xlators/performance/io-cache/src/io-cache.c
@@ -1,5 +1,5 @@
/*
- Copyright (c) 2007-2009 Z RESEARCH, Inc. <http://www.zresearch.com>
+ Copyright (c) 2007-2009 Gluster, Inc. <http://www.gluster.com>
This file is part of GlusterFS.
GlusterFS is free software; you can redistribute it and/or modify
diff --git a/xlators/performance/io-cache/src/io-cache.h b/xlators/performance/io-cache/src/io-cache.h
index 5d0590d3311..8f628d207bd 100644
--- a/xlators/performance/io-cache/src/io-cache.h
+++ b/xlators/performance/io-cache/src/io-cache.h
@@ -1,5 +1,5 @@
/*
- Copyright (c) 2007-2009 Z RESEARCH, Inc. <http://www.zresearch.com>
+ Copyright (c) 2007-2009 Gluster, Inc. <http://www.gluster.com>
This file is part of GlusterFS.
GlusterFS is free software; you can redistribute it and/or modify
diff --git a/xlators/performance/io-cache/src/ioc-inode.c b/xlators/performance/io-cache/src/ioc-inode.c
index 60207720278..380558e06aa 100644
--- a/xlators/performance/io-cache/src/ioc-inode.c
+++ b/xlators/performance/io-cache/src/ioc-inode.c
@@ -1,5 +1,5 @@
/*
- Copyright (c) 2007-2009 Z RESEARCH, Inc. <http://www.zresearch.com>
+ Copyright (c) 2007-2009 Gluster, Inc. <http://www.gluster.com>
This file is part of GlusterFS.
GlusterFS is free software; you can redistribute it and/or modify
diff --git a/xlators/performance/io-cache/src/page.c b/xlators/performance/io-cache/src/page.c
index be880344441..9db42bca1aa 100644
--- a/xlators/performance/io-cache/src/page.c
+++ b/xlators/performance/io-cache/src/page.c
@@ -1,5 +1,5 @@
/*
- Copyright (c) 2007-2009 Z RESEARCH, Inc. <http://www.zresearch.com>
+ Copyright (c) 2007-2009 Gluster, Inc. <http://www.gluster.com>
This file is part of GlusterFS.
GlusterFS is free software; you can redistribute it and/or modify
@@ -290,7 +290,7 @@ ioc_cache_still_valid (ioc_inode_t *ioc_inode, struct stat *stbuf)
#endif
#if 0
- /* talk with avati@zresearch.com to enable this section */
+ /* talk with avati@gluster.com to enable this section */
if (!ioc_inode->mtime && stbuf) {
cache_still_valid = 1;
ioc_inode->mtime = stbuf->st_mtime;
'>250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002
/*
  Copyright (c) 2012-2014 DataLab, s.l. <http://www.datalab.es>
  This file is part of GlusterFS.

  This file is licensed to you under your choice of the GNU Lesser
  General Public License, version 3 or any later version (LGPLv3 or
  later), or the GNU General Public License, version 2 (GPLv2), in all
  cases as published by the Free Software Foundation.
*/

#include <fnmatch.h>

#include "libxlator.h"
#include "byte-order.h"

#include "ec-types.h"
#include "ec-helpers.h"
#include "ec-common.h"
#include "ec-combine.h"
#include "ec-messages.h"
#include "quota-common-utils.h"

#define EC_QUOTA_PREFIX "trusted.glusterfs.quota."

#define EC_MISSING_DATA ((data_t *)1ULL)

struct _ec_dict_info;
typedef struct _ec_dict_info ec_dict_info_t;

struct _ec_dict_combine;
typedef struct _ec_dict_combine ec_dict_combine_t;

struct _ec_dict_info {
    dict_t *dict;
    int32_t count;
};

struct _ec_dict_combine {
    ec_cbk_data_t *cbk;
    int32_t which;
};

int32_t
ec_combine_write(ec_fop_data_t *fop, ec_cbk_data_t *dst, ec_cbk_data_t *src)
{
    int valid = 0;

    if (!fop || !dst || !src)
        return 0;

    switch (fop->id) {
        case GF_FOP_REMOVEXATTR:
        case GF_FOP_FREMOVEXATTR:
        case GF_FOP_SETXATTR:
        case GF_FOP_FSETXATTR:
            return 1;

        case GF_FOP_SYMLINK:
        case GF_FOP_LINK:
        case GF_FOP_CREATE:
        case GF_FOP_MKNOD:
        case GF_FOP_MKDIR:
            valid = 3;
            break;
        case GF_FOP_UNLINK:
        case GF_FOP_RMDIR:
        case GF_FOP_SETATTR:
        case GF_FOP_FSETATTR:
        case GF_FOP_TRUNCATE:
        case GF_FOP_FTRUNCATE:
        case GF_FOP_WRITE:
        case GF_FOP_FALLOCATE:
        case GF_FOP_DISCARD:
        case GF_FOP_ZEROFILL:
            valid = 2;
            break;
        case GF_FOP_RENAME:
            valid = 5;
            break;
        default:
            gf_msg_callingfn(fop->xl->name, GF_LOG_WARNING, EINVAL,
                             EC_MSG_INVALID_FOP, "Invalid fop %d", fop->id);
            return 0;
            break;
    }

    if (!ec_iatt_combine(fop, dst->iatt, src->iatt, valid)) {
        gf_msg(fop->xl->name, GF_LOG_NOTICE, 0, EC_MSG_IATT_MISMATCH,
               "Mismatching iatt in "
               "answers of '%s'",
               gf_fop_list[fop->id]);
        return 0;
    }
    return 1;
}

void
ec_iatt_time_merge(int64_t *dst_sec, uint32_t *dst_nsec, int64_t src_sec,
                   uint32_t src_nsec)
{
    if ((*dst_sec < src_sec) ||
        ((*dst_sec == src_sec) && (*dst_nsec < src_nsec))) {
        *dst_sec = src_sec;
        *dst_nsec = src_nsec;
    }
}

static gf_boolean_t
ec_iatt_is_trusted(ec_fop_data_t *fop, struct iatt *iatt)
{
    uint64_t ino;
    int32_t i;

    /* Only the top level fop will have fop->locks filled. */
    while (fop->parent != NULL) {
        fop = fop->parent;
    }

    /* Lookups are special requests always done without locks taken but they
     * require to be able to identify differences between bricks. Special
     * handling of these differences is already done in lookup specific code
     * so we shouldn't ignore any difference here and consider all iatt
     * structures as trusted. */
    if (fop->id == GF_FOP_LOOKUP) {
        return _gf_true;
    }

    /* Check if the iatt references an inode locked by the current fop */
    for (i = 0; i < fop->lock_count; i++) {
        ino = gfid_to_ino(fop->locks[i].lock->loc.inode->gfid);
        if (iatt->ia_ino == ino) {
            return _gf_true;
        }
    }

    return _gf_false;
}

int32_t
ec_iatt_combine(ec_fop_data_t *fop, struct iatt *dst, struct iatt *src,
                int32_t count)
{
    int32_t i;
    gf_boolean_t failed = _gf_false;

    for (i = 0; i < count; i++) {
        /* Check for basic fields. These fields must be equal always, even if
         * the inode is not locked because in these cases the parent inode
         * will be locked and differences in these fields require changes in
         * the parent directory. */
        if ((dst[i].ia_ino != src[i].ia_ino) ||
            (((dst[i].ia_type == IA_IFBLK) || (dst[i].ia_type == IA_IFCHR)) &&
             (dst[i].ia_rdev != src[i].ia_rdev)) ||
            (gf_uuid_compare(dst[i].ia_gfid, src[i].ia_gfid) != 0)) {
            failed = _gf_true;
        }
        /* Check for not so stable fields. These fields can change if the
         * inode is not locked. */
        if (!failed && ((dst[i].ia_uid != src[i].ia_uid) ||
                        (dst[i].ia_gid != src[i].ia_gid) ||
                        (st_mode_from_ia(dst[i].ia_prot, dst[i].ia_type) !=
                         st_mode_from_ia(src[i].ia_prot, src[i].ia_type)))) {
            if (ec_iatt_is_trusted(fop, dst)) {
                /* If the iatt contains information from an inode that is
                 * locked, these differences are real problems, so we need to
                 * report them. Otherwise we ignore them and don't care which
                 * data is returned. */
                failed = _gf_true;
            } else {
                gf_msg_debug(fop->xl->name, 0,
                             "Ignoring iatt differences because inode is not "
                             "locked");
            }
        }
        if (failed) {
            gf_msg(fop->xl->name, GF_LOG_WARNING, 0, EC_MSG_IATT_COMBINE_FAIL,
                   "Failed to combine iatt (inode: %lu-%lu, links: %u-%u, "
                   "uid: %u-%u, gid: %u-%u, rdev: %lu-%lu, size: %lu-%lu, "
                   "mode: %o-%o)",
                   dst[i].ia_ino, src[i].ia_ino, dst[i].ia_nlink,
                   src[i].ia_nlink, dst[i].ia_uid, src[i].ia_uid, dst[i].ia_gid,
                   src[i].ia_gid, dst[i].ia_rdev, src[i].ia_rdev,
                   dst[i].ia_size, src[i].ia_size,
                   st_mode_from_ia(dst[i].ia_prot, dst[i].ia_type),
                   st_mode_from_ia(src[i].ia_prot, dst[i].ia_type));

            return 0;
        }
    }

    while (count-- > 0) {
        dst[count].ia_blocks += src[count].ia_blocks;
        if (dst[count].ia_blksize < src[count].ia_blksize) {
            dst[count].ia_blksize = src[count].ia_blksize;
        }

        ec_iatt_time_merge(&dst[count].ia_atime, &dst[count].ia_atime_nsec,
                           src[count].ia_atime, src[count].ia_atime_nsec);
        ec_iatt_time_merge(&dst[count].ia_mtime, &dst[count].ia_mtime_nsec,
                           src[count].ia_mtime, src[count].ia_mtime_nsec);
        ec_iatt_time_merge(&dst[count].ia_ctime, &dst[count].ia_ctime_nsec,
                           src[count].ia_ctime, src[count].ia_ctime_nsec);
    }

    return 1;
}

void
ec_iatt_rebuild(ec_t *ec, struct iatt *iatt, int32_t count, int32_t answers)
{
    uint64_t blocks;

    while (count-- > 0) {
        blocks = iatt[count].ia_blocks * ec->fragments + answers - 1;
        blocks /= answers;
        iatt[count].ia_blocks = blocks;
    }
}

gf_boolean_t
ec_xattr_match(dict_t *dict, char *key, data_t *value, void *arg)
{
    if ((fnmatch(GF_XATTR_STIME_PATTERN, key, 0) == 0) ||
        (strcmp(key, GET_LINK_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_INODELK_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_ENTRYLK_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_OPEN_FD_COUNT) == 0)) {
        return _gf_false;
    }

    return _gf_true;
}

gf_boolean_t
ec_value_ignore(char *key)
{
    if ((strcmp(key, GF_CONTENT_KEY) == 0) ||
        (strcmp(key, GF_XATTR_PATHINFO_KEY) == 0) ||
        (strcmp(key, GF_XATTR_USER_PATHINFO_KEY) == 0) ||
        (strcmp(key, GF_XATTR_LOCKINFO_KEY) == 0) ||
        (strcmp(key, GLUSTERFS_OPEN_FD_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_INODELK_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_ENTRYLK_COUNT) == 0) ||
        (strncmp(key, GF_XATTR_CLRLK_CMD, SLEN(GF_XATTR_CLRLK_CMD)) == 0) ||
        (strcmp(key, DHT_IATT_IN_XDATA_KEY) == 0) ||
        (strncmp(key, EC_QUOTA_PREFIX, SLEN(EC_QUOTA_PREFIX)) == 0) ||
        (fnmatch(MARKER_XATTR_PREFIX ".*." XTIME, key, 0) == 0) ||
        (fnmatch(GF_XATTR_MARKER_KEY ".*", key, 0) == 0) ||
        (XATTR_IS_NODE_UUID(key))) {
        return _gf_true;
    }

    return _gf_false;
}

int32_t
ec_dict_compare(dict_t *dict1, dict_t *dict2)
{
    if (are_dicts_equal(dict1, dict2, ec_xattr_match, ec_value_ignore))
        return 1;
    return 0;
}

static uint32_t
ec_dict_list(data_t **list, ec_cbk_data_t *cbk, int32_t which, char *key,
             gf_boolean_t global)
{
    ec_t *ec = cbk->fop->xl->private;
    ec_cbk_data_t *ans = NULL;
    dict_t *dict = NULL;
    data_t *data;
    uint32_t count;
    int32_t i;

    for (i = 0; i < ec->nodes; i++) {
        /* We initialize the list with EC_MISSING_DATA if we are
         * returning a global list or the current subvolume belongs
         * to the group of the accepted answer. Note that if some
         * subvolume is known to be down before issuing the request,
         * we won't have any answer from it, so we set here the
         * appropriate default value. */
        if (global || ((cbk->mask & (1ULL << i)) != 0)) {
            list[i] = EC_MISSING_DATA;
        } else {
            list[i] = NULL;
        }
    }

    count = 0;
    list_for_each_entry(ans, &cbk->fop->answer_list, answer_list)
    {
        if (global || ((cbk->mask & ans->mask) != 0)) {
            dict = (which == EC_COMBINE_XDATA) ? ans->xdata : ans->dict;
            data = dict_get(dict, key);
            if (data != NULL) {
                list[ans->idx] = data;
                count++;
            }
        }
    }

    return count;
}

int32_t
ec_concat_prepare(xlator_t *xl, char **str, char **sep, char **post,
                  const char *fmt, va_list args)
{
    char *tmp;
    int32_t len;

    len = gf_vasprintf(str, fmt, args);
    if (len < 0) {
        return -ENOMEM;
    }

    tmp = strchr(*str, '{');
    if (tmp == NULL) {
        goto out;
    }
    *tmp++ = 0;
    *sep = tmp;
    tmp = strchr(tmp, '}');
    if (tmp == NULL) {
        goto out;
    }
    *tmp++ = 0;
    *post = tmp;

    return 0;

out:
    gf_msg(xl->name, GF_LOG_ERROR, EINVAL, EC_MSG_INVALID_FORMAT,
           "Invalid concat format");

    GF_FREE(*str);

    return -EINVAL;
}

static int32_t
ec_dict_data_concat(const char *fmt, ec_cbk_data_t *cbk, int32_t which,
                    char *key, char *new_key, const char *def,
                    gf_boolean_t global, ...)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    char *str = NULL, *pre = NULL, *sep, *post;
    dict_t *dict;
    va_list args;
    int32_t i, num, len, deflen, prelen, postlen, seplen, tmp;
    int32_t err;

    ec_dict_list(data, cbk, which, key, global);

    va_start(args, global);
    err = ec_concat_prepare(cbk->fop->xl, &pre, &sep, &post, fmt, args);
    va_end(args);

    if (err != 0) {
        return err;
    }

    prelen = strlen(pre);
    seplen = strlen(sep);
    postlen = strlen(post);

    deflen = 0;
    if (def != NULL) {
        deflen = strlen(def);
    }

    len = prelen + postlen + 1;
    num = -1;
    for (i = 0; i < ec->nodes; i++) {
        if (data[i] == NULL) {
            continue;
        }
        if (data[i] == EC_MISSING_DATA) {
            if (def == NULL) {
                continue;
            }
            len += deflen;
        } else {
            len += data[i]->len - 1;
        }
        if (num >= 0) {
            len += seplen;
        }
        num++;
    }

    err = -ENOMEM;

    str = GF_MALLOC(len, gf_common_mt_char);
    if (str == NULL) {
        goto out;
    }

    memcpy(str, pre, prelen);
    len = prelen;
    for (i = 0; i < ec->nodes; i++) {
        if (data[i] == NULL) {
            continue;
        }
        if (data[i] == EC_MISSING_DATA) {
            if (deflen == 0) {
                continue;
            }
            tmp = deflen;
            memcpy(str + len, def, tmp);
        } else {
            tmp = data[i]->len - 1;
            memcpy(str + len, data[i]->data, tmp);
        }
        len += tmp;
        if (i < num) {
            memcpy(str + len, sep, seplen);
            len += seplen;
        }
    }
    memcpy(str + len, post, postlen + 1);

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    if (new_key) {
        key = new_key;
    }
    err = dict_set_dynstr(dict, key, str);
    if (err != 0) {
        goto out;
    }

    str = NULL;

out:
    GF_FREE(str);
    GF_FREE(pre);

    return err;
}

int32_t
ec_dict_data_merge(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict, *lockinfo, *tmp = NULL;
    char *ptr = NULL;
    int32_t i, len;
    int32_t err;

    ec_dict_list(data, cbk, which, key, _gf_false);

    lockinfo = dict_new();
    if (lockinfo == NULL) {
        return -ENOMEM;
    }

    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA)) {
            continue;
        }

        tmp = dict_new();
        if (tmp == NULL) {
            err = -ENOMEM;

            goto out;
        }
        err = dict_unserialize(data[i]->data, data[i]->len, &tmp);
        if (err != 0) {
            goto out;
        }
        if (dict_copy(tmp, lockinfo) == NULL) {
            err = -ENOMEM;

            goto out;
        }

        dict_unref(tmp);
    }

    tmp = NULL;

    len = dict_serialized_length(lockinfo);
    if (len < 0) {
        err = len;

        goto out;
    }
    ptr = GF_MALLOC(len, gf_common_mt_char);
    if (ptr == NULL) {
        err = -ENOMEM;

        goto out;
    }
    err = dict_serialize(lockinfo, ptr);
    if (err != 0) {
        goto out;
    }
    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    err = dict_set_dynptr(dict, key, ptr, len);
    if (err != 0) {
        goto out;
    }

    ptr = NULL;

out:
    GF_FREE(ptr);
    dict_unref(lockinfo);
    if (tmp != NULL) {
        dict_unref(tmp);
    }

    return err;
}

int32_t
ec_dict_data_uuid(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_cbk_data_t *ans, *min;
    dict_t *src, *dst;
    data_t *data;

    min = cbk;
    for (ans = cbk->next; ans != NULL; ans = ans->next) {
        if (ans->idx < min->idx) {
            min = ans;
        }
    }

    if (min != cbk) {
        src = (which == EC_COMBINE_XDATA) ? min->xdata : min->dict;
        dst = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;

        data = dict_get(src, key);
        if (data == NULL) {
            return -ENOENT;
        }
        if (dict_set(dst, key, data) != 0) {
            return -ENOMEM;
        }
    }

    return 0;
}

int32_t
ec_dict_data_iatt(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict;
    struct iatt *stbuf, *tmp;
    int32_t i, ret;

    ec_dict_list(data, cbk, which, key, _gf_false);

    stbuf = NULL;
    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA)) {
            continue;
        }
        tmp = data_to_iatt(data[i], key);
        if (tmp == NULL) {
            ret = -EINVAL;
            goto out;
        }
        if (stbuf == NULL) {
            stbuf = GF_MALLOC(sizeof(struct iatt), gf_common_mt_char);
            if (stbuf == NULL) {
                ret = -ENOMEM;
                goto out;
            }
            *stbuf = *tmp;
        } else {
            if (!ec_iatt_combine(cbk->fop, stbuf, tmp, 1)) {
                ret = -EINVAL;
                goto out;
            }
        }
    }

    if ((stbuf != NULL) && (stbuf->ia_type == IA_IFREG)) {
        ec_iatt_rebuild(ec, stbuf, 1, cbk->count);
        /* TODO: not sure if an iatt could come in xdata from a fop that takes
         *       no locks. */
        if (!ec_get_inode_size(cbk->fop, cbk->fop->locks[0].lock->loc.inode,
                               &stbuf->ia_size)) {
            ret = -EINVAL;
            goto out;
        }
    }

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    ret = dict_set_iatt(dict, key, stbuf, false);
    if (ret >= 0) {
        stbuf = NULL;
    }

out:
    GF_FREE(stbuf);

    return ret;
}

int32_t
ec_dict_data_max32(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict;
    int32_t i;
    uint32_t max, tmp;

    ec_dict_list(data, cbk, which, key, _gf_false);

    max = 0;
    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA)) {
            continue;
        }

        tmp = data_to_uint32(data[i]);
        if (max < tmp) {
            max = tmp;
        }
    }

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    return dict_set_uint32(dict, key, max);
}

int32_t
ec_dict_data_max64(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict;
    int32_t i;
    uint64_t max, tmp;

    ec_dict_list(data, cbk, which, key, _gf_false);

    max = 0;
    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA)) {
            continue;
        }

        tmp = data_to_uint64(data[i]);
        if (max < tmp) {
            max = tmp;
        }
    }

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    return dict_set_uint64(dict, key, max);
}

int32_t
ec_dict_data_quota(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict = NULL;
    int32_t i = 0;
    quota_meta_t size = {
        0,
    };
    quota_meta_t max_size = {
        0,
    };

    if (ec_dict_list(data, cbk, which, key, _gf_false) == 0) {
        return 0;
    }

    /* Quota size xattr is managed outside of the control of the ec xlator.
     * This means that it might not be updated at the same time on all
     * bricks and we can receive slightly different values. If that's the
     * case, we take the maximum of all received values.
     */
    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA) ||
            (quota_data_to_meta(data[i], QUOTA_SIZE_KEY, &size) < 0)) {
            continue;
        }

        if (size.size > max_size.size)
            max_size.size = size.size;
        if (size.file_count > max_size.file_count)
            max_size.file_count = size.file_count;
        if (size.dir_count > max_size.dir_count)
            max_size.dir_count = size.dir_count;
    }

    max_size.size *= ec->fragments;

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    return quota_dict_set_meta(dict, key, &max_size, IA_IFDIR);
}

int32_t
ec_dict_data_stime(ec_cbk_data_t *cbk, int32_t which, char *key)
{
    ec_t *ec = cbk->fop->xl->private;
    data_t *data[ec->nodes];
    dict_t *dict;
    int32_t i, err;

    ec_dict_list(data, cbk, which, key, _gf_false);

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    for (i = 0; i < ec->nodes; i++) {
        if ((data[i] == NULL) || (data[i] == EC_MISSING_DATA)) {
            continue;
        }
        err = gf_get_max_stime(cbk->fop->xl, dict, key, data[i]);
        if (err != 0) {
            gf_msg(cbk->fop->xl->name, GF_LOG_ERROR, -err,
                   EC_MSG_STIME_COMBINE_FAIL, "STIME combination failed");

            return err;
        }
    }

    return 0;
}

int32_t
ec_dict_data_combine(dict_t *dict, char *key, data_t *value, void *arg)
{
    ec_dict_combine_t *data = arg;

    if ((strcmp(key, GF_XATTR_PATHINFO_KEY) == 0) ||
        (strcmp(key, GF_XATTR_USER_PATHINFO_KEY) == 0)) {
        return ec_dict_data_concat("(<EC:%s> { })", data->cbk, data->which, key,
                                   NULL, NULL, _gf_false,
                                   data->cbk->fop->xl->name);
    }

    if (strncmp(key, GF_XATTR_CLRLK_CMD, SLEN(GF_XATTR_CLRLK_CMD)) == 0) {
        return ec_dict_data_concat("{\n}", data->cbk, data->which, key, NULL,
                                   NULL, _gf_false);
    }

    if (strncmp(key, GF_XATTR_LOCKINFO_KEY, SLEN(GF_XATTR_LOCKINFO_KEY)) == 0) {
        return ec_dict_data_merge(data->cbk, data->which, key);
    }

    if (strcmp(key, GET_LINK_COUNT) == 0) {
        return ec_dict_data_max32(data->cbk, data->which, key);
    }

    if (strcmp(key, GLUSTERFS_OPEN_FD_COUNT) == 0) {
        return ec_dict_data_max32(data->cbk, data->which, key);
    }
    if ((strcmp(key, GLUSTERFS_INODELK_COUNT) == 0) ||
        (strcmp(key, GLUSTERFS_ENTRYLK_COUNT) == 0)) {
        return ec_dict_data_max32(data->cbk, data->which, key);
    }

    if (strcmp(key, QUOTA_SIZE_KEY) == 0) {
        return ec_dict_data_quota(data->cbk, data->which, key);
    }
    /* Ignore all other quota attributes */
    if (strncmp(key, EC_QUOTA_PREFIX, SLEN(EC_QUOTA_PREFIX)) == 0) {
        return 0;
    }

    if (XATTR_IS_NODE_UUID(key)) {
        if (data->cbk->fop->int32) {
            /* List of node uuid is requested */
            return ec_dict_data_concat("{ }", data->cbk, data->which, key,
                                       GF_XATTR_LIST_NODE_UUIDS_KEY, UUID0_STR,
                                       _gf_true);
        } else {
            return ec_dict_data_uuid(data->cbk, data->which, key);
        }
    }

    if (fnmatch(GF_XATTR_STIME_PATTERN, key, FNM_NOESCAPE) == 0) {
        return ec_dict_data_stime(data->cbk, data->which, key);
    }

    if (fnmatch(MARKER_XATTR_PREFIX ".*." XTIME, key, FNM_NOESCAPE) == 0) {
        return ec_dict_data_max64(data->cbk, data->which, key);
    }

    if (strcmp(key, GF_PRESTAT) == 0 || strcmp(key, GF_POSTSTAT) == 0) {
        return ec_dict_data_iatt(data->cbk, data->which, key);
    }

    return 0;
}

int32_t
ec_dict_combine(ec_cbk_data_t *cbk, int32_t which)
{
    dict_t *dict = NULL;
    ec_dict_combine_t data;
    int32_t err = 0;

    data.cbk = cbk;
    data.which = which;

    dict = (which == EC_COMBINE_XDATA) ? cbk->xdata : cbk->dict;
    if (dict != NULL) {
        err = dict_foreach(dict, ec_dict_data_combine, &data);
        if (err != 0) {
            gf_msg(cbk->fop->xl->name, GF_LOG_ERROR, -err,
                   EC_MSG_DICT_COMBINE_FAIL, "Dictionary combination failed");

            return err;
        }
    }

    return 0;
}

int32_t
ec_vector_compare(struct iovec *dst_vector, int32_t dst_count,
                  struct iovec *src_vector, int32_t src_count)
{
    int32_t dst_size = 0, src_size = 0;

    if (dst_count > 0) {
        dst_size = iov_length(dst_vector, dst_count);
    }
    if (src_count > 0) {
        src_size = iov_length(src_vector, src_count);
    }

    return (dst_size == src_size);
}

int32_t
ec_flock_compare(struct gf_flock *dst, struct gf_flock *src)
{
    if ((dst->l_type != src->l_type) || (dst->l_whence != src->l_whence) ||
        (dst->l_start != src->l_start) || (dst->l_len != src->l_len) ||
        (dst->l_pid != src->l_pid) ||
        !is_same_lkowner(&dst->l_owner, &src->l_owner)) {
        return 0;
    }

    return 1;
}

void
ec_statvfs_combine(struct statvfs *dst, struct statvfs *src)
{
    if (dst->f_bsize < src->f_bsize) {
        dst->f_bsize = src->f_bsize;
    }

    if (dst->f_frsize < src->f_frsize) {
        dst->f_blocks *= dst->f_frsize;
        dst->f_blocks /= src->f_frsize;

        dst->f_bfree *= dst->f_frsize;
        dst->f_bfree /= src->f_frsize;

        dst->f_bavail *= dst->f_frsize;
        dst->f_bavail /= src->f_frsize;

        dst->f_frsize = src->f_frsize;
    } else if (dst->f_frsize > src->f_frsize) {
        src->f_blocks *= src->f_frsize;
        src->f_blocks /= dst->f_frsize;

        src->f_bfree *= src->f_frsize;
        src->f_bfree /= dst->f_frsize;

        src->f_bavail *= src->f_frsize;
        src->f_bavail /= dst->f_frsize;
    }
    if (dst->f_blocks > src->f_blocks) {
        dst->f_blocks = src->f_blocks;
    }
    if (dst->f_bfree > src->f_bfree) {
        dst->f_bfree = src->f_bfree;
    }
    if (dst->f_bavail > src->f_bavail) {
        dst->f_bavail = src->f_bavail;
    }

    if (dst->f_files < src->f_files) {
        dst->f_files = src->f_files;
    }
    if (dst->f_ffree > src->f_ffree) {
        dst->f_ffree = src->f_ffree;
    }
    if (dst->f_favail > src->f_favail) {
        dst->f_favail = src->f_favail;
    }
    if (dst->f_namemax > src->f_namemax) {
        dst->f_namemax = src->f_namemax;
    }

    if (dst->f_flag != src->f_flag) {
        gf_msg_debug(THIS->name, 0,
                     "Mismatching file system flags "
                     "(%lX, %lX)",
                     dst->f_flag, src->f_flag);
    }
    dst->f_flag &= src->f_flag;
}

int32_t
ec_combine_check(ec_cbk_data_t *dst, ec_cbk_data_t *src, ec_combine_f combine)
{
    ec_fop_data_t *fop = dst->fop;

    if (dst->op_ret != src->op_ret) {
        gf_msg_debug(fop->xl->name, 0,
                     "Mismatching return code in "
                     "answers of '%s': %d <-> %d",
                     ec_fop_name(fop->id), dst->op_ret, src->op_ret);

        return 0;
    }
    if (dst->op_ret < 0) {
        if (dst->op_errno != src->op_errno) {
            gf_msg_debug(fop->xl->name, 0,
                         "Mismatching errno code in "
                         "answers of '%s': %d <-> %d",
                         ec_fop_name(fop->id), dst->op_errno, src->op_errno);

            return 0;
        }
    }

    if (!ec_dict_compare(dst->xdata, src->xdata)) {
        gf_msg(fop->xl->name, GF_LOG_DEBUG, 0, EC_MSG_XDATA_MISMATCH,
               "Mismatching xdata in answers "
               "of '%s'",
               ec_fop_name(fop->id));

        return 0;
    }

    if ((dst->op_ret >= 0) && (combine != NULL)) {
        return combine(fop, dst, src);
    }

    return 1;
}

void
ec_combine(ec_cbk_data_t *newcbk, ec_combine_f combine)
{
    ec_fop_data_t *fop = newcbk->fop;
    ec_cbk_data_t *cbk = NULL, *tmp = NULL;
    struct list_head *item = NULL;
    int32_t needed = 0;
    char str[32];

    LOCK(&fop->lock);

    fop->received |= newcbk->mask;

    item = fop->cbk_list.prev;
    list_for_each_entry(cbk, &fop->cbk_list, list)
    {
        if (ec_combine_check(newcbk, cbk, combine)) {
            newcbk->count += cbk->count;
            newcbk->mask |= cbk->mask;

            item = cbk->list.prev;
            while (item != &fop->cbk_list) {
                tmp = list_entry(item, ec_cbk_data_t, list);
                if (tmp->count >= newcbk->count) {
                    break;
                }
                item = item->prev;
            }
            list_del(&cbk->list);

            newcbk->next = cbk;

            break;
        }
    }
    list_add(&newcbk->list, item);

    ec_trace("ANSWER", fop, "combine=%s[%d]",
             ec_bin(str, sizeof(str), newcbk->mask, 0), newcbk->count);

    cbk = list_entry(fop->cbk_list.next, ec_cbk_data_t, list);
    if ((fop->mask ^ fop->remaining) == fop->received) {
        needed = fop->minimum - cbk->count;
    }

    UNLOCK(&fop->lock);

    if (needed > 0) {
        ec_dispatch_next(fop, newcbk->idx);
    }
}