diff options
author | Brian Behlendorf <[email protected]> | 2010-05-28 13:45:14 -0700 |
---|---|---|
committer | Brian Behlendorf <[email protected]> | 2010-05-28 13:45:14 -0700 |
commit | 428870ff734fdaccc342b33fc53cf94724409a46 (patch) | |
tree | 164e83c0ceda52a843795ed7cd9e95637d02c177 /cmd/zdb | |
parent | 6119cb885a976e175a6e827894accf657ff1984f (diff) |
Update core ZFS code from build 121 to build 141.
Diffstat (limited to 'cmd/zdb')
-rw-r--r-- | cmd/zdb/zdb.c | 2243 | ||||
-rw-r--r-- | cmd/zdb/zdb_il.c | 141 |
2 files changed, 1521 insertions, 863 deletions
diff --git a/cmd/zdb/zdb.c b/cmd/zdb/zdb.c index 292bb519a..ff73072f8 100644 --- a/cmd/zdb/zdb.c +++ b/cmd/zdb/zdb.c @@ -19,8 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2009 Sun Microsystems, Inc. All rights reserved. - * Use is subject to license terms. + * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. */ #include <stdio.h> @@ -34,6 +33,9 @@ #include <sys/zap.h> #include <sys/fs/zfs.h> #include <sys/zfs_znode.h> +#include <sys/zfs_sa.h> +#include <sys/sa.h> +#include <sys/sa_impl.h> #include <sys/vdev.h> #include <sys/vdev_impl.h> #include <sys/metaslab_impl.h> @@ -51,10 +53,25 @@ #include <sys/zio_compress.h> #include <sys/zfs_fuid.h> #include <sys/arc.h> +#include <sys/ddt.h> #undef ZFS_MAXNAMELEN #undef verify #include <libzfs.h> +#define ZDB_COMPRESS_NAME(idx) ((idx) < ZIO_COMPRESS_FUNCTIONS ? \ + zio_compress_table[(idx)].ci_name : "UNKNOWN") +#define ZDB_CHECKSUM_NAME(idx) ((idx) < ZIO_CHECKSUM_FUNCTIONS ? \ + zio_checksum_table[(idx)].ci_name : "UNKNOWN") +#define ZDB_OT_NAME(idx) ((idx) < DMU_OT_NUMTYPES ? \ + dmu_ot[(idx)].ot_name : "UNKNOWN") +#define ZDB_OT_TYPE(idx) ((idx) < DMU_OT_NUMTYPES ? (idx) : DMU_OT_NUMTYPES) + +#ifndef lint +extern int zfs_recover; +#else +int zfs_recover; +#endif + const char cmdname[] = "zdb"; uint8_t dump_opt[256]; @@ -64,8 +81,6 @@ extern void dump_intent_log(zilog_t *); uint64_t *zopt_object = NULL; int zopt_objects = 0; libzfs_handle_t *g_zfs; -boolean_t zdb_sig_user_data = B_TRUE; -int zdb_sig_cksumalg = ZIO_CHECKSUM_SHA256; /* * These libumem hooks provide a reasonable set of defaults for the allocator's @@ -87,39 +102,56 @@ static void usage(void) { (void) fprintf(stderr, - "Usage: %s [-udibcsvL] [-U cachefile_path] [-t txg]\n" - "\t [-S user:cksumalg] " - "dataset [object...]\n" - " %s -C [pool]\n" - " %s -l dev\n" - " %s -R pool:vdev:offset:size:flags\n" - " %s [-p path_to_vdev_dir]\n" - " %s -e pool | GUID | devid ...\n", - cmdname, cmdname, cmdname, cmdname, cmdname, cmdname); - - (void) fprintf(stderr, " -u uberblock\n"); - (void) fprintf(stderr, " -d datasets\n"); - (void) fprintf(stderr, " -C cached pool configuration\n"); - (void) fprintf(stderr, " -i intent logs\n"); - (void) fprintf(stderr, " -b block statistics\n"); - (void) fprintf(stderr, " -m metaslabs\n"); - (void) fprintf(stderr, " -c checksum all metadata (twice for " + "Usage: %s [-CumdibcsDvhL] poolname [object...]\n" + " %s [-div] dataset [object...]\n" + " %s -m [-L] poolname [vdev [metaslab...]]\n" + " %s -R poolname vdev:offset:size[:flags]\n" + " %s -S poolname\n" + " %s -l [-u] device\n" + " %s -C\n\n", + cmdname, cmdname, cmdname, cmdname, cmdname, cmdname, cmdname); + + (void) fprintf(stderr, " Dataset name must include at least one " + "separator character '/' or '@'\n"); + (void) fprintf(stderr, " If dataset name is specified, only that " + "dataset is dumped\n"); + (void) fprintf(stderr, " If object numbers are specified, only " + "those objects are dumped\n\n"); + (void) fprintf(stderr, " Options to control amount of output:\n"); + (void) fprintf(stderr, " -u uberblock\n"); + (void) fprintf(stderr, " -d dataset(s)\n"); + (void) fprintf(stderr, " -i intent logs\n"); + (void) fprintf(stderr, " -C config (or cachefile if alone)\n"); + (void) fprintf(stderr, " -h pool history\n"); + (void) fprintf(stderr, " -b block statistics\n"); + (void) fprintf(stderr, " -m metaslabs\n"); + (void) fprintf(stderr, " -c checksum all metadata (twice for " "all data) blocks\n"); - (void) fprintf(stderr, " -s report stats on zdb's I/O\n"); - (void) fprintf(stderr, " -S <user|all>:<cksum_alg|all> -- " - "dump blkptr signatures\n"); - (void) fprintf(stderr, " -v verbose (applies to all others)\n"); + (void) fprintf(stderr, " -s report stats on zdb's I/O\n"); + (void) fprintf(stderr, " -D dedup statistics\n"); + (void) fprintf(stderr, " -S simulate dedup to measure effect\n"); + (void) fprintf(stderr, " -v verbose (applies to all others)\n"); (void) fprintf(stderr, " -l dump label contents\n"); (void) fprintf(stderr, " -L disable leak tracking (do not " "load spacemaps)\n"); - (void) fprintf(stderr, " -U cachefile_path -- use alternate " - "cachefile\n"); (void) fprintf(stderr, " -R read and display block from a " - "device\n"); - (void) fprintf(stderr, " -e Pool is exported/destroyed/" - "has altroot\n"); - (void) fprintf(stderr, " -p <Path to vdev dir> (use with -e)\n"); - (void) fprintf(stderr, " -t <txg> highest txg to use when " + "device\n\n"); + (void) fprintf(stderr, " Below options are intended for use " + "with other options (except -l):\n"); + (void) fprintf(stderr, " -A ignore assertions (-A), enable " + "panic recovery (-AA) or both (-AAA)\n"); + (void) fprintf(stderr, " -F attempt automatic rewind within " + "safe range of transaction groups\n"); + (void) fprintf(stderr, " -U <cachefile_path> -- use alternate " + "cachefile\n"); + (void) fprintf(stderr, " -X attempt extreme rewind (does not " + "work with dataset)\n"); + (void) fprintf(stderr, " -e pool is exported/destroyed/" + "has altroot/not in a cachefile\n"); + (void) fprintf(stderr, " -p <path> -- use one or more with " + "-e to specify path to vdev dir\n"); + (void) fprintf(stderr, " -P print numbers parsable\n"); + (void) fprintf(stderr, " -t <txg> -- highest txg to use when " "searching for uberblocks\n"); (void) fprintf(stderr, "Specify an option more than once (e.g. -bb) " "to make only that option verbose\n"); @@ -146,68 +178,6 @@ fatal(const char *fmt, ...) exit(1); } -static void -dump_nvlist(nvlist_t *list, int indent) -{ - nvpair_t *elem = NULL; - - while ((elem = nvlist_next_nvpair(list, elem)) != NULL) { - switch (nvpair_type(elem)) { - case DATA_TYPE_STRING: - { - char *value; - - VERIFY(nvpair_value_string(elem, &value) == 0); - (void) printf("%*s%s='%s'\n", indent, "", - nvpair_name(elem), value); - } - break; - - case DATA_TYPE_UINT64: - { - uint64_t value; - - VERIFY(nvpair_value_uint64(elem, &value) == 0); - (void) printf("%*s%s=%llu\n", indent, "", - nvpair_name(elem), (u_longlong_t)value); - } - break; - - case DATA_TYPE_NVLIST: - { - nvlist_t *value; - - VERIFY(nvpair_value_nvlist(elem, &value) == 0); - (void) printf("%*s%s\n", indent, "", - nvpair_name(elem)); - dump_nvlist(value, indent + 4); - } - break; - - case DATA_TYPE_NVLIST_ARRAY: - { - nvlist_t **value; - uint_t c, count; - - VERIFY(nvpair_value_nvlist_array(elem, &value, - &count) == 0); - - for (c = 0; c < count; c++) { - (void) printf("%*s%s[%u]\n", indent, "", - nvpair_name(elem), c); - dump_nvlist(value[c], indent + 8); - } - } - break; - - default: - - (void) printf("bad config type %d for %s\n", - nvpair_type(elem), nvpair_name(elem)); - } - } -} - /* ARGSUSED */ static void dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size) @@ -227,6 +197,15 @@ dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size) nvlist_free(nv); } +static void +zdb_nicenum(uint64_t num, char *buf) +{ + if (dump_opt['P']) + (void) sprintf(buf, "%llu", (longlong_t)num); + else + nicenum(num, buf); +} + const char dump_zap_stars[] = "****************************************"; const int dump_zap_width = sizeof (dump_zap_stars) - 1; @@ -325,6 +304,13 @@ dump_none(objset_t *os, uint64_t object, void *data, size_t size) } /*ARGSUSED*/ +static void +dump_unknown(objset_t *os, uint64_t object, void *data, size_t size) +{ + (void) printf("\tUNKNOWN OBJECT TYPE\n"); +} + +/*ARGSUSED*/ void dump_uint8(objset_t *os, uint64_t object, void *data, size_t size) { @@ -388,6 +374,79 @@ dump_zap(objset_t *os, uint64_t object, void *data, size_t size) /*ARGSUSED*/ static void +dump_ddt_zap(objset_t *os, uint64_t object, void *data, size_t size) +{ + dump_zap_stats(os, object); + /* contents are printed elsewhere, properly decoded */ +} + +/*ARGSUSED*/ +static void +dump_sa_attrs(objset_t *os, uint64_t object, void *data, size_t size) +{ + zap_cursor_t zc; + zap_attribute_t attr; + + dump_zap_stats(os, object); + (void) printf("\n"); + + for (zap_cursor_init(&zc, os, object); + zap_cursor_retrieve(&zc, &attr) == 0; + zap_cursor_advance(&zc)) { + (void) printf("\t\t%s = ", attr.za_name); + if (attr.za_num_integers == 0) { + (void) printf("\n"); + continue; + } + (void) printf(" %llx : [%d:%d:%d]\n", + (u_longlong_t)attr.za_first_integer, + (int)ATTR_LENGTH(attr.za_first_integer), + (int)ATTR_BSWAP(attr.za_first_integer), + (int)ATTR_NUM(attr.za_first_integer)); + } + zap_cursor_fini(&zc); +} + +/*ARGSUSED*/ +static void +dump_sa_layouts(objset_t *os, uint64_t object, void *data, size_t size) +{ + zap_cursor_t zc; + zap_attribute_t attr; + uint16_t *layout_attrs; + int i; + + dump_zap_stats(os, object); + (void) printf("\n"); + + for (zap_cursor_init(&zc, os, object); + zap_cursor_retrieve(&zc, &attr) == 0; + zap_cursor_advance(&zc)) { + (void) printf("\t\t%s = [", attr.za_name); + if (attr.za_num_integers == 0) { + (void) printf("\n"); + continue; + } + + VERIFY(attr.za_integer_length == 2); + layout_attrs = umem_zalloc(attr.za_num_integers * + attr.za_integer_length, UMEM_NOFAIL); + + VERIFY(zap_lookup(os, object, attr.za_name, + attr.za_integer_length, + attr.za_num_integers, layout_attrs) == 0); + + for (i = 0; i != attr.za_num_integers; i++) + (void) printf(" %d ", (int)layout_attrs[i]); + (void) printf("]\n"); + umem_free(layout_attrs, + attr.za_num_integers * attr.za_integer_length); + } + zap_cursor_fini(&zc); +} + +/*ARGSUSED*/ +static void dump_zpldir(objset_t *os, uint64_t object, void *data, size_t size) { zap_cursor_t zc; @@ -441,17 +500,17 @@ dump_spacemap(objset_t *os, space_map_obj_t *smo, space_map_t *sm) */ alloc = 0; for (offset = 0; offset < smo->smo_objsize; offset += sizeof (entry)) { - VERIFY(0 == dmu_read(os, smo->smo_object, offset, + VERIFY3U(0, ==, dmu_read(os, smo->smo_object, offset, sizeof (entry), &entry, DMU_READ_PREFETCH)); if (SM_DEBUG_DECODE(entry)) { - (void) printf("\t\t[%4llu] %s: txg %llu, pass %llu\n", + (void) printf("\t [%6llu] %s: txg %llu, pass %llu\n", (u_longlong_t)(offset / sizeof (entry)), ddata[SM_DEBUG_ACTION_DECODE(entry)], (u_longlong_t)SM_DEBUG_TXG_DECODE(entry), (u_longlong_t)SM_DEBUG_SYNCPASS_DECODE(entry)); } else { - (void) printf("\t\t[%4llu] %c range:" - " %08llx-%08llx size: %06llx\n", + (void) printf("\t [%6llu] %c range:" + " %010llx-%010llx size: %06llx\n", (u_longlong_t)(offset / sizeof (entry)), SM_TYPE_DECODE(entry) == SM_ALLOC ? 'A' : 'F', (u_longlong_t)((SM_OFFSET_DECODE(entry) << @@ -476,14 +535,14 @@ dump_spacemap(objset_t *os, space_map_obj_t *smo, space_map_t *sm) static void dump_metaslab_stats(metaslab_t *msp) { - char maxbuf[5]; + char maxbuf[32]; space_map_t *sm = &msp->ms_map; avl_tree_t *t = sm->sm_pp_root; int free_pct = sm->sm_space * 100 / sm->sm_size; - nicenum(space_map_maxsize(sm), maxbuf); + zdb_nicenum(space_map_maxsize(sm), maxbuf); - (void) printf("\t %20s %10lu %7s %6s %4s %4d%%\n", + (void) printf("\t %25s %10lu %7s %6s %4s %4d%%\n", "segments", avl_numnodes(t), "maxsize", maxbuf, "freepct", free_pct); } @@ -491,54 +550,86 @@ dump_metaslab_stats(metaslab_t *msp) static void dump_metaslab(metaslab_t *msp) { - char freebuf[5]; - space_map_obj_t *smo = &msp->ms_smo; vdev_t *vd = msp->ms_group->mg_vd; spa_t *spa = vd->vdev_spa; + space_map_t *sm = &msp->ms_map; + space_map_obj_t *smo = &msp->ms_smo; + char freebuf[32]; - nicenum(msp->ms_map.sm_size - smo->smo_alloc, freebuf); + zdb_nicenum(sm->sm_size - smo->smo_alloc, freebuf); (void) printf( - "\tvdev %5llu offset %12llx spacemap %6llu free %5s\n", - (u_longlong_t)vd->vdev_id, (u_longlong_t)msp->ms_map.sm_start, - (u_longlong_t)smo->smo_object, freebuf); + "\tmetaslab %6llu offset %12llx spacemap %6llu free %5s\n", + (u_longlong_t)(sm->sm_start / sm->sm_size), + (u_longlong_t)sm->sm_start, (u_longlong_t)smo->smo_object, freebuf); - if (dump_opt['m'] > 1) { + if (dump_opt['m'] > 1 && !dump_opt['L']) { mutex_enter(&msp->ms_lock); - VERIFY(space_map_load(&msp->ms_map, zfs_metaslab_ops, - SM_FREE, &msp->ms_smo, spa->spa_meta_objset) == 0); + space_map_load_wait(sm); + if (!sm->sm_loaded) + VERIFY(space_map_load(sm, zfs_metaslab_ops, + SM_FREE, smo, spa->spa_meta_objset) == 0); dump_metaslab_stats(msp); - space_map_unload(&msp->ms_map); + space_map_unload(sm); mutex_exit(&msp->ms_lock); } if (dump_opt['d'] > 5 || dump_opt['m'] > 2) { - ASSERT(msp->ms_map.sm_size == (1ULL << vd->vdev_ms_shift)); + ASSERT(sm->sm_size == (1ULL << vd->vdev_ms_shift)); mutex_enter(&msp->ms_lock); - dump_spacemap(spa->spa_meta_objset, smo, &msp->ms_map); + dump_spacemap(spa->spa_meta_objset, smo, sm); mutex_exit(&msp->ms_lock); } +} +static void +print_vdev_metaslab_header(vdev_t *vd) +{ + (void) printf("\tvdev %10llu\n\t%-10s%5llu %-19s %-15s %-10s\n", + (u_longlong_t)vd->vdev_id, + "metaslabs", (u_longlong_t)vd->vdev_ms_count, + "offset", "spacemap", "free"); + (void) printf("\t%15s %19s %15s %10s\n", + "---------------", "-------------------", + "---------------", "-------------"); } static void dump_metaslabs(spa_t *spa) { - vdev_t *rvd = spa->spa_root_vdev; - vdev_t *vd; - int c, m; + vdev_t *vd, *rvd = spa->spa_root_vdev; + uint64_t m, c = 0, children = rvd->vdev_children; (void) printf("\nMetaslabs:\n"); - for (c = 0; c < rvd->vdev_children; c++) { - vd = rvd->vdev_child[c]; + if (!dump_opt['d'] && zopt_objects > 0) { + c = zopt_object[0]; - (void) printf("\t%-10s %-19s %-15s %-10s\n", - "vdev", "offset", "spacemap", "free"); - (void) printf("\t%10s %19s %15s %10s\n", - "----------", "-------------------", - "---------------", "-------------"); + if (c >= children) + (void) fatal("bad vdev id: %llu", (u_longlong_t)c); + + if (zopt_objects > 1) { + vd = rvd->vdev_child[c]; + print_vdev_metaslab_header(vd); + + for (m = 1; m < zopt_objects; m++) { + if (zopt_object[m] < vd->vdev_ms_count) + dump_metaslab( + vd->vdev_ms[zopt_object[m]]); + else + (void) fprintf(stderr, "bad metaslab " + "number %llu\n", + (u_longlong_t)zopt_object[m]); + } + (void) printf("\n"); + return; + } + children = c + 1; + } + for (; c < children; c++) { + vd = rvd->vdev_child[c]; + print_vdev_metaslab_header(vd); for (m = 0; m < vd->vdev_ms_count; m++) dump_metaslab(vd->vdev_ms[m]); @@ -547,6 +638,133 @@ dump_metaslabs(spa_t *spa) } static void +dump_dde(const ddt_t *ddt, const ddt_entry_t *dde, uint64_t index) +{ + const ddt_phys_t *ddp = dde->dde_phys; + const ddt_key_t *ddk = &dde->dde_key; + char *types[4] = { "ditto", "single", "double", "triple" }; + char blkbuf[BP_SPRINTF_LEN]; + blkptr_t blk; + + for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) { + if (ddp->ddp_phys_birth == 0) + continue; + ddt_bp_create(ddt->ddt_checksum, ddk, ddp, &blk); + sprintf_blkptr(blkbuf, &blk); + (void) printf("index %llx refcnt %llu %s %s\n", + (u_longlong_t)index, (u_longlong_t)ddp->ddp_refcnt, + types[p], blkbuf); + } +} + +static void +dump_dedup_ratio(const ddt_stat_t *dds) +{ + double rL, rP, rD, D, dedup, compress, copies; + + if (dds->dds_blocks == 0) + return; + + rL = (double)dds->dds_ref_lsize; + rP = (double)dds->dds_ref_psize; + rD = (double)dds->dds_ref_dsize; + D = (double)dds->dds_dsize; + + dedup = rD / D; + compress = rL / rP; + copies = rD / rP; + + (void) printf("dedup = %.2f, compress = %.2f, copies = %.2f, " + "dedup * compress / copies = %.2f\n\n", + dedup, compress, copies, dedup * compress / copies); +} + +static void +dump_ddt(ddt_t *ddt, enum ddt_type type, enum ddt_class class) +{ + char name[DDT_NAMELEN]; + ddt_entry_t dde; + uint64_t walk = 0; + dmu_object_info_t doi; + uint64_t count, dspace, mspace; + int error; + + error = ddt_object_info(ddt, type, class, &doi); + + if (error == ENOENT) + return; + ASSERT(error == 0); + + count = ddt_object_count(ddt, type, class); + dspace = doi.doi_physical_blocks_512 << 9; + mspace = doi.doi_fill_count * doi.doi_data_block_size; + + ASSERT(count != 0); /* we should have destroyed it */ + + ddt_object_name(ddt, type, class, name); + + (void) printf("%s: %llu entries, size %llu on disk, %llu in core\n", + name, + (u_longlong_t)count, + (u_longlong_t)(dspace / count), + (u_longlong_t)(mspace / count)); + + if (dump_opt['D'] < 3) + return; + + zpool_dump_ddt(NULL, &ddt->ddt_histogram[type][class]); + + if (dump_opt['D'] < 4) + return; + + if (dump_opt['D'] < 5 && class == DDT_CLASS_UNIQUE) + return; + + (void) printf("%s contents:\n\n", name); + + while ((error = ddt_object_walk(ddt, type, class, &walk, &dde)) == 0) + dump_dde(ddt, &dde, walk); + + ASSERT(error == ENOENT); + + (void) printf("\n"); +} + +static void +dump_all_ddts(spa_t *spa) +{ + ddt_histogram_t ddh_total = { 0 }; + ddt_stat_t dds_total = { 0 }; + + for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) { + ddt_t *ddt = spa->spa_ddt[c]; + for (enum ddt_type type = 0; type < DDT_TYPES; type++) { + for (enum ddt_class class = 0; class < DDT_CLASSES; + class++) { + dump_ddt(ddt, type, class); + } + } + } + + ddt_get_dedup_stats(spa, &dds_total); + + if (dds_total.dds_blocks == 0) { + (void) printf("All DDTs are empty\n"); + return; + } + + (void) printf("\n"); + + if (dump_opt['D'] > 1) { + (void) printf("DDT histogram (aggregated over all DDTs):\n"); + ddt_get_dedup_histogram(spa, &ddh_total); + zpool_dump_ddt(&dds_total, &ddh_total); + } + + dump_dedup_ratio(&dds_total); +} + +static void dump_dtl_seg(space_map_t *sm, uint64_t start, uint64_t size) { char *prefix = (void *)sm; @@ -566,7 +784,7 @@ dump_dtl(vdev_t *vd, int indent) char *name[DTL_TYPES] = { "missing", "partial", "scrub", "outage" }; char prefix[256]; - spa_vdev_state_enter(spa); + spa_vdev_state_enter(spa, SCL_NONE); required = vdev_dtl_required(vd); (void) spa_vdev_state_exit(spa, NULL, 0); @@ -596,6 +814,68 @@ dump_dtl(vdev_t *vd, int indent) dump_dtl(vd->vdev_child[c], indent + 4); } +static void +dump_history(spa_t *spa) +{ + nvlist_t **events = NULL; + char buf[SPA_MAXBLOCKSIZE]; + uint64_t resid, len, off = 0; + uint_t num = 0; + int error; + time_t tsec; + struct tm t; + char tbuf[30]; + char internalstr[MAXPATHLEN]; + + do { + len = sizeof (buf); + + if ((error = spa_history_get(spa, &off, &len, buf)) != 0) { + (void) fprintf(stderr, "Unable to read history: " + "error %d\n", error); + return; + } + + if (zpool_history_unpack(buf, len, &resid, &events, &num) != 0) + break; + + off -= resid; + } while (len != 0); + + (void) printf("\nHistory:\n"); + for (int i = 0; i < num; i++) { + uint64_t time, txg, ievent; + char *cmd, *intstr; + + if (nvlist_lookup_uint64(events[i], ZPOOL_HIST_TIME, + &time) != 0) + continue; + if (nvlist_lookup_string(events[i], ZPOOL_HIST_CMD, + &cmd) != 0) { + if (nvlist_lookup_uint64(events[i], + ZPOOL_HIST_INT_EVENT, &ievent) != 0) + continue; + verify(nvlist_lookup_uint64(events[i], + ZPOOL_HIST_TXG, &txg) == 0); + verify(nvlist_lookup_string(events[i], + ZPOOL_HIST_INT_STR, &intstr) == 0); + if (ievent >= LOG_END) + continue; + + (void) snprintf(internalstr, + sizeof (internalstr), + "[internal %s txg:%lld] %s", + zfs_history_event_names[ievent], txg, + intstr); + cmd = internalstr; + } + tsec = time; + (void) localtime_r(&tsec, &t); + (void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t); + (void) printf("%s %s\n", tbuf, cmd); + } +} + /*ARGSUSED*/ static void dump_dnode(objset_t *os, uint64_t object, void *data, size_t size) @@ -603,35 +883,48 @@ dump_dnode(objset_t *os, uint64_t object, void *data, size_t size) } static uint64_t -blkid2offset(const dnode_phys_t *dnp, int level, uint64_t blkid) +blkid2offset(const dnode_phys_t *dnp, const blkptr_t *bp, const zbookmark_t *zb) { - if (level < 0) - return (blkid); + if (dnp == NULL) { + ASSERT(zb->zb_level < 0); + if (zb->zb_object == 0) + return (zb->zb_blkid); + return (zb->zb_blkid * BP_GET_LSIZE(bp)); + } + + ASSERT(zb->zb_level >= 0); - return ((blkid << (level * (dnp->dn_indblkshift - SPA_BLKPTRSHIFT))) * + return ((zb->zb_blkid << + (zb->zb_level * (dnp->dn_indblkshift - SPA_BLKPTRSHIFT))) * dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT); } static void -sprintf_blkptr_compact(char *blkbuf, blkptr_t *bp, int alldvas) +sprintf_blkptr_compact(char *blkbuf, const blkptr_t *bp) { - dva_t *dva = bp->blk_dva; - int ndvas = alldvas ? BP_GET_NDVAS(bp) : 1; - int i; + const dva_t *dva = bp->blk_dva; + int ndvas = dump_opt['d'] > 5 ? BP_GET_NDVAS(bp) : 1; + + if (dump_opt['b'] >= 5) { + sprintf_blkptr(blkbuf, bp); + return; + } blkbuf[0] = '\0'; - for (i = 0; i < ndvas; i++) + for (int i = 0; i < ndvas; i++) (void) sprintf(blkbuf + strlen(blkbuf), "%llu:%llx:%llx ", (u_longlong_t)DVA_GET_VDEV(&dva[i]), (u_longlong_t)DVA_GET_OFFSET(&dva[i]), (u_longlong_t)DVA_GET_ASIZE(&dva[i])); - (void) sprintf(blkbuf + strlen(blkbuf), "%llxL/%llxP F=%llu B=%llu", + (void) sprintf(blkbuf + strlen(blkbuf), + "%llxL/%llxP F=%llu B=%llu/%llu", (u_longlong_t)BP_GET_LSIZE(bp), (u_longlong_t)BP_GET_PSIZE(bp), (u_longlong_t)bp->blk_fill, - (u_longlong_t)bp->blk_birth); + (u_longlong_t)bp->blk_birth, + (u_longlong_t)BP_PHYSICAL_BIRTH(bp)); } static void @@ -644,8 +937,7 @@ print_indirect(blkptr_t *bp, const zbookmark_t *zb, ASSERT3U(BP_GET_TYPE(bp), ==, dnp->dn_type); ASSERT3U(BP_GET_LEVEL(bp), ==, zb->zb_level); - (void) printf("%16llx ", - (u_longlong_t)blkid2offset(dnp, zb->zb_level, zb->zb_blkid)); + (void) printf("%16llx ", (u_longlong_t)blkid2offset(dnp, bp, zb)); ASSERT(zb->zb_level >= 0); @@ -657,23 +949,15 @@ print_indirect(blkptr_t *bp, const zbookmark_t *zb, } } - sprintf_blkptr_compact(blkbuf, bp, dump_opt['d'] > 5 ? 1 : 0); + sprintf_blkptr_compact(blkbuf, bp); (void) printf("%s\n", blkbuf); } -#define SET_BOOKMARK(zb, objset, object, level, blkid) \ -{ \ - (zb)->zb_objset = objset; \ - (zb)->zb_object = object; \ - (zb)->zb_level = level; \ - (zb)->zb_blkid = blkid; \ -} - static int visit_indirect(spa_t *spa, const dnode_phys_t *dnp, blkptr_t *bp, const zbookmark_t *zb) { - int err; + int err = 0; if (bp->blk_birth == 0) return (0); @@ -692,6 +976,7 @@ visit_indirect(spa_t *spa, const dnode_phys_t *dnp, ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &flags, zb); if (err) return (err); + ASSERT(buf->b_data); /* recursively visit blocks below this */ cbp = buf->b_data; @@ -724,11 +1009,11 @@ dump_indirect(dnode_t *dn) (void) printf("Indirect blocks:\n"); - SET_BOOKMARK(&czb, dmu_objset_id(&dn->dn_objset->os), + SET_BOOKMARK(&czb, dmu_objset_id(dn->dn_objset), dn->dn_object, dnp->dn_nlevels - 1, 0); for (j = 0; j < dnp->dn_nblkptr; j++) { czb.zb_blkid = j; - (void) visit_indirect(dmu_objset_spa(&dn->dn_objset->os), dnp, + (void) visit_indirect(dmu_objset_spa(dn->dn_objset), dnp, &dnp->dn_blkptr[j], &czb); } @@ -741,7 +1026,7 @@ dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size) { dsl_dir_phys_t *dd = data; time_t crtime; - char nice[6]; + char nice[32]; if (dd == NULL) return; @@ -758,15 +1043,15 @@ dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size) (u_longlong_t)dd->dd_origin_obj); (void) printf("\t\tchild_dir_zapobj = %llu\n", (u_longlong_t)dd->dd_child_dir_zapobj); - nicenum(dd->dd_used_bytes, nice); + zdb_nicenum(dd->dd_used_bytes, nice); (void) printf("\t\tused_bytes = %s\n", nice); - nicenum(dd->dd_compressed_bytes, nice); + zdb_nicenum(dd->dd_compressed_bytes, nice); (void) printf("\t\tcompressed_bytes = %s\n", nice); - nicenum(dd->dd_uncompressed_bytes, nice); + zdb_nicenum(dd->dd_uncompressed_bytes, nice); (void) printf("\t\tuncompressed_bytes = %s\n", nice); - nicenum(dd->dd_quota, nice); + zdb_nicenum(dd->dd_quota, nice); (void) printf("\t\tquota = %s\n", nice); - nicenum(dd->dd_reserved, nice); + zdb_nicenum(dd->dd_reserved, nice); (void) printf("\t\treserved = %s\n", nice); (void) printf("\t\tprops_zapobj = %llu\n", (u_longlong_t)dd->dd_props_zapobj); @@ -776,7 +1061,7 @@ dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size) (u_longlong_t)dd->dd_flags); #define DO(which) \ - nicenum(dd->dd_used_breakdown[DD_USED_ ## which], nice); \ + zdb_nicenum(dd->dd_used_breakdown[DD_USED_ ## which], nice); \ (void) printf("\t\tused_breakdown[" #which "] = %s\n", nice) DO(HEAD); DO(SNAP); @@ -792,7 +1077,7 @@ dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size) { dsl_dataset_phys_t *ds = data; time_t crtime; - char used[6], compressed[6], uncompressed[6], unique[6]; + char used[32], compressed[32], uncompressed[32], unique[32]; char blkbuf[BP_SPRINTF_LEN]; if (ds == NULL) @@ -800,11 +1085,11 @@ dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size) ASSERT(size == sizeof (*ds)); crtime = ds->ds_creation_time; - nicenum(ds->ds_used_bytes, used); - nicenum(ds->ds_compressed_bytes, compressed); - nicenum(ds->ds_uncompressed_bytes, uncompressed); - nicenum(ds->ds_unique_bytes, unique); - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, &ds->ds_bp); + zdb_nicenum(ds->ds_used_bytes, used); + zdb_nicenum(ds->ds_compressed_bytes, compressed); + zdb_nicenum(ds->ds_uncompressed_bytes, uncompressed); + zdb_nicenum(ds->ds_unique_bytes, unique); + sprintf_blkptr(blkbuf, &ds->ds_bp); (void) printf("\t\tdir_obj = %llu\n", (u_longlong_t)ds->ds_dir_obj); @@ -842,63 +1127,88 @@ dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size) (void) printf("\t\tbp = %s\n", blkbuf); } +/* ARGSUSED */ +static int +dump_bpobj_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) +{ + char blkbuf[BP_SPRINTF_LEN]; + + ASSERT(bp->blk_birth != 0); + sprintf_blkptr_compact(blkbuf, bp); + (void) printf("\t%s\n", blkbuf); + return (0); +} + static void -dump_bplist(objset_t *mos, uint64_t object, char *name) +dump_bpobj(bpobj_t *bpo, char *name) { - bplist_t bpl = { 0 }; - blkptr_t blk, *bp = &blk; - uint64_t itor = 0; - char bytes[6]; - char comp[6]; - char uncomp[6]; + char bytes[32]; + char comp[32]; + char uncomp[32]; if (dump_opt['d'] < 3) return; - mutex_init(&bpl.bpl_lock, NULL, MUTEX_DEFAULT, NULL); - VERIFY(0 == bplist_open(&bpl, mos, object)); - if (bplist_empty(&bpl)) { - bplist_close(&bpl); - mutex_destroy(&bpl.bpl_lock); - return; - } - - nicenum(bpl.bpl_phys->bpl_bytes, bytes); - if (bpl.bpl_dbuf->db_size == sizeof (bplist_phys_t)) { - nicenum(bpl.bpl_phys->bpl_comp, comp); - nicenum(bpl.bpl_phys->bpl_uncomp, uncomp); - (void) printf("\n %s: %llu entries, %s (%s/%s comp)\n", - name, (u_longlong_t)bpl.bpl_phys->bpl_entries, + zdb_nicenum(bpo->bpo_phys->bpo_bytes, bytes); + if (bpo->bpo_havesubobj) { + zdb_nicenum(bpo->bpo_phys->bpo_comp, comp); + zdb_nicenum(bpo->bpo_phys->bpo_uncomp, uncomp); + (void) printf("\n %s: %llu local blkptrs, %llu subobjs, " + "%s (%s/%s comp)\n", + name, (u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs, + (u_longlong_t)bpo->bpo_phys->bpo_num_subobjs, bytes, comp, uncomp); } else { - (void) printf("\n %s: %llu entries, %s\n", - name, (u_longlong_t)bpl.bpl_phys->bpl_entries, bytes); + (void) printf("\n %s: %llu blkptrs, %s\n", + name, (u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs, bytes); } - if (dump_opt['d'] < 5) { - bplist_close(&bpl); - mutex_destroy(&bpl.bpl_lock); + if (dump_opt['d'] < 5) return; - } (void) printf("\n"); - while (bplist_iterate(&bpl, &itor, bp) == 0) { - char blkbuf[BP_SPRINTF_LEN]; + (void) bpobj_iterate_nofree(bpo, dump_bpobj_cb, NULL, NULL); +} - ASSERT(bp->blk_birth != 0); - sprintf_blkptr_compact(blkbuf, bp, dump_opt['d'] > 5 ? 1 : 0); - (void) printf("\tItem %3llu: %s\n", - (u_longlong_t)itor - 1, blkbuf); - } +static void +dump_deadlist(dsl_deadlist_t *dl) +{ + dsl_deadlist_entry_t *dle; + char bytes[32]; + char comp[32]; + char uncomp[32]; + + if (dump_opt['d'] < 3) + return; - bplist_close(&bpl); - mutex_destroy(&bpl.bpl_lock); + zdb_nicenum(dl->dl_phys->dl_used, bytes); + zdb_nicenum(dl->dl_phys->dl_comp, comp); + zdb_nicenum(dl->dl_phys->dl_uncomp, uncomp); + (void) printf("\n Deadlist: %s (%s/%s comp)\n", + bytes, comp, uncomp); + + if (dump_opt['d'] < 4) + return; + + (void) printf("\n"); + + for (dle = avl_first(&dl->dl_tree); dle; + dle = AVL_NEXT(&dl->dl_tree, dle)) { + (void) printf(" mintxg %llu -> obj %llu\n", + (longlong_t)dle->dle_mintxg, + (longlong_t)dle->dle_bpobj.bpo_object); + + if (dump_opt['d'] >= 5) + dump_bpobj(&dle->dle_bpobj, ""); + } } static avl_tree_t idx_tree; static avl_tree_t domain_tree; static boolean_t fuid_table_loaded; +static boolean_t sa_loaded; +sa_attr_type_t *sa_attr_table; static void fuid_table_destroy() @@ -931,12 +1241,12 @@ print_idstr(uint64_t id, const char *id_type) } static void -dump_uidgid(objset_t *os, znode_phys_t *zp) +dump_uidgid(objset_t *os, uint64_t uid, uint64_t gid) { uint32_t uid_idx, gid_idx; - uid_idx = FUID_INDEX(zp->zp_uid); - gid_idx = FUID_INDEX(zp->zp_gid); + uid_idx = FUID_INDEX(uid); + gid_idx = FUID_INDEX(gid); /* Load domain table, if not already loaded */ if (!fuid_table_loaded && (uid_idx || gid_idx)) { @@ -951,50 +1261,107 @@ dump_uidgid(objset_t *os, znode_phys_t *zp) fuid_table_loaded = B_TRUE; } - print_idstr(zp->zp_uid, "uid"); - print_idstr(zp->zp_gid, "gid"); + print_idstr(uid, "uid"); + print_idstr(gid, "gid"); } /*ARGSUSED*/ static void dump_znode(objset_t *os, uint64_t object, void *data, size_t size) { - znode_phys_t *zp = data; - time_t z_crtime, z_atime, z_mtime, z_ctime; char path[MAXPATHLEN * 2]; /* allow for xattr and failure prefix */ + sa_handle_t *hdl; + uint64_t xattr, rdev, gen; + uint64_t uid, gid, mode, fsize, parent, links; + uint64_t pflags; + uint64_t acctm[2], modtm[2], chgtm[2], crtm[2]; + time_t z_crtime, z_atime, z_mtime, z_ctime; + sa_bulk_attr_t bulk[12]; + int idx = 0; int error; - ASSERT(size >= sizeof (znode_phys_t)); + if (!sa_loaded) { + uint64_t sa_attrs = 0; + uint64_t version; + + VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZPL_VERSION_STR, + 8, 1, &version) == 0); + if (version >= ZPL_VERSION_SA) { + VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, + 8, 1, &sa_attrs) == 0); + } + sa_attr_table = sa_setup(os, sa_attrs, + zfs_attr_table, ZPL_END); + sa_loaded = B_TRUE; + } + + if (sa_handle_get(os, object, NULL, SA_HDL_PRIVATE, &hdl)) { + (void) printf("Failed to get handle for SA znode\n"); + return; + } + + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_UID], NULL, &uid, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GID], NULL, &gid, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_LINKS], NULL, + &links, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GEN], NULL, &gen, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MODE], NULL, + &mode, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_PARENT], + NULL, &parent, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_SIZE], NULL, + &fsize, 8); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_ATIME], NULL, + acctm, 16); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MTIME], NULL, + modtm, 16); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CRTIME], NULL, + crtm, 16); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CTIME], NULL, + chgtm, 16); + SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_FLAGS], NULL, + &pflags, 8); + + if (sa_bulk_lookup(hdl, bulk, idx)) { + (void) sa_handle_destroy(hdl); + return; + } error = zfs_obj_to_path(os, object, path, sizeof (path)); if (error != 0) { (void) snprintf(path, sizeof (path), "\?\?\?<object#%llu>", (u_longlong_t)object); } - if (dump_opt['d'] < 3) { (void) printf("\t%s\n", path); + (void) sa_handle_destroy(hdl); return; } - z_crtime = (time_t)zp->zp_crtime[0]; - z_atime = (time_t)zp->zp_atime[0]; - z_mtime = (time_t)zp->zp_mtime[0]; - z_ctime = (time_t)zp->zp_ctime[0]; + z_crtime = (time_t)crtm[0]; + z_atime = (time_t)acctm[0]; + z_mtime = (time_t)modtm[0]; + z_ctime = (time_t)chgtm[0]; (void) printf("\tpath %s\n", path); - dump_uidgid(os, zp); + dump_uidgid(os, uid, gid); (void) printf("\tatime %s", ctime(&z_atime)); (void) printf("\tmtime %s", ctime(&z_mtime)); (void) printf("\tctime %s", ctime(&z_ctime)); (void) printf("\tcrtime %s", ctime(&z_crtime)); - (void) printf("\tgen %llu\n", (u_longlong_t)zp->zp_gen); - (void) printf("\tmode %llo\n", (u_longlong_t)zp->zp_mode); - (void) printf("\tsize %llu\n", (u_longlong_t)zp->zp_size); - (void) printf("\tparent %llu\n", (u_longlong_t)zp->zp_parent); - (void) printf("\tlinks %llu\n", (u_longlong_t)zp->zp_links); - (void) printf("\txattr %llu\n", (u_longlong_t)zp->zp_xattr); - (void) printf("\trdev 0x%016llx\n", (u_longlong_t)zp->zp_rdev); + (void) printf("\tgen %llu\n", (u_longlong_t)gen); + (void) printf("\tmode %llo\n", (u_longlong_t)mode); + (void) printf("\tsize %llu\n", (u_longlong_t)fsize); + (void) printf("\tparent %llu\n", (u_longlong_t)parent); + (void) printf("\tlinks %llu\n", (u_longlong_t)links); + (void) printf("\tpflags %llx\n", (u_longlong_t)pflags); + if (sa_lookup(hdl, sa_attr_table[ZPL_XATTR], &xattr, + sizeof (uint64_t)) == 0) + (void) printf("\txattr %llu\n", (u_longlong_t)xattr); + if (sa_lookup(hdl, sa_attr_table[ZPL_RDEV], &rdev, + sizeof (uint64_t)) == 0) + (void) printf("\trdev 0x%016llx\n", (u_longlong_t)rdev); + sa_handle_destroy(hdl); } /*ARGSUSED*/ @@ -1009,7 +1376,7 @@ dump_dmu_objset(objset_t *os, uint64_t object, void *data, size_t size) { } -static object_viewer_t *object_viewer[DMU_OT_NUMTYPES] = { +static object_viewer_t *object_viewer[DMU_OT_NUMTYPES + 1] = { dump_none, /* unallocated */ dump_zap, /* object directory */ dump_uint64, /* object array */ @@ -1052,6 +1419,19 @@ static object_viewer_t *object_viewer[DMU_OT_NUMTYPES] = { dump_zap, /* ZFS user/group used */ dump_zap, /* ZFS user/group quota */ dump_zap, /* snapshot refcount tags */ + dump_ddt_zap, /* DDT ZAP object */ + dump_zap, /* DDT statistics */ + dump_znode, /* SA object */ + dump_zap, /* SA Master Node */ + dump_sa_attrs, /* SA attribute registration */ + dump_sa_layouts, /* SA attribute layouts */ + dump_zap, /* DSL scrub translations */ + dump_none, /* fake dedup BP */ + dump_zap, /* deadlist */ + dump_none, /* deadlist hdr */ + dump_zap, /* dsl clones */ + dump_none, /* bpobj subobjs */ + dump_unknown, /* Unknown type, must be last */ }; static void @@ -1062,18 +1442,20 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) dnode_t *dn; void *bonus = NULL; size_t bsize = 0; - char iblk[6], dblk[6], lsize[6], asize[6], bonus_size[6], segsize[6]; + char iblk[32], dblk[32], lsize[32], asize[32], fill[32]; + char bonus_size[32]; char aux[50]; int error; if (*print_header) { - (void) printf("\n Object lvl iblk dblk lsize" - " asize type\n"); + (void) printf("\n%10s %3s %5s %5s %5s %5s %6s %s\n", + "Object", "lvl", "iblk", "dblk", "dsize", "lsize", + "%full", "type"); *print_header = 0; } if (object == 0) { - dn = os->os->os_meta_dnode; + dn = os->os_meta_dnode; } else { error = dmu_bonus_hold(os, object, FTAG, &db); if (error) @@ -1085,46 +1467,51 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) } dmu_object_info_from_dnode(dn, &doi); - nicenum(doi.doi_metadata_block_size, iblk); - nicenum(doi.doi_data_block_size, dblk); - nicenum(doi.doi_data_block_size * (doi.doi_max_block_offset + 1), - lsize); - nicenum(doi.doi_physical_blks << 9, asize); - nicenum(doi.doi_bonus_size, bonus_size); + zdb_nicenum(doi.doi_metadata_block_size, iblk); + zdb_nicenum(doi.doi_data_block_size, dblk); + zdb_nicenum(doi.doi_max_offset, lsize); + zdb_nicenum(doi.doi_physical_blocks_512 << 9, asize); + zdb_nicenum(doi.doi_bonus_size, bonus_size); + (void) sprintf(fill, "%6.2f", 100.0 * doi.doi_fill_count * + doi.doi_data_block_size / (object == 0 ? DNODES_PER_BLOCK : 1) / + doi.doi_max_offset); aux[0] = '\0'; if (doi.doi_checksum != ZIO_CHECKSUM_INHERIT || verbosity >= 6) { (void) snprintf(aux + strlen(aux), sizeof (aux), " (K=%s)", - zio_checksum_table[doi.doi_checksum].ci_name); + ZDB_CHECKSUM_NAME(doi.doi_checksum)); } if (doi.doi_compress != ZIO_COMPRESS_INHERIT || verbosity >= 6) { (void) snprintf(aux + strlen(aux), sizeof (aux), " (Z=%s)", - zio_compress_table[doi.doi_compress].ci_name); + ZDB_COMPRESS_NAME(doi.doi_compress)); } - (void) printf("%10lld %3u %5s %5s %5s %5s %s%s\n", - (u_longlong_t)object, doi.doi_indirection, iblk, dblk, lsize, - asize, dmu_ot[doi.doi_type].ot_name, aux); + (void) printf("%10lld %3u %5s %5s %5s %5s %6s %s%s\n", + (u_longlong_t)object, doi.doi_indirection, iblk, dblk, + asize, lsize, fill, ZDB_OT_NAME(doi.doi_type), aux); if (doi.doi_bonus_type != DMU_OT_NONE && verbosity > 3) { - (void) printf("%10s %3s %5s %5s %5s %5s %s\n", - "", "", "", "", bonus_size, "bonus", - dmu_ot[doi.doi_bonus_type].ot_name); + (void) printf("%10s %3s %5s %5s %5s %5s %6s %s\n", + "", "", "", "", "", bonus_size, "bonus", + ZDB_OT_NAME(doi.doi_bonus_type)); } if (verbosity >= 4) { - (void) printf("\tdnode flags: %s%s\n", + (void) printf("\tdnode flags: %s%s%s\n", (dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) ? "USED_BYTES " : "", (dn->dn_phys->dn_flags & DNODE_FLAG_USERUSED_ACCOUNTED) ? - "USERUSED_ACCOUNTED " : ""); + "USERUSED_ACCOUNTED " : "", + (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) ? + "SPILL_BLKPTR" : ""); (void) printf("\tdnode maxblkid: %llu\n", (longlong_t)dn->dn_phys->dn_maxblkid); - object_viewer[doi.doi_bonus_type](os, object, bonus, bsize); - object_viewer[doi.doi_type](os, object, NULL, 0); + object_viewer[ZDB_OT_TYPE(doi.doi_bonus_type)](os, object, + bonus, bsize); + object_viewer[ZDB_OT_TYPE(doi.doi_type)](os, object, NULL, 0); *print_header = 1; } @@ -1146,6 +1533,7 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) } for (;;) { + char segsize[32]; error = dnode_next_offset(dn, 0, &start, minlvl, blkfill, 0); if (error) @@ -1153,7 +1541,7 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) end = start; error = dnode_next_offset(dn, DNODE_FIND_HOLE, &end, minlvl, blkfill, 0); - nicenum(end - start, segsize); + zdb_nicenum(end - start, segsize); (void) printf("\t\tsegment [%016llx, %016llx)" " size %5s\n", (u_longlong_t)start, (u_longlong_t)end, segsize); @@ -1176,7 +1564,7 @@ dump_dir(objset_t *os) dmu_objset_stats_t dds; uint64_t object, object_count; uint64_t refdbytes, usedobjs, scratch; - char numbuf[8]; + char numbuf[32]; char blkbuf[BP_SPRINTF_LEN + 20]; char osname[MAXNAMELEN]; char *type = "UNKNOWN"; @@ -1191,21 +1579,20 @@ dump_dir(objset_t *os) if (dds.dds_type == DMU_OST_META) { dds.dds_creation_txg = TXG_INITIAL; - usedobjs = os->os->os_rootbp->blk_fill; - refdbytes = os->os->os_spa->spa_dsl_pool-> + usedobjs = os->os_rootbp->blk_fill; + refdbytes = os->os_spa->spa_dsl_pool-> dp_mos_dir->dd_phys->dd_used_bytes; } else { dmu_objset_space(os, &refdbytes, &scratch, &usedobjs, &scratch); } - ASSERT3U(usedobjs, ==, os->os->os_rootbp->blk_fill); + ASSERT3U(usedobjs, ==, os->os_rootbp->blk_fill); - nicenum(refdbytes, numbuf); + zdb_nicenum(refdbytes, numbuf); if (verbosity >= 4) { - (void) sprintf(blkbuf + strlen(blkbuf), ", rootbp "); - (void) sprintf_blkptr(blkbuf + strlen(blkbuf), - BP_SPRINTF_LEN - strlen(blkbuf), os->os->os_rootbp); + (void) sprintf(blkbuf, ", rootbp "); + (void) sprintf_blkptr(blkbuf + strlen(blkbuf), os->os_rootbp); } else { blkbuf[0] = '\0'; } @@ -1218,18 +1605,6 @@ dump_dir(objset_t *os) (u_longlong_t)dds.dds_creation_txg, numbuf, (u_longlong_t)usedobjs, blkbuf); - dump_intent_log(dmu_objset_zil(os)); - - if (dmu_objset_ds(os) != NULL) - dump_bplist(dmu_objset_pool(os)->dp_meta_objset, - dmu_objset_ds(os)->ds_phys->ds_deadlist_obj, "Deadlist"); - - if (verbosity < 2) - return; - - if (os->os->os_rootbp->blk_birth == 0) - return; - if (zopt_objects != 0) { for (i = 0; i < zopt_objects; i++) dump_object(os, zopt_object[i], verbosity, @@ -1238,10 +1613,22 @@ dump_dir(objset_t *os) return; } + if (dump_opt['i'] != 0 || verbosity >= 2) + dump_intent_log(dmu_objset_zil(os)); + + if (dmu_objset_ds(os) != NULL) + dump_deadlist(&dmu_objset_ds(os)->ds_deadlist); + + if (verbosity < 2) + return; + + if (os->os_rootbp->blk_birth == 0) + return; + dump_object(os, 0, verbosity, &print_header); object_count = 0; - if (os->os->os_userused_dnode && - os->os->os_userused_dnode->dn_type != 0) { + if (os->os_userused_dnode && + os->os_userused_dnode->dn_type != 0) { dump_object(os, DMU_USERUSED_OBJECT, verbosity, &print_header); dump_object(os, DMU_GROUPUSED_OBJECT, verbosity, &print_header); } @@ -1263,11 +1650,11 @@ dump_dir(objset_t *os) } static void -dump_uberblock(uberblock_t *ub) +dump_uberblock(uberblock_t *ub, const char *header, const char *footer) { time_t timestamp = ub->ub_timestamp; - (void) printf("Uberblock\n\n"); + (void) printf(header ? header : ""); (void) printf("\tmagic = %016llx\n", (u_longlong_t)ub->ub_magic); (void) printf("\tversion = %llu\n", (u_longlong_t)ub->ub_version); (void) printf("\ttxg = %llu\n", (u_longlong_t)ub->ub_txg); @@ -1276,25 +1663,34 @@ dump_uberblock(uberblock_t *ub) (u_longlong_t)ub->ub_timestamp, asctime(localtime(×tamp))); if (dump_opt['u'] >= 3) { char blkbuf[BP_SPRINTF_LEN]; - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, &ub->ub_rootbp); + sprintf_blkptr(blkbuf, &ub->ub_rootbp); (void) printf("\trootbp = %s\n", blkbuf); } - (void) printf("\n"); + (void) printf(footer ? footer : ""); } static void -dump_config(const char *pool) +dump_config(spa_t *spa) { - spa_t *spa = NULL; + dmu_buf_t *db; + size_t nvsize = 0; + int error = 0; + + + error = dmu_bonus_hold(spa->spa_meta_objset, + spa->spa_config_object, FTAG, &db); - mutex_enter(&spa_namespace_lock); - while ((spa = spa_next(spa)) != NULL) { - if (pool == NULL) - (void) printf("%s\n", spa_name(spa)); - if (pool == NULL || strcmp(pool, spa_name(spa)) == 0) - dump_nvlist(spa->spa_config, 4); + if (error == 0) { + nvsize = *(uint64_t *)db->db_data; + dmu_buf_rele(db, FTAG); + + (void) printf("\nMOS Configuration:\n"); + dump_packed_nvlist(spa->spa_meta_objset, + spa->spa_config_object, (void *)&nvsize, 1); + } else { + (void) fprintf(stderr, "dmu_bonus_hold(%llu) failed, errno %d", + (u_longlong_t)spa->spa_config_object, error); } - mutex_exit(&spa_namespace_lock); } static void @@ -1343,33 +1739,75 @@ dump_cachefile(const char *cachefile) nvlist_free(config); } +#define ZDB_MAX_UB_HEADER_SIZE 32 + +static void +dump_label_uberblocks(vdev_label_t *lbl, uint64_t ashift) +{ + vdev_t vd; + vdev_t *vdp = &vd; + char header[ZDB_MAX_UB_HEADER_SIZE]; + + vd.vdev_ashift = ashift; + vdp->vdev_top = vdp; + + for (int i = 0; i < VDEV_UBERBLOCK_COUNT(vdp); i++) { + uint64_t uoff = VDEV_UBERBLOCK_OFFSET(vdp, i); + uberblock_t *ub = (void *)((char *)lbl + uoff); + + if (uberblock_verify(ub)) + continue; + (void) snprintf(header, ZDB_MAX_UB_HEADER_SIZE, + "Uberblock[%d]\n", i); + dump_uberblock(ub, header, ""); + } +} + static void dump_label(const char *dev) { int fd; vdev_label_t label; - char *buf = label.vl_vdev_phys.vp_nvlist; + char *path, *buf = label.vl_vdev_phys.vp_nvlist; size_t buflen = sizeof (label.vl_vdev_phys.vp_nvlist); struct stat64 statbuf; - uint64_t psize; - int l; + uint64_t psize, ashift; + int len = strlen(dev) + 1; + + if (strncmp(dev, "/dev/dsk/", 9) == 0) { + len++; + path = malloc(len); + (void) snprintf(path, len, "%s%s", "/dev/rdsk/", dev + 9); + } else { + path = strdup(dev); + } - if ((fd = open64(dev, O_RDONLY)) < 0) { - (void) printf("cannot open '%s': %s\n", dev, strerror(errno)); + if ((fd = open64(path, O_RDONLY)) < 0) { + (void) printf("cannot open '%s': %s\n", path, strerror(errno)); + free(path); exit(1); } if (fstat64(fd, &statbuf) != 0) { - (void) printf("failed to stat '%s': %s\n", dev, + (void) printf("failed to stat '%s': %s\n", path, strerror(errno)); + free(path); + (void) close(fd); + exit(1); + } + + if (S_ISBLK(statbuf.st_mode)) { + (void) printf("cannot use '%s': character device required\n", + path); + free(path); + (void) close(fd); exit(1); } psize = statbuf.st_size; psize = P2ALIGN(psize, (uint64_t)sizeof (vdev_label_t)); - for (l = 0; l < VDEV_LABELS; l++) { - + for (int l = 0; l < VDEV_LABELS; l++) { nvlist_t *config = NULL; (void) printf("--------------------------------------------\n"); @@ -1384,105 +1822,47 @@ dump_label(const char *dev) if (nvlist_unpack(buf, buflen, &config, 0) != 0) { (void) printf("failed to unpack label %d\n", l); - continue; + ashift = SPA_MINBLOCKSHIFT; + } else { + nvlist_t *vdev_tree = NULL; + + dump_nvlist(config, 4); + if ((nvlist_lookup_nvlist(config, + ZPOOL_CONFIG_VDEV_TREE, &vdev_tree) != 0) || + (nvlist_lookup_uint64(vdev_tree, + ZPOOL_CONFIG_ASHIFT, &ashift) != 0)) + ashift = SPA_MINBLOCKSHIFT; + nvlist_free(config); } - dump_nvlist(config, 4); - nvlist_free(config); + if (dump_opt['u']) + dump_label_uberblocks(&label, ashift); } + + free(path); + (void) close(fd); } /*ARGSUSED*/ static int -dump_one_dir(char *dsname, void *arg) +dump_one_dir(const char *dsname, void *arg) { int error; objset_t *os; - error = dmu_objset_open(dsname, DMU_OST_ANY, - DS_MODE_USER | DS_MODE_READONLY, &os); + error = dmu_objset_own(dsname, DMU_OST_ANY, B_TRUE, FTAG, &os); if (error) { - (void) printf("Could not open %s\n", dsname); + (void) printf("Could not open %s, error %d\n", dsname, error); return (0); } dump_dir(os); - dmu_objset_close(os); + dmu_objset_disown(os, FTAG); fuid_table_destroy(); + sa_loaded = B_FALSE; return (0); } -static void -zdb_leak(space_map_t *sm, uint64_t start, uint64_t size) -{ - vdev_t *vd = sm->sm_ppd; - - (void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n", - (u_longlong_t)vd->vdev_id, (u_longlong_t)start, (u_longlong_t)size); -} - -/* ARGSUSED */ -static void -zdb_space_map_load(space_map_t *sm) -{ -} - -static void -zdb_space_map_unload(space_map_t *sm) -{ - space_map_vacate(sm, zdb_leak, sm); -} - -/* ARGSUSED */ -static void -zdb_space_map_claim(space_map_t *sm, uint64_t start, uint64_t size) -{ -} - -static space_map_ops_t zdb_space_map_ops = { - zdb_space_map_load, - zdb_space_map_unload, - NULL, /* alloc */ - zdb_space_map_claim, - NULL, /* free */ - NULL /* maxsize */ -}; - -static void -zdb_leak_init(spa_t *spa) -{ - vdev_t *rvd = spa->spa_root_vdev; - - for (int c = 0; c < rvd->vdev_children; c++) { - vdev_t *vd = rvd->vdev_child[c]; - for (int m = 0; m < vd->vdev_ms_count; m++) { - metaslab_t *msp = vd->vdev_ms[m]; - mutex_enter(&msp->ms_lock); - VERIFY(space_map_load(&msp->ms_map, &zdb_space_map_ops, - SM_ALLOC, &msp->ms_smo, spa->spa_meta_objset) == 0); - msp->ms_map.sm_ppd = vd; - mutex_exit(&msp->ms_lock); - } - } -} - -static void -zdb_leak_fini(spa_t *spa) -{ - vdev_t *rvd = spa->spa_root_vdev; - - for (int c = 0; c < rvd->vdev_children; c++) { - vdev_t *vd = rvd->vdev_child[c]; - for (int m = 0; m < vd->vdev_ms_count; m++) { - metaslab_t *msp = vd->vdev_ms[m]; - mutex_enter(&msp->ms_lock); - space_map_unload(&msp->ms_map); - mutex_exit(&msp->ms_lock); - } - } -} - /* - * Verify that the sum of the sizes of all blocks in the pool adds up - * to the SPA's sa_alloc total. + * Block statistics. */ typedef struct zdb_blkstats { uint64_t zb_asize; @@ -1491,24 +1871,45 @@ typedef struct zdb_blkstats { uint64_t zb_count; } zdb_blkstats_t; -#define DMU_OT_DEFERRED DMU_OT_NONE -#define DMU_OT_TOTAL DMU_OT_NUMTYPES +/* + * Extended object types to report deferred frees and dedup auto-ditto blocks. + */ +#define ZDB_OT_DEFERRED (DMU_OT_NUMTYPES + 0) +#define ZDB_OT_DITTO (DMU_OT_NUMTYPES + 1) +#define ZDB_OT_TOTAL (DMU_OT_NUMTYPES + 2) + +static char *zdb_ot_extname[] = { + "deferred free", + "dedup ditto", + "Total", +}; #define ZB_TOTAL DN_MAX_LEVELS typedef struct zdb_cb { - zdb_blkstats_t zcb_type[ZB_TOTAL + 1][DMU_OT_TOTAL + 1]; + zdb_blkstats_t zcb_type[ZB_TOTAL + 1][ZDB_OT_TOTAL + 1]; + uint64_t zcb_dedup_asize; + uint64_t zcb_dedup_blocks; uint64_t zcb_errors[256]; int zcb_readfails; int zcb_haderrors; + spa_t *zcb_spa; } zdb_cb_t; static void -zdb_count_block(spa_t *spa, zdb_cb_t *zcb, blkptr_t *bp, dmu_object_type_t type) +zdb_count_block(zdb_cb_t *zcb, zilog_t *zilog, const blkptr_t *bp, + dmu_object_type_t type) { + uint64_t refcnt = 0; + + ASSERT(type < ZDB_OT_TOTAL); + + if (zilog && zil_bp_tree_add(zilog, bp) != 0) + return; + for (int i = 0; i < 4; i++) { int l = (i < 2) ? BP_GET_LEVEL(bp) : ZB_TOTAL; - int t = (i & 1) ? type : DMU_OT_TOTAL; + int t = (i & 1) ? type : ZDB_OT_TOTAL; zdb_blkstats_t *zb = &zcb->zcb_type[l][t]; zb->zb_asize += BP_GET_ASIZE(bp); @@ -1517,127 +1918,258 @@ zdb_count_block(spa_t *spa, zdb_cb_t *zcb, blkptr_t *bp, dmu_object_type_t type) zb->zb_count++; } - if (dump_opt['S']) { - boolean_t print_sig; - - print_sig = !zdb_sig_user_data || (BP_GET_LEVEL(bp) == 0 && - BP_GET_TYPE(bp) == DMU_OT_PLAIN_FILE_CONTENTS); - - if (BP_GET_CHECKSUM(bp) < zdb_sig_cksumalg) - print_sig = B_FALSE; - - if (print_sig) { - (void) printf("%llu\t%lld\t%lld\t%s\t%s\t%s\t" - "%llx:%llx:%llx:%llx\n", - (u_longlong_t)BP_GET_LEVEL(bp), - (longlong_t)BP_GET_PSIZE(bp), - (longlong_t)BP_GET_NDVAS(bp), - dmu_ot[BP_GET_TYPE(bp)].ot_name, - zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name, - zio_compress_table[BP_GET_COMPRESS(bp)].ci_name, - (u_longlong_t)bp->blk_cksum.zc_word[0], - (u_longlong_t)bp->blk_cksum.zc_word[1], - (u_longlong_t)bp->blk_cksum.zc_word[2], - (u_longlong_t)bp->blk_cksum.zc_word[3]); + if (dump_opt['L']) + return; + + if (BP_GET_DEDUP(bp)) { + ddt_t *ddt; + ddt_entry_t *dde; + + ddt = ddt_select(zcb->zcb_spa, bp); + ddt_enter(ddt); + dde = ddt_lookup(ddt, bp, B_FALSE); + + if (dde == NULL) { + refcnt = 0; + } else { + ddt_phys_t *ddp = ddt_phys_select(dde, bp); + ddt_phys_decref(ddp); + refcnt = ddp->ddp_refcnt; + if (ddt_phys_total_refcnt(dde) == 0) + ddt_remove(ddt, dde); } + ddt_exit(ddt); } - if (!dump_opt['L']) - VERIFY(zio_wait(zio_claim(NULL, spa, spa_first_txg(spa), bp, - NULL, NULL, ZIO_FLAG_MUSTSUCCEED)) == 0); + VERIFY3U(zio_wait(zio_claim(NULL, zcb->zcb_spa, + refcnt ? 0 : spa_first_txg(zcb->zcb_spa), + bp, NULL, NULL, ZIO_FLAG_CANFAIL)), ==, 0); } +/* ARGSUSED */ static int -zdb_blkptr_cb(spa_t *spa, blkptr_t *bp, const zbookmark_t *zb, - const dnode_phys_t *dnp, void *arg) +zdb_blkptr_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, arc_buf_t *pbuf, + const zbookmark_t *zb, const dnode_phys_t *dnp, void *arg) { zdb_cb_t *zcb = arg; char blkbuf[BP_SPRINTF_LEN]; dmu_object_type_t type; - boolean_t is_l0_metadata; + boolean_t is_metadata; if (bp == NULL) return (0); type = BP_GET_TYPE(bp); - zdb_count_block(spa, zcb, bp, type); + zdb_count_block(zcb, zilog, bp, type); - /* - * if we do metadata-only checksumming there's no need to checksum - * indirect blocks here because it is done during traverse - */ - is_l0_metadata = (BP_GET_LEVEL(bp) == 0 && type < DMU_OT_NUMTYPES && - dmu_ot[type].ot_metadata); + is_metadata = (BP_GET_LEVEL(bp) != 0 || dmu_ot[type].ot_metadata); - if (dump_opt['c'] > 1 || dump_opt['S'] || - (dump_opt['c'] && is_l0_metadata)) { - int ioerr, size; - void *data; + if (dump_opt['c'] > 1 || (dump_opt['c'] && is_metadata)) { + int ioerr; + size_t size = BP_GET_PSIZE(bp); + void *data = malloc(size); + int flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB | ZIO_FLAG_RAW; + + /* If it's an intent log block, failure is expected. */ + if (zb->zb_level == ZB_ZIL_LEVEL) + flags |= ZIO_FLAG_SPECULATIVE; - size = BP_GET_LSIZE(bp); - data = malloc(size); ioerr = zio_wait(zio_read(NULL, spa, bp, data, size, - NULL, NULL, ZIO_PRIORITY_ASYNC_READ, - ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB, zb)); + NULL, NULL, ZIO_PRIORITY_ASYNC_READ, flags, zb)); + free(data); - /* We expect io errors on intent log */ - if (ioerr && type != DMU_OT_INTENT_LOG) { + if (ioerr && !(flags & ZIO_FLAG_SPECULATIVE)) { zcb->zcb_haderrors = 1; zcb->zcb_errors[ioerr]++; if (dump_opt['b'] >= 2) - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, bp); + sprintf_blkptr(blkbuf, bp); else blkbuf[0] = '\0'; - if (!dump_opt['S']) { - (void) printf("zdb_blkptr_cb: " - "Got error %d reading " - "<%llu, %llu, %lld, %llx> %s -- skipping\n", - ioerr, - (u_longlong_t)zb->zb_objset, - (u_longlong_t)zb->zb_object, - (u_longlong_t)zb->zb_level, - (u_longlong_t)zb->zb_blkid, - blkbuf); - } + (void) printf("zdb_blkptr_cb: " + "Got error %d reading " + "<%llu, %llu, %lld, %llx> %s -- skipping\n", + ioerr, + (u_longlong_t)zb->zb_objset, + (u_longlong_t)zb->zb_object, + (u_longlong_t)zb->zb_level, + (u_longlong_t)zb->zb_blkid, + blkbuf); } } zcb->zcb_readfails = 0; if (dump_opt['b'] >= 4) { - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, bp); - (void) printf("objset %llu object %llu offset 0x%llx %s\n", + sprintf_blkptr(blkbuf, bp); + (void) printf("objset %llu object %llu " + "level %lld offset 0x%llx %s\n", (u_longlong_t)zb->zb_objset, (u_longlong_t)zb->zb_object, - (u_longlong_t)blkid2offset(dnp, zb->zb_level, zb->zb_blkid), + (longlong_t)zb->zb_level, + (u_longlong_t)blkid2offset(dnp, bp, zb), blkbuf); } return (0); } +static void +zdb_leak(space_map_t *sm, uint64_t start, uint64_t size) +{ + vdev_t *vd = sm->sm_ppd; + + (void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n", + (u_longlong_t)vd->vdev_id, (u_longlong_t)start, (u_longlong_t)size); +} + +/* ARGSUSED */ +static void +zdb_space_map_load(space_map_t *sm) +{ +} + +static void +zdb_space_map_unload(space_map_t *sm) +{ + space_map_vacate(sm, zdb_leak, sm); +} + +/* ARGSUSED */ +static void +zdb_space_map_claim(space_map_t *sm, uint64_t start, uint64_t size) +{ +} + +static space_map_ops_t zdb_space_map_ops = { + zdb_space_map_load, + zdb_space_map_unload, + NULL, /* alloc */ + zdb_space_map_claim, + NULL, /* free */ + NULL /* maxsize */ +}; + +static void +zdb_ddt_leak_init(spa_t *spa, zdb_cb_t *zcb) +{ + ddt_bookmark_t ddb = { 0 }; + ddt_entry_t dde; + int error; + + while ((error = ddt_walk(spa, &ddb, &dde)) == 0) { + blkptr_t blk; + ddt_phys_t *ddp = dde.dde_phys; + + if (ddb.ddb_class == DDT_CLASS_UNIQUE) + return; + + ASSERT(ddt_phys_total_refcnt(&dde) > 1); + + for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) { + if (ddp->ddp_phys_birth == 0) + continue; + ddt_bp_create(ddb.ddb_checksum, + &dde.dde_key, ddp, &blk); + if (p == DDT_PHYS_DITTO) { + zdb_count_block(zcb, NULL, &blk, ZDB_OT_DITTO); + } else { + zcb->zcb_dedup_asize += + BP_GET_ASIZE(&blk) * (ddp->ddp_refcnt - 1); + zcb->zcb_dedup_blocks++; + } + } + if (!dump_opt['L']) { + ddt_t *ddt = spa->spa_ddt[ddb.ddb_checksum]; + ddt_enter(ddt); + VERIFY(ddt_lookup(ddt, &blk, B_TRUE) != NULL); + ddt_exit(ddt); + } + } + + ASSERT(error == ENOENT); +} + +static void +zdb_leak_init(spa_t *spa, zdb_cb_t *zcb) +{ + zcb->zcb_spa = spa; + + if (!dump_opt['L']) { + vdev_t *rvd = spa->spa_root_vdev; + for (int c = 0; c < rvd->vdev_children; c++) { + vdev_t *vd = rvd->vdev_child[c]; + for (int m = 0; m < vd->vdev_ms_count; m++) { + metaslab_t *msp = vd->vdev_ms[m]; + mutex_enter(&msp->ms_lock); + space_map_unload(&msp->ms_map); + VERIFY(space_map_load(&msp->ms_map, + &zdb_space_map_ops, SM_ALLOC, &msp->ms_smo, + spa->spa_meta_objset) == 0); + msp->ms_map.sm_ppd = vd; + mutex_exit(&msp->ms_lock); + } + } + } + + spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); + + zdb_ddt_leak_init(spa, zcb); + + spa_config_exit(spa, SCL_CONFIG, FTAG); +} + +static void +zdb_leak_fini(spa_t *spa) +{ + if (!dump_opt['L']) { + vdev_t *rvd = spa->spa_root_vdev; + for (int c = 0; c < rvd->vdev_children; c++) { + vdev_t *vd = rvd->vdev_child[c]; + for (int m = 0; m < vd->vdev_ms_count; m++) { + metaslab_t *msp = vd->vdev_ms[m]; + mutex_enter(&msp->ms_lock); + space_map_unload(&msp->ms_map); + mutex_exit(&msp->ms_lock); + } + } + } +} + +/* ARGSUSED */ +static int +count_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) +{ + zdb_cb_t *zcb = arg; + + if (dump_opt['b'] >= 4) { + char blkbuf[BP_SPRINTF_LEN]; + sprintf_blkptr(blkbuf, bp); + (void) printf("[%s] %s\n", + "deferred free", blkbuf); + } + zdb_count_block(zcb, NULL, bp, ZDB_OT_DEFERRED); + return (0); +} + static int dump_block_stats(spa_t *spa) { zdb_cb_t zcb = { 0 }; zdb_blkstats_t *zb, *tzb; - uint64_t alloc, space, logalloc; - vdev_t *rvd = spa->spa_root_vdev; + uint64_t norm_alloc, norm_space, total_alloc, total_found; + int flags = TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA | TRAVERSE_HARD; int leaks = 0; - int c, e; - if (!dump_opt['S']) { - (void) printf("\nTraversing all blocks %s%s%s%s%s...\n", - (dump_opt['c'] || !dump_opt['L']) ? "to verify " : "", - (dump_opt['c'] == 1) ? "metadata " : "", - dump_opt['c'] ? "checksums " : "", - (dump_opt['c'] && !dump_opt['L']) ? "and verify " : "", - !dump_opt['L'] ? "nothing leaked " : ""); - } + (void) printf("\nTraversing all blocks %s%s%s%s%s...\n", + (dump_opt['c'] || !dump_opt['L']) ? "to verify " : "", + (dump_opt['c'] == 1) ? "metadata " : "", + dump_opt['c'] ? "checksums " : "", + (dump_opt['c'] && !dump_opt['L']) ? "and verify " : "", + !dump_opt['L'] ? "nothing leaked " : ""); /* * Load all space maps as SM_ALLOC maps, then traverse the pool @@ -1647,39 +2179,25 @@ dump_block_stats(spa_t *spa) * it's not part of any space map) is a double allocation, * reference to a freed block, or an unclaimed log block. */ - if (!dump_opt['L']) - zdb_leak_init(spa); + zdb_leak_init(spa, &zcb); /* * If there's a deferred-free bplist, process that first. */ - if (spa->spa_sync_bplist_obj != 0) { - bplist_t *bpl = &spa->spa_sync_bplist; - blkptr_t blk; - uint64_t itor = 0; - - VERIFY(0 == bplist_open(bpl, spa->spa_meta_objset, - spa->spa_sync_bplist_obj)); - - while (bplist_iterate(bpl, &itor, &blk) == 0) { - if (dump_opt['b'] >= 4) { - char blkbuf[BP_SPRINTF_LEN]; - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, &blk); - (void) printf("[%s] %s\n", - "deferred free", blkbuf); - } - zdb_count_block(spa, &zcb, &blk, DMU_OT_DEFERRED); - } + (void) bpobj_iterate_nofree(&spa->spa_deferred_bpobj, + count_block_cb, &zcb, NULL); + (void) bpobj_iterate_nofree(&spa->spa_dsl_pool->dp_free_bpobj, + count_block_cb, &zcb, NULL); - bplist_close(bpl); - } + if (dump_opt['c'] > 1) + flags |= TRAVERSE_PREFETCH_DATA; - zcb.zcb_haderrors |= traverse_pool(spa, zdb_blkptr_cb, &zcb); + zcb.zcb_haderrors |= traverse_pool(spa, 0, flags, zdb_blkptr_cb, &zcb); - if (zcb.zcb_haderrors && !dump_opt['S']) { + if (zcb.zcb_haderrors) { (void) printf("\nError counts:\n\n"); (void) printf("\t%5s %s\n", "errno", "count"); - for (e = 0; e < 256; e++) { + for (int e = 0; e < 256; e++) { if (zcb.zcb_errors[e] != 0) { (void) printf("\t%5d %llu\n", e, (u_longlong_t)zcb.zcb_errors[e]); @@ -1690,43 +2208,27 @@ dump_block_stats(spa_t *spa) /* * Report any leaked segments. */ - if (!dump_opt['L']) - zdb_leak_fini(spa); - - /* - * If we're interested in printing out the blkptr signatures, - * return now as we don't print out anything else (including - * errors and leaks). - */ - if (dump_opt['S']) - return (zcb.zcb_haderrors ? 3 : 0); - - alloc = spa_get_alloc(spa); - space = spa_get_space(spa); + zdb_leak_fini(spa); - /* - * Log blocks allocated from a separate log device don't count - * as part of the normal pool space; factor them in here. - */ - logalloc = 0; + tzb = &zcb.zcb_type[ZB_TOTAL][ZDB_OT_TOTAL]; - for (c = 0; c < rvd->vdev_children; c++) - if (rvd->vdev_child[c]->vdev_islog) - logalloc += rvd->vdev_child[c]->vdev_stat.vs_alloc; + norm_alloc = metaslab_class_get_alloc(spa_normal_class(spa)); + norm_space = metaslab_class_get_space(spa_normal_class(spa)); - tzb = &zcb.zcb_type[ZB_TOTAL][DMU_OT_TOTAL]; + total_alloc = norm_alloc + metaslab_class_get_alloc(spa_log_class(spa)); + total_found = tzb->zb_asize - zcb.zcb_dedup_asize; - if (tzb->zb_asize == alloc + logalloc) { + if (total_found == total_alloc) { if (!dump_opt['L']) (void) printf("\n\tNo leaks (block sum matches space" " maps exactly)\n"); } else { (void) printf("block traversal size %llu != alloc %llu " "(%s %lld)\n", - (u_longlong_t)tzb->zb_asize, - (u_longlong_t)alloc + logalloc, + (u_longlong_t)total_found, + (u_longlong_t)total_alloc, (dump_opt['L']) ? "unreachable" : "leaked", - (longlong_t)(alloc + logalloc - tzb->zb_asize)); + (longlong_t)(total_alloc - total_found)); leaks = 1; } @@ -1736,33 +2238,41 @@ dump_block_stats(spa_t *spa) (void) printf("\n"); (void) printf("\tbp count: %10llu\n", (u_longlong_t)tzb->zb_count); - (void) printf("\tbp logical: %10llu\t avg: %6llu\n", + (void) printf("\tbp logical: %10llu avg: %6llu\n", (u_longlong_t)tzb->zb_lsize, (u_longlong_t)(tzb->zb_lsize / tzb->zb_count)); - (void) printf("\tbp physical: %10llu\t avg:" - " %6llu\tcompression: %6.2f\n", + (void) printf("\tbp physical: %10llu avg:" + " %6llu compression: %6.2f\n", (u_longlong_t)tzb->zb_psize, (u_longlong_t)(tzb->zb_psize / tzb->zb_count), (double)tzb->zb_lsize / tzb->zb_psize); - (void) printf("\tbp allocated: %10llu\t avg:" - " %6llu\tcompression: %6.2f\n", + (void) printf("\tbp allocated: %10llu avg:" + " %6llu compression: %6.2f\n", (u_longlong_t)tzb->zb_asize, (u_longlong_t)(tzb->zb_asize / tzb->zb_count), (double)tzb->zb_lsize / tzb->zb_asize); - (void) printf("\tSPA allocated: %10llu\tused: %5.2f%%\n", - (u_longlong_t)alloc, 100.0 * alloc / space); + (void) printf("\tbp deduped: %10llu ref>1:" + " %6llu deduplication: %6.2f\n", + (u_longlong_t)zcb.zcb_dedup_asize, + (u_longlong_t)zcb.zcb_dedup_blocks, + (double)zcb.zcb_dedup_asize / tzb->zb_asize + 1.0); + (void) printf("\tSPA allocated: %10llu used: %5.2f%%\n", + (u_longlong_t)norm_alloc, 100.0 * norm_alloc / norm_space); if (dump_opt['b'] >= 2) { int l, t, level; (void) printf("\nBlocks\tLSIZE\tPSIZE\tASIZE" "\t avg\t comp\t%%Total\tType\n"); - for (t = 0; t <= DMU_OT_NUMTYPES; t++) { - char csize[6], lsize[6], psize[6], asize[6], avg[6]; + for (t = 0; t <= ZDB_OT_TOTAL; t++) { + char csize[32], lsize[32], psize[32], asize[32]; + char avg[32]; char *typename; - typename = t == DMU_OT_DEFERRED ? "deferred free" : - t == DMU_OT_TOTAL ? "Total" : dmu_ot[t].ot_name; + if (t < DMU_OT_NUMTYPES) + typename = dmu_ot[t].ot_name; + else + typename = zdb_ot_extname[t - DMU_OT_NUMTYPES]; if (zcb.zcb_type[ZB_TOTAL][t].zb_asize == 0) { (void) printf("%6s\t%5s\t%5s\t%5s" @@ -1792,11 +2302,11 @@ dump_block_stats(spa_t *spa) zcb.zcb_type[ZB_TOTAL][t].zb_asize) continue; - nicenum(zb->zb_count, csize); - nicenum(zb->zb_lsize, lsize); - nicenum(zb->zb_psize, psize); - nicenum(zb->zb_asize, asize); - nicenum(zb->zb_asize / zb->zb_count, avg); + zdb_nicenum(zb->zb_count, csize); + zdb_nicenum(zb->zb_lsize, lsize); + zdb_nicenum(zb->zb_psize, psize); + zdb_nicenum(zb->zb_asize, asize); + zdb_nicenum(zb->zb_asize / zb->zb_count, avg); (void) printf("%6s\t%5s\t%5s\t%5s\t%5s" "\t%5.2f\t%6.2f\t", @@ -1824,36 +2334,157 @@ dump_block_stats(spa_t *spa) return (0); } +typedef struct zdb_ddt_entry { + ddt_key_t zdde_key; + uint64_t zdde_ref_blocks; + uint64_t zdde_ref_lsize; + uint64_t zdde_ref_psize; + uint64_t zdde_ref_dsize; + avl_node_t zdde_node; +} zdb_ddt_entry_t; + +/* ARGSUSED */ +static int +zdb_ddt_add_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, + arc_buf_t *pbuf, const zbookmark_t *zb, const dnode_phys_t *dnp, void *arg) +{ + avl_tree_t *t = arg; + avl_index_t where; + zdb_ddt_entry_t *zdde, zdde_search; + + if (bp == NULL) + return (0); + + if (dump_opt['S'] > 1 && zb->zb_level == ZB_ROOT_LEVEL) { + (void) printf("traversing objset %llu, %llu objects, " + "%lu blocks so far\n", + (u_longlong_t)zb->zb_objset, + (u_longlong_t)bp->blk_fill, + avl_numnodes(t)); + } + + if (BP_IS_HOLE(bp) || BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_OFF || + BP_GET_LEVEL(bp) > 0 || dmu_ot[BP_GET_TYPE(bp)].ot_metadata) + return (0); + + ddt_key_fill(&zdde_search.zdde_key, bp); + + zdde = avl_find(t, &zdde_search, &where); + + if (zdde == NULL) { + zdde = umem_zalloc(sizeof (*zdde), UMEM_NOFAIL); + zdde->zdde_key = zdde_search.zdde_key; + avl_insert(t, zdde, where); + } + + zdde->zdde_ref_blocks += 1; + zdde->zdde_ref_lsize += BP_GET_LSIZE(bp); + zdde->zdde_ref_psize += BP_GET_PSIZE(bp); + zdde->zdde_ref_dsize += bp_get_dsize_sync(spa, bp); + + return (0); +} + +static void +dump_simulated_ddt(spa_t *spa) +{ + avl_tree_t t; + void *cookie = NULL; + zdb_ddt_entry_t *zdde; + ddt_histogram_t ddh_total = { 0 }; + ddt_stat_t dds_total = { 0 }; + + avl_create(&t, ddt_entry_compare, + sizeof (zdb_ddt_entry_t), offsetof(zdb_ddt_entry_t, zdde_node)); + + spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); + + (void) traverse_pool(spa, 0, TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA, + zdb_ddt_add_cb, &t); + + spa_config_exit(spa, SCL_CONFIG, FTAG); + + while ((zdde = avl_destroy_nodes(&t, &cookie)) != NULL) { + ddt_stat_t dds; + uint64_t refcnt = zdde->zdde_ref_blocks; + ASSERT(refcnt != 0); + + dds.dds_blocks = zdde->zdde_ref_blocks / refcnt; + dds.dds_lsize = zdde->zdde_ref_lsize / refcnt; + dds.dds_psize = zdde->zdde_ref_psize / refcnt; + dds.dds_dsize = zdde->zdde_ref_dsize / refcnt; + + dds.dds_ref_blocks = zdde->zdde_ref_blocks; + dds.dds_ref_lsize = zdde->zdde_ref_lsize; + dds.dds_ref_psize = zdde->zdde_ref_psize; + dds.dds_ref_dsize = zdde->zdde_ref_dsize; + + ddt_stat_add(&ddh_total.ddh_stat[highbit(refcnt) - 1], &dds, 0); + + umem_free(zdde, sizeof (*zdde)); + } + + avl_destroy(&t); + + ddt_histogram_stat(&dds_total, &ddh_total); + + (void) printf("Simulated DDT histogram:\n"); + + zpool_dump_ddt(&dds_total, &ddh_total); + + dump_dedup_ratio(&dds_total); +} + static void dump_zpool(spa_t *spa) { dsl_pool_t *dp = spa_get_dsl(spa); int rc = 0; + if (dump_opt['S']) { + dump_simulated_ddt(spa); + return; + } + + if (!dump_opt['e'] && dump_opt['C'] > 1) { + (void) printf("\nCached configuration:\n"); + dump_nvlist(spa->spa_config, 8); + } + + if (dump_opt['C']) + dump_config(spa); + if (dump_opt['u']) - dump_uberblock(&spa->spa_uberblock); + dump_uberblock(&spa->spa_uberblock, "\nUberblock:\n", "\n"); + + if (dump_opt['D']) + dump_all_ddts(spa); + + if (dump_opt['d'] > 2 || dump_opt['m']) + dump_metaslabs(spa); - if (dump_opt['d'] || dump_opt['i'] || dump_opt['m']) { + if (dump_opt['d'] || dump_opt['i']) { dump_dir(dp->dp_meta_objset); if (dump_opt['d'] >= 3) { - dump_bplist(dp->dp_meta_objset, - spa->spa_sync_bplist_obj, "Deferred frees"); + dump_bpobj(&spa->spa_deferred_bpobj, "Deferred frees"); + if (spa_version(spa) >= SPA_VERSION_DEADLISTS) { + dump_bpobj(&spa->spa_dsl_pool->dp_free_bpobj, + "Pool frees"); + } dump_dtl(spa->spa_root_vdev, 0); } - - if (dump_opt['d'] >= 3 || dump_opt['m']) - dump_metaslabs(spa); - - (void) dmu_objset_find(spa_name(spa), dump_one_dir, NULL, - DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN); + (void) dmu_objset_find(spa_name(spa), dump_one_dir, + NULL, DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN); } - - if (dump_opt['b'] || dump_opt['c'] || dump_opt['S']) + if (dump_opt['b'] || dump_opt['c']) rc = dump_block_stats(spa); if (dump_opt['s']) show_pool_stats(spa); + if (dump_opt['h']) + dump_history(spa); + if (rc != 0) exit(rc); } @@ -1872,51 +2503,13 @@ int flagbits[256]; static void zdb_print_blkptr(blkptr_t *bp, int flags) { - dva_t *dva = bp->blk_dva; - int d; + char blkbuf[BP_SPRINTF_LEN]; if (flags & ZDB_FLAG_BSWAP) byteswap_uint64_array((void *)bp, sizeof (blkptr_t)); - /* - * Super-ick warning: This code is also duplicated in - * cmd/mdb/common/modules/zfs/zfs.c . Yeah, I hate code - * replication, too. - */ - for (d = 0; d < BP_GET_NDVAS(bp); d++) { - (void) printf("\tDVA[%d]: vdev_id %lld / %llx\n", d, - (longlong_t)DVA_GET_VDEV(&dva[d]), - (longlong_t)DVA_GET_OFFSET(&dva[d])); - (void) printf("\tDVA[%d]: GANG: %-5s GRID: %04llx\t" - "ASIZE: %llx\n", d, - DVA_GET_GANG(&dva[d]) ? "TRUE" : "FALSE", - (longlong_t)DVA_GET_GRID(&dva[d]), - (longlong_t)DVA_GET_ASIZE(&dva[d])); - (void) printf("\tDVA[%d]: :%llu:%llx:%llx:%s%s%s%s\n", d, - (u_longlong_t)DVA_GET_VDEV(&dva[d]), - (longlong_t)DVA_GET_OFFSET(&dva[d]), - (longlong_t)BP_GET_PSIZE(bp), - BP_SHOULD_BYTESWAP(bp) ? "e" : "", - !DVA_GET_GANG(&dva[d]) && BP_GET_LEVEL(bp) != 0 ? - "d" : "", - DVA_GET_GANG(&dva[d]) ? "g" : "", - BP_GET_COMPRESS(bp) != 0 ? "d" : ""); - } - (void) printf("\tLSIZE: %-16llx\t\tPSIZE: %llx\n", - (longlong_t)BP_GET_LSIZE(bp), (longlong_t)BP_GET_PSIZE(bp)); - (void) printf("\tENDIAN: %6s\t\t\t\t\tTYPE: %s\n", - BP_GET_BYTEORDER(bp) ? "LITTLE" : "BIG", - dmu_ot[BP_GET_TYPE(bp)].ot_name); - (void) printf("\tBIRTH: %-16llx LEVEL: %-2llu\tFILL: %llx\n", - (u_longlong_t)bp->blk_birth, (u_longlong_t)BP_GET_LEVEL(bp), - (u_longlong_t)bp->blk_fill); - (void) printf("\tCKFUNC: %-16s\t\tCOMP: %s\n", - zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name, - zio_compress_table[BP_GET_COMPRESS(bp)].ci_name); - (void) printf("\tCKSUM: %llx:%llx:%llx:%llx\n", - (u_longlong_t)bp->blk_cksum.zc_word[0], - (u_longlong_t)bp->blk_cksum.zc_word[1], - (u_longlong_t)bp->blk_cksum.zc_word[2], - (u_longlong_t)bp->blk_cksum.zc_word[3]); + + sprintf_blkptr(blkbuf, bp); + (void) printf("%s\n", blkbuf); } static void @@ -1939,7 +2532,7 @@ zdb_dump_block_raw(void *buf, uint64_t size, int flags) { if (flags & ZDB_FLAG_BSWAP) byteswap_uint64_array(buf, size); - (void) write(2, buf, size); + (void) write(1, buf, size); } static void @@ -2042,31 +2635,30 @@ name: * flags - A string of characters specifying options * b: Decode a blkptr at given offset within block * *c: Calculate and display checksums - * *d: Decompress data before dumping + * d: Decompress data before dumping * e: Byteswap data before dumping - * *g: Display data as a gang block header - * *i: Display as an indirect block + * g: Display data as a gang block header + * i: Display as an indirect block * p: Do I/O to physical offset * r: Dump raw data to stdout * * * = not yet implemented */ static void -zdb_read_block(char *thing, spa_t **spap) +zdb_read_block(char *thing, spa_t *spa) { - spa_t *spa = *spap; + blkptr_t blk, *bp = &blk; + dva_t *dva = bp->blk_dva; int flags = 0; - uint64_t offset = 0, size = 0, blkptr_offset = 0; + uint64_t offset = 0, size = 0, psize = 0, lsize = 0, blkptr_offset = 0; zio_t *zio; vdev_t *vd; - void *buf; - char *s, *p, *dup, *pool, *vdev, *flagstr; - int i, error, zio_flags; + void *pbuf, *lbuf, *buf; + char *s, *p, *dup, *vdev, *flagstr; + int i, error; dup = strdup(thing); s = strtok(dup, ":"); - pool = s ? s : ""; - s = strtok(NULL, ":"); vdev = s ? s : ""; s = strtok(NULL, ":"); offset = strtoull(s ? s : "", NULL, 16); @@ -2100,7 +2692,7 @@ zdb_read_block(char *thing, spa_t **spap) flags |= bit; /* If it's not something with an argument, keep going */ - if ((bit & (ZDB_FLAG_CHECKSUM | ZDB_FLAG_DECOMPRESS | + if ((bit & (ZDB_FLAG_CHECKSUM | ZDB_FLAG_PRINT_BLKPTR)) == 0) continue; @@ -2115,16 +2707,6 @@ zdb_read_block(char *thing, spa_t **spap) } } - if (spa == NULL || strcmp(spa_name(spa), pool) != 0) { - if (spa) - spa_close(spa, (void *)zdb_read_block); - error = spa_open(pool, spap, (void *)zdb_read_block); - if (error) - fatal("Failed to open pool '%s': %s", - pool, strerror(error)); - spa = *spap; - } - vd = zdb_vdev_lookup(spa->spa_root_vdev, vdev); if (vd == NULL) { (void) printf("***Invalid vdev: %s\n", vdev); @@ -2132,22 +2714,58 @@ zdb_read_block(char *thing, spa_t **spap) return; } else { if (vd->vdev_path) - (void) printf("Found vdev: %s\n", vd->vdev_path); + (void) fprintf(stderr, "Found vdev: %s\n", + vd->vdev_path); else - (void) printf("Found vdev type: %s\n", + (void) fprintf(stderr, "Found vdev type: %s\n", vd->vdev_ops->vdev_op_type); } - buf = umem_alloc(size, UMEM_NOFAIL); + psize = size; + lsize = size; - zio_flags = ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE | - ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY; + pbuf = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL); + lbuf = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL); + + BP_ZERO(bp); + + DVA_SET_VDEV(&dva[0], vd->vdev_id); + DVA_SET_OFFSET(&dva[0], offset); + DVA_SET_GANG(&dva[0], !!(flags & ZDB_FLAG_GBH)); + DVA_SET_ASIZE(&dva[0], vdev_psize_to_asize(vd, psize)); + + BP_SET_BIRTH(bp, TXG_INITIAL, TXG_INITIAL); + + BP_SET_LSIZE(bp, lsize); + BP_SET_PSIZE(bp, psize); + BP_SET_COMPRESS(bp, ZIO_COMPRESS_OFF); + BP_SET_CHECKSUM(bp, ZIO_CHECKSUM_OFF); + BP_SET_TYPE(bp, DMU_OT_NONE); + BP_SET_LEVEL(bp, 0); + BP_SET_DEDUP(bp, 0); + BP_SET_BYTEORDER(bp, ZFS_HOST_BYTEORDER); spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); zio = zio_root(spa, NULL, NULL, 0); - /* XXX todo - cons up a BP so RAID-Z will be happy */ - zio_nowait(zio_vdev_child_io(zio, NULL, vd, offset, buf, size, - ZIO_TYPE_READ, ZIO_PRIORITY_SYNC_READ, zio_flags, NULL, NULL)); + + if (vd == vd->vdev_top) { + /* + * Treat this as a normal block read. + */ + zio_nowait(zio_read(zio, spa, bp, pbuf, psize, NULL, NULL, + ZIO_PRIORITY_SYNC_READ, + ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW, NULL)); + } else { + /* + * Treat this as a vdev child I/O. + */ + zio_nowait(zio_vdev_child_io(zio, bp, vd, offset, pbuf, psize, + ZIO_TYPE_READ, ZIO_PRIORITY_SYNC_READ, + ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE | + ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY | + ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW, NULL, NULL)); + } + error = zio_wait(zio); spa_config_exit(spa, SCL_STATE, FTAG); @@ -2156,6 +2774,52 @@ zdb_read_block(char *thing, spa_t **spap) goto out; } + if (flags & ZDB_FLAG_DECOMPRESS) { + /* + * We don't know how the data was compressed, so just try + * every decompress function at every inflated blocksize. + */ + enum zio_compress c; + void *pbuf2 = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL); + void *lbuf2 = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL); + + bcopy(pbuf, pbuf2, psize); + + VERIFY(random_get_pseudo_bytes((uint8_t *)pbuf + psize, + SPA_MAXBLOCKSIZE - psize) == 0); + + VERIFY(random_get_pseudo_bytes((uint8_t *)pbuf2 + psize, + SPA_MAXBLOCKSIZE - psize) == 0); + + for (lsize = SPA_MAXBLOCKSIZE; lsize > psize; + lsize -= SPA_MINBLOCKSIZE) { + for (c = 0; c < ZIO_COMPRESS_FUNCTIONS; c++) { + if (zio_decompress_data(c, pbuf, lbuf, + psize, lsize) == 0 && + zio_decompress_data(c, pbuf2, lbuf2, + psize, lsize) == 0 && + bcmp(lbuf, lbuf2, lsize) == 0) + break; + } + if (c != ZIO_COMPRESS_FUNCTIONS) + break; + lsize -= SPA_MINBLOCKSIZE; + } + + umem_free(pbuf2, SPA_MAXBLOCKSIZE); + umem_free(lbuf2, SPA_MAXBLOCKSIZE); + + if (lsize <= psize) { + (void) printf("Decompress of %s failed\n", thing); + goto out; + } + buf = lbuf; + size = lsize; + } else { + buf = pbuf; + size = psize; + } + if (flags & ZDB_FLAG_PRINT_BLKPTR) zdb_print_blkptr((blkptr_t *)(void *) ((uintptr_t)buf + (uintptr_t)blkptr_offset), flags); @@ -2170,134 +2834,92 @@ zdb_read_block(char *thing, spa_t **spap) zdb_dump_block(thing, buf, size, flags); out: - umem_free(buf, size); + umem_free(pbuf, SPA_MAXBLOCKSIZE); + umem_free(lbuf, SPA_MAXBLOCKSIZE); free(dup); } static boolean_t -nvlist_string_match(nvlist_t *config, char *name, char *tgt) +pool_match(nvlist_t *cfg, char *tgt) { + uint64_t v, guid = strtoull(tgt, NULL, 0); char *s; - if (nvlist_lookup_string(config, name, &s) != 0) - return (B_FALSE); - - return (strcmp(s, tgt) == 0); -} - -static boolean_t -nvlist_uint64_match(nvlist_t *config, char *name, uint64_t tgt) -{ - uint64_t val; - - if (nvlist_lookup_uint64(config, name, &val) != 0) - return (B_FALSE); - - return (val == tgt); -} - -static boolean_t -vdev_child_guid_match(nvlist_t *vdev, uint64_t guid) -{ - nvlist_t **child; - uint_t c, children; - - verify(nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, - &child, &children) == 0); - for (c = 0; c < children; ++c) - if (nvlist_uint64_match(child[c], ZPOOL_CONFIG_GUID, guid)) - return (B_TRUE); - return (B_FALSE); -} - -static boolean_t -vdev_child_string_match(nvlist_t *vdev, char *tgt) -{ - nvlist_t **child; - uint_t c, children; - - verify(nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, - &child, &children) == 0); - for (c = 0; c < children; ++c) { - if (nvlist_string_match(child[c], ZPOOL_CONFIG_PATH, tgt) || - nvlist_string_match(child[c], ZPOOL_CONFIG_DEVID, tgt)) - return (B_TRUE); - } - return (B_FALSE); -} - -static boolean_t -vdev_guid_match(nvlist_t *config, uint64_t guid) -{ - nvlist_t *nvroot; - - verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, - &nvroot) == 0); - - return (nvlist_uint64_match(nvroot, ZPOOL_CONFIG_GUID, guid) || - vdev_child_guid_match(nvroot, guid)); -} - -static boolean_t -vdev_string_match(nvlist_t *config, char *tgt) -{ - nvlist_t *nvroot; - - verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, - &nvroot) == 0); - - return (vdev_child_string_match(nvroot, tgt)); -} - -static boolean_t -pool_match(nvlist_t *config, char *tgt) -{ - uint64_t guid = strtoull(tgt, NULL, 0); - if (guid != 0) { - return ( - nvlist_uint64_match(config, ZPOOL_CONFIG_POOL_GUID, guid) || - vdev_guid_match(config, guid)); + if (nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_POOL_GUID, &v) == 0) + return (v == guid); } else { - return ( - nvlist_string_match(config, ZPOOL_CONFIG_POOL_NAME, tgt) || - vdev_string_match(config, tgt)); + if (nvlist_lookup_string(cfg, ZPOOL_CONFIG_POOL_NAME, &s) == 0) + return (strcmp(s, tgt) == 0); } + return (B_FALSE); } -static int -find_exported_zpool(char *pool_id, nvlist_t **configp, char *vdev_dir) +static char * +find_zpool(char **target, nvlist_t **configp, int dirc, char **dirv) { nvlist_t *pools; - int error = ENOENT; nvlist_t *match = NULL; + char *name = NULL; + char *sepp = NULL; + char sep; + int count = 0; + importargs_t args = { 0 }; - if (vdev_dir != NULL) - pools = zpool_find_import_activeok(g_zfs, 1, &vdev_dir); - else - pools = zpool_find_import_activeok(g_zfs, 0, NULL); + args.paths = dirc; + args.path = dirv; + args.can_be_active = B_TRUE; + + if ((sepp = strpbrk(*target, "/@")) != NULL) { + sep = *sepp; + *sepp = '\0'; + } + + pools = zpool_search_import(g_zfs, &args); if (pools != NULL) { nvpair_t *elem = NULL; - while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) { verify(nvpair_value_nvlist(elem, configp) == 0); - if (pool_match(*configp, pool_id)) { + if (pool_match(*configp, *target)) { + count++; if (match != NULL) { - (void) fatal( - "More than one matching pool - " - "specify guid/devid/device path."); + /* print previously found config */ + if (name != NULL) { + (void) printf("%s\n", name); + dump_nvlist(match, 8); + name = NULL; + } + (void) printf("%s\n", + nvpair_name(elem)); + dump_nvlist(*configp, 8); } else { match = *configp; - error = 0; + name = nvpair_name(elem); } } } } + if (count > 1) + (void) fatal("\tMatched %d pools - use pool GUID " + "instead of pool name or \n" + "\tpool name part of a dataset name to select pool", count); - *configp = error ? NULL : match; + if (sepp) + *sepp = sep; + /* + * If pool GUID was specified for pool id, replace it with pool name + */ + if (name && (strstr(*target, name) != *target)) { + int sz = 1 + strlen(name) + ((sepp) ? strlen(sepp) : 0); - return (error); + *target = umem_alloc(sz, UMEM_NOFAIL); + (void) snprintf(*target, sz, "%s%s", name, sepp ? sepp : ""); + } + + *configp = name ? match : NULL; + + return (name); } int @@ -2305,83 +2927,85 @@ main(int argc, char **argv) { int i, c; struct rlimit rl = { 1024, 1024 }; - spa_t *spa; + spa_t *spa = NULL; objset_t *os = NULL; - char *endstr; int dump_all = 1; int verbose = 0; - int error; - int exported = 0; - char *vdev_dir = NULL; + int error = 0; + char **searchdirs = NULL; + int nsearch = 0; + char *target; + nvlist_t *policy = NULL; + uint64_t max_txg = UINT64_MAX; + int rewind = ZPOOL_NEVER_REWIND; (void) setrlimit(RLIMIT_NOFILE, &rl); (void) enable_extended_FILE_stdio(-1, -1); dprintf_setup(&argc, argv); - while ((c = getopt(argc, argv, "udibcmsvCLS:U:lRep:t:")) != -1) { + while ((c = getopt(argc, argv, "bcdhilmsuCDRSAFLXevp:t:U:P")) != -1) { switch (c) { - case 'u': - case 'd': - case 'i': case 'b': case 'c': + case 'd': + case 'h': + case 'i': + case 'l': case 'm': case 's': + case 'u': case 'C': - case 'l': + case 'D': case 'R': + case 'S': dump_opt[c]++; dump_all = 0; break; + case 'A': + case 'F': case 'L': + case 'X': + case 'e': + case 'P': dump_opt[c]++; break; case 'v': verbose++; break; - case 'U': - spa_config_path = optarg; - break; - case 'e': - exported = 1; - break; case 'p': - vdev_dir = optarg; - break; - case 'S': - dump_opt[c]++; - dump_all = 0; - zdb_sig_user_data = (strncmp(optarg, "user:", 5) == 0); - if (!zdb_sig_user_data && strncmp(optarg, "all:", 4)) - usage(); - endstr = strchr(optarg, ':') + 1; - if (strcmp(endstr, "fletcher2") == 0) - zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_2; - else if (strcmp(endstr, "fletcher4") == 0) - zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_4; - else if (strcmp(endstr, "sha256") == 0) - zdb_sig_cksumalg = ZIO_CHECKSUM_SHA256; - else if (strcmp(endstr, "all") == 0) - zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_2; - else - usage(); + if (searchdirs == NULL) { + searchdirs = umem_alloc(sizeof (char *), + UMEM_NOFAIL); + } else { + char **tmp = umem_alloc((nsearch + 1) * + sizeof (char *), UMEM_NOFAIL); + bcopy(searchdirs, tmp, nsearch * + sizeof (char *)); + umem_free(searchdirs, + nsearch * sizeof (char *)); + searchdirs = tmp; + } + searchdirs[nsearch++] = optarg; break; case 't': - ub_max_txg = strtoull(optarg, NULL, 0); - if (ub_max_txg < TXG_INITIAL) { + max_txg = strtoull(optarg, NULL, 0); + if (max_txg < TXG_INITIAL) { (void) fprintf(stderr, "incorrect txg " "specified: %s\n", optarg); usage(); } break; + case 'U': + spa_config_path = optarg; + break; default: usage(); break; } } - if (vdev_dir != NULL && exported == 0) { + if (!dump_opt['e'] && searchdirs != NULL) { (void) fprintf(stderr, "-p option requires use of -e\n"); usage(); } @@ -2390,18 +3014,26 @@ main(int argc, char **argv) g_zfs = libzfs_init(); ASSERT(g_zfs != NULL); + if (dump_all) + verbose = MAX(verbose, 1); + for (c = 0; c < 256; c++) { - if (dump_all && c != 'l' && c != 'R') + if (dump_all && !strchr("elAFLRSXP", c)) dump_opt[c] = 1; if (dump_opt[c]) dump_opt[c] += verbose; } + aok = (dump_opt['A'] == 1) || (dump_opt['A'] > 2); + zfs_recover = (dump_opt['A'] > 1); + argc -= optind; argv += optind; + if (argc < 2 && dump_opt['R']) + usage(); if (argc < 1) { - if (dump_opt['C']) { + if (!dump_opt['e'] && dump_opt['C']) { dump_cachefile(spa_config_path); return (0); } @@ -2413,99 +3045,104 @@ main(int argc, char **argv) return (0); } - if (dump_opt['R']) { - flagbits['b'] = ZDB_FLAG_PRINT_BLKPTR; - flagbits['c'] = ZDB_FLAG_CHECKSUM; - flagbits['d'] = ZDB_FLAG_DECOMPRESS; - flagbits['e'] = ZDB_FLAG_BSWAP; - flagbits['g'] = ZDB_FLAG_GBH; - flagbits['i'] = ZDB_FLAG_INDIRECT; - flagbits['p'] = ZDB_FLAG_PHYS; - flagbits['r'] = ZDB_FLAG_RAW; - - spa = NULL; - while (argv[0]) { - zdb_read_block(argv[0], &spa); - argv++; - argc--; - } - if (spa) - spa_close(spa, (void *)zdb_read_block); - return (0); - } + if (dump_opt['X'] || dump_opt['F']) + rewind = ZPOOL_DO_REWIND | + (dump_opt['X'] ? ZPOOL_EXTREME_REWIND : 0); - if (dump_opt['C']) - dump_config(argv[0]); + if (nvlist_alloc(&policy, NV_UNIQUE_NAME_TYPE, 0) != 0 || + nvlist_add_uint64(policy, ZPOOL_REWIND_REQUEST_TXG, max_txg) != 0 || + nvlist_add_uint32(policy, ZPOOL_REWIND_REQUEST, rewind) != 0) + fatal("internal error: %s", strerror(ENOMEM)); error = 0; - if (exported) { - /* - * Check to see if the name refers to an exported zpool - */ - char *slash; - nvlist_t *exported_conf = NULL; - - if ((slash = strchr(argv[0], '/')) != NULL) - *slash = '\0'; - - error = find_exported_zpool(argv[0], &exported_conf, vdev_dir); - if (error == 0) { - nvlist_t *nvl = NULL; - - if (vdev_dir != NULL) { - if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) - error = ENOMEM; - else if (nvlist_add_string(nvl, - zpool_prop_to_name(ZPOOL_PROP_ALTROOT), - vdev_dir) != 0) - error = ENOMEM; - } + target = argv[0]; - if (error == 0) - error = spa_import_verbatim(argv[0], - exported_conf, nvl); + if (dump_opt['e']) { + nvlist_t *cfg = NULL; + char *name = find_zpool(&target, &cfg, nsearch, searchdirs); - nvlist_free(nvl); + error = ENOENT; + if (name) { + if (dump_opt['C'] > 1) { + (void) printf("\nConfiguration for import:\n"); + dump_nvlist(cfg, 8); + } + if (nvlist_add_nvlist(cfg, + ZPOOL_REWIND_POLICY, policy) != 0) { + fatal("can't open '%s': %s", + target, strerror(ENOMEM)); + } + if ((error = spa_import(name, cfg, NULL)) != 0) + error = spa_import_verbatim(name, cfg, NULL); } - - if (slash != NULL) - *slash = '/'; } if (error == 0) { - if (strchr(argv[0], '/') != NULL) { - error = dmu_objset_open(argv[0], DMU_OST_ANY, - DS_MODE_USER | DS_MODE_READONLY, &os); + if (strpbrk(target, "/@") == NULL || dump_opt['R']) { + error = spa_open_rewind(target, &spa, FTAG, policy, + NULL); + if (error) { + /* + * If we're missing the log device then + * try opening the pool after clearing the + * log state. + */ + mutex_enter(&spa_namespace_lock); + if ((spa = spa_lookup(target)) != NULL && + spa->spa_log_state == SPA_LOG_MISSING) { + spa->spa_log_state = SPA_LOG_CLEAR; + error = 0; + } + mutex_exit(&spa_namespace_lock); + + if (!error) { + error = spa_open_rewind(target, &spa, + FTAG, policy, NULL); + } + } } else { - error = spa_open(argv[0], &spa, FTAG); + error = dmu_objset_own(target, DMU_OST_ANY, + B_TRUE, FTAG, &os); } } + nvlist_free(policy); if (error) - fatal("can't open %s: %s", argv[0], strerror(error)); + fatal("can't open '%s': %s", target, strerror(error)); argv++; - if (--argc > 0) { - zopt_objects = argc; - zopt_object = calloc(zopt_objects, sizeof (uint64_t)); - for (i = 0; i < zopt_objects; i++) { - errno = 0; - zopt_object[i] = strtoull(argv[i], NULL, 0); - if (zopt_object[i] == 0 && errno != 0) - fatal("bad object number %s: %s", - argv[i], strerror(errno)); + argc--; + if (!dump_opt['R']) { + if (argc > 0) { + zopt_objects = argc; + zopt_object = calloc(zopt_objects, sizeof (uint64_t)); + for (i = 0; i < zopt_objects; i++) { + errno = 0; + zopt_object[i] = strtoull(argv[i], NULL, 0); + if (zopt_object[i] == 0 && errno != 0) + fatal("bad number %s: %s", + argv[i], strerror(errno)); + } } - } - - if (os != NULL) { - dump_dir(os); - dmu_objset_close(os); + (os != NULL) ? dump_dir(os) : dump_zpool(spa); } else { - dump_zpool(spa); - spa_close(spa, FTAG); + flagbits['b'] = ZDB_FLAG_PRINT_BLKPTR; + flagbits['c'] = ZDB_FLAG_CHECKSUM; + flagbits['d'] = ZDB_FLAG_DECOMPRESS; + flagbits['e'] = ZDB_FLAG_BSWAP; + flagbits['g'] = ZDB_FLAG_GBH; + flagbits['i'] = ZDB_FLAG_INDIRECT; + flagbits['p'] = ZDB_FLAG_PHYS; + flagbits['r'] = ZDB_FLAG_RAW; + + for (i = 0; i < argc; i++) + zdb_read_block(argv[i], spa); } + (os != NULL) ? dmu_objset_disown(os, FTAG) : spa_close(spa, FTAG); + fuid_table_destroy(); + sa_loaded = B_FALSE; libzfs_fini(g_zfs); kernel_fini(); diff --git a/cmd/zdb/zdb_il.c b/cmd/zdb/zdb_il.c index cc08ef514..a0ed985f5 100644 --- a/cmd/zdb/zdb_il.c +++ b/cmd/zdb/zdb_il.c @@ -40,12 +40,14 @@ extern uint8_t dump_opt[256]; +static char prefix[4] = "\t\t\t"; + static void print_log_bp(const blkptr_t *bp, const char *prefix) { char blkbuf[BP_SPRINTF_LEN]; - sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, bp); + sprintf_blkptr(blkbuf, bp); (void) printf("%s%s\n", prefix, blkbuf); } @@ -54,19 +56,29 @@ static void zil_prt_rec_create(zilog_t *zilog, int txtype, lr_create_t *lr) { time_t crtime = lr->lr_crtime[0]; - char *name = (char *)(lr + 1); - char *link = name + strlen(name) + 1; + char *name, *link; + lr_attr_t *lrattr; - if (txtype == TX_SYMLINK) - (void) printf("\t\t\t%s -> %s\n", name, link); - else - (void) printf("\t\t\t%s\n", name); + name = (char *)(lr + 1); + + if (lr->lr_common.lrc_txtype == TX_CREATE_ATTR || + lr->lr_common.lrc_txtype == TX_MKDIR_ATTR) { + lrattr = (lr_attr_t *)(lr + 1); + name += ZIL_XVAT_SIZE(lrattr->lr_attr_masksize); + } + + if (txtype == TX_SYMLINK) { + link = name + strlen(name) + 1; + (void) printf("%s%s -> %s\n", prefix, name, link); + } else if (txtype != TX_MKXATTR) { + (void) printf("%s%s\n", prefix, name); + } - (void) printf("\t\t\t%s", ctime(&crtime)); - (void) printf("\t\t\tdoid %llu, foid %llu, mode %llo\n", + (void) printf("%s%s", prefix, ctime(&crtime)); + (void) printf("%sdoid %llu, foid %llu, mode %llo\n", prefix, (u_longlong_t)lr->lr_doid, (u_longlong_t)lr->lr_foid, (longlong_t)lr->lr_mode); - (void) printf("\t\t\tuid %llu, gid %llu, gen %llu, rdev 0x%llx\n", + (void) printf("%suid %llu, gid %llu, gen %llu, rdev 0x%llx\n", prefix, (u_longlong_t)lr->lr_uid, (u_longlong_t)lr->lr_gid, (u_longlong_t)lr->lr_gen, (u_longlong_t)lr->lr_rdev); } @@ -75,7 +87,7 @@ zil_prt_rec_create(zilog_t *zilog, int txtype, lr_create_t *lr) static void zil_prt_rec_remove(zilog_t *zilog, int txtype, lr_remove_t *lr) { - (void) printf("\t\t\tdoid %llu, name %s\n", + (void) printf("%sdoid %llu, name %s\n", prefix, (u_longlong_t)lr->lr_doid, (char *)(lr + 1)); } @@ -83,7 +95,7 @@ zil_prt_rec_remove(zilog_t *zilog, int txtype, lr_remove_t *lr) static void zil_prt_rec_link(zilog_t *zilog, int txtype, lr_link_t *lr) { - (void) printf("\t\t\tdoid %llu, link_obj %llu, name %s\n", + (void) printf("%sdoid %llu, link_obj %llu, name %s\n", prefix, (u_longlong_t)lr->lr_doid, (u_longlong_t)lr->lr_link_obj, (char *)(lr + 1)); } @@ -95,9 +107,9 @@ zil_prt_rec_rename(zilog_t *zilog, int txtype, lr_rename_t *lr) char *snm = (char *)(lr + 1); char *tnm = snm + strlen(snm) + 1; - (void) printf("\t\t\tsdoid %llu, tdoid %llu\n", + (void) printf("%ssdoid %llu, tdoid %llu\n", prefix, (u_longlong_t)lr->lr_sdoid, (u_longlong_t)lr->lr_tdoid); - (void) printf("\t\t\tsrc %s tgt %s\n", snm, tnm); + (void) printf("%ssrc %s tgt %s\n", prefix, snm, tnm); } /* ARGSUSED */ @@ -106,43 +118,48 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, lr_write_t *lr) { char *data, *dlimit; blkptr_t *bp = &lr->lr_blkptr; + zbookmark_t zb; char buf[SPA_MAXBLOCKSIZE]; int verbose = MAX(dump_opt['d'], dump_opt['i']); int error; - (void) printf("\t\t\tfoid %llu, offset 0x%llx," - " length 0x%llx, blkoff 0x%llx\n", - (u_longlong_t)lr->lr_foid, (longlong_t)lr->lr_offset, - (u_longlong_t)lr->lr_length, (u_longlong_t)lr->lr_blkoff); + (void) printf("%sfoid %llu, offset %llx, length %llx\n", prefix, + (u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_offset, + (u_longlong_t)lr->lr_length); - if (verbose < 5) + if (txtype == TX_WRITE2 || verbose < 5) return; if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) { - (void) printf("\t\t\thas blkptr, %s\n", + (void) printf("%shas blkptr, %s\n", prefix, bp->blk_birth >= spa_first_txg(zilog->zl_spa) ? "will claim" : "won't claim"); - print_log_bp(bp, "\t\t\t"); + print_log_bp(bp, prefix); + + if (BP_IS_HOLE(bp)) { + (void) printf("\t\t\tLSIZE 0x%llx\n", + (u_longlong_t)BP_GET_LSIZE(bp)); + } if (bp->blk_birth == 0) { bzero(buf, sizeof (buf)); - } else { - zbookmark_t zb; - - ASSERT3U(bp->blk_cksum.zc_word[ZIL_ZC_OBJSET], ==, - dmu_objset_id(zilog->zl_os)); - - zb.zb_objset = bp->blk_cksum.zc_word[ZIL_ZC_OBJSET]; - zb.zb_object = 0; - zb.zb_level = -1; - zb.zb_blkid = bp->blk_cksum.zc_word[ZIL_ZC_SEQ]; - - error = zio_wait(zio_read(NULL, zilog->zl_spa, - bp, buf, BP_GET_LSIZE(bp), NULL, NULL, - ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL, &zb)); - if (error) - return; + (void) printf("%s<hole>\n", prefix); + return; } - data = buf + lr->lr_blkoff; + if (bp->blk_birth < zilog->zl_header->zh_claim_txg) { + (void) printf("%s<block already committed>\n", prefix); + return; + } + + SET_BOOKMARK(&zb, dmu_objset_id(zilog->zl_os), + lr->lr_foid, ZB_ZIL_LEVEL, + lr->lr_offset / BP_GET_LSIZE(bp)); + + error = zio_wait(zio_read(NULL, zilog->zl_spa, + bp, buf, BP_GET_LSIZE(bp), NULL, NULL, + ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL, &zb)); + if (error) + return; + data = buf; } else { data = (char *)(lr + 1); } @@ -150,7 +167,7 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, lr_write_t *lr) dlimit = data + MIN(lr->lr_length, (verbose < 6 ? 20 : SPA_MAXBLOCKSIZE)); - (void) printf("\t\t\t"); + (void) printf("%s", prefix); while (data < dlimit) { if (isprint(*data)) (void) printf("%c ", *data); @@ -165,7 +182,7 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, lr_write_t *lr) static void zil_prt_rec_truncate(zilog_t *zilog, int txtype, lr_truncate_t *lr) { - (void) printf("\t\t\tfoid %llu, offset 0x%llx, length 0x%llx\n", + (void) printf("%sfoid %llu, offset 0x%llx, length 0x%llx\n", prefix, (u_longlong_t)lr->lr_foid, (longlong_t)lr->lr_offset, (u_longlong_t)lr->lr_length); } @@ -177,38 +194,38 @@ zil_prt_rec_setattr(zilog_t *zilog, int txtype, lr_setattr_t *lr) time_t atime = (time_t)lr->lr_atime[0]; time_t mtime = (time_t)lr->lr_mtime[0]; - (void) printf("\t\t\tfoid %llu, mask 0x%llx\n", + (void) printf("%sfoid %llu, mask 0x%llx\n", prefix, (u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_mask); if (lr->lr_mask & AT_MODE) { - (void) printf("\t\t\tAT_MODE %llo\n", + (void) printf("%sAT_MODE %llo\n", prefix, (longlong_t)lr->lr_mode); } if (lr->lr_mask & AT_UID) { - (void) printf("\t\t\tAT_UID %llu\n", + (void) printf("%sAT_UID %llu\n", prefix, (u_longlong_t)lr->lr_uid); } if (lr->lr_mask & AT_GID) { - (void) printf("\t\t\tAT_GID %llu\n", + (void) printf("%sAT_GID %llu\n", prefix, (u_longlong_t)lr->lr_gid); } if (lr->lr_mask & AT_SIZE) { - (void) printf("\t\t\tAT_SIZE %llu\n", + (void) printf("%sAT_SIZE %llu\n", prefix, (u_longlong_t)lr->lr_size); } if (lr->lr_mask & AT_ATIME) { - (void) printf("\t\t\tAT_ATIME %llu.%09llu %s", + (void) printf("%sAT_ATIME %llu.%09llu %s", prefix, (u_longlong_t)lr->lr_atime[0], (u_longlong_t)lr->lr_atime[1], ctime(&atime)); } if (lr->lr_mask & AT_MTIME) { - (void) printf("\t\t\tAT_MTIME %llu.%09llu %s", + (void) printf("%sAT_MTIME %llu.%09llu %s", prefix, (u_longlong_t)lr->lr_mtime[0], (u_longlong_t)lr->lr_mtime[1], ctime(&mtime)); @@ -219,7 +236,7 @@ zil_prt_rec_setattr(zilog_t *zilog, int txtype, lr_setattr_t *lr) static void zil_prt_rec_acl(zilog_t *zilog, int txtype, lr_acl_t *lr) { - (void) printf("\t\t\tfoid %llu, aclcnt %llu\n", + (void) printf("%sfoid %llu, aclcnt %llu\n", prefix, (u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_aclcnt); } @@ -251,10 +268,11 @@ static zil_rec_info_t zil_rec_info[TX_MAX_TYPE] = { { zil_prt_rec_create, "TX_MKDIR_ACL " }, { zil_prt_rec_create, "TX_MKDIR_ATTR " }, { zil_prt_rec_create, "TX_MKDIR_ACL_ATTR " }, + { zil_prt_rec_write, "TX_WRITE2 " }, }; /* ARGSUSED */ -static void +static int print_log_record(zilog_t *zilog, lr_t *lr, void *arg, uint64_t claim_txg) { int txtype; @@ -278,23 +296,24 @@ print_log_record(zilog_t *zilog, lr_t *lr, void *arg, uint64_t claim_txg) zil_rec_info[txtype].zri_count++; zil_rec_info[0].zri_count++; + + return (0); } /* ARGSUSED */ -static void +static int print_log_block(zilog_t *zilog, blkptr_t *bp, void *arg, uint64_t claim_txg) { - char blkbuf[BP_SPRINTF_LEN]; + char blkbuf[BP_SPRINTF_LEN + 10]; int verbose = MAX(dump_opt['d'], dump_opt['i']); char *claim; if (verbose <= 3) - return; + return (0); if (verbose >= 5) { (void) strcpy(blkbuf, ", "); - sprintf_blkptr(blkbuf + strlen(blkbuf), - BP_SPRINTF_LEN - strlen(blkbuf), bp); + sprintf_blkptr(blkbuf + strlen(blkbuf), bp); } else { blkbuf[0] = '\0'; } @@ -308,6 +327,8 @@ print_log_block(zilog_t *zilog, blkptr_t *bp, void *arg, uint64_t claim_txg) (void) printf("\tBlock seqno %llu, %s%s\n", (u_longlong_t)bp->blk_cksum.zc_word[ZIL_ZC_SEQ], claim, blkbuf); + + return (0); } static void @@ -340,17 +361,17 @@ dump_intent_log(zilog_t *zilog) int verbose = MAX(dump_opt['d'], dump_opt['i']); int i; - if (zh->zh_log.blk_birth == 0 || verbose < 2) + if (zh->zh_log.blk_birth == 0 || verbose < 1) return; - (void) printf("\n ZIL header: claim_txg %llu, claim_seq %llu", - (u_longlong_t)zh->zh_claim_txg, (u_longlong_t)zh->zh_claim_seq); + (void) printf("\n ZIL header: claim_txg %llu, " + "claim_blk_seq %llu, claim_lr_seq %llu", + (u_longlong_t)zh->zh_claim_txg, + (u_longlong_t)zh->zh_claim_blk_seq, + (u_longlong_t)zh->zh_claim_lr_seq); (void) printf(" replay_seq %llu, flags 0x%llx\n", (u_longlong_t)zh->zh_replay_seq, (u_longlong_t)zh->zh_flags); - if (verbose >= 4) - print_log_bp(&zh->zh_log, "\n\tfirst block: "); - for (i = 0; i < TX_MAX_TYPE; i++) zil_rec_info[i].zri_count = 0; |