modules/zfs/zfs.c

	zfs.c revision 3f1f80124f2b2b91c4c06303305e5badae5228e8
/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright 2010 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */

#include <mdb/mdb_ctf.h>
#include <sys/zfs_context.h>
#include <sys/mdb_modapi.h>
#include <sys/dbuf.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_pool.h>
#include <sys/metaslab_impl.h>
#include <sys/space_map.h>
#include <sys/list.h>
#include <sys/spa_impl.h>
#include <sys/vdev_impl.h>
#include <sys/zap_leaf.h>
#include <sys/zap_impl.h>
#include <ctype.h>

#ifndef _KERNEL
#include "../genunix/list.h"
#endif

#ifdef _KERNEL
#define ZFS_OBJ_NAME    "zfs"
#else
#define ZFS_OBJ_NAME    "libzpool.so.1"
#endif

static int
getmember(uintptr_t addr, const char *type, mdb_ctf_id_t *idp,
    const char *member, int len, void *buf)
{
    mdb_ctf_id_t id;
    ulong_t off;
    char name[64];

    if (idp == NULL) {
        if (mdb_ctf_lookup_by_name(type, &id) == -1) {
            mdb_warn("couldn't find type %s", type);
            return (DCMD_ERR);
        }
        idp = &id;
    } else {
        type = name;
        mdb_ctf_type_name(*idp, name, sizeof (name));
    }

    if (mdb_ctf_offsetof(*idp, member, &off) == -1) {
        mdb_warn("couldn't find member %s of type %s\n", member, type);
        return (DCMD_ERR);
    }
    if (off % 8 != 0) {
        mdb_warn("member %s of type %s is unsupported bitfield",
            member, type);
        return (DCMD_ERR);
    }
    off /= 8;

    if (mdb_vread(buf, len, addr + off) == -1) {
        mdb_warn("failed to read %s from %s at %p",
            member, type, addr + off);
        return (DCMD_ERR);
    }
    /* mdb_warn("read %s from %s at %p+%llx\n", member, type, addr, off); */

    return (0);
}

#define GETMEMB(addr, type, member, dest) \
    getmember(addr, #type, NULL, #member, sizeof (dest), &(dest))

#define GETMEMBID(addr, ctfid, member, dest) \
    getmember(addr, NULL, ctfid, #member, sizeof (dest), &(dest))

static int
getrefcount(uintptr_t addr, mdb_ctf_id_t *id,
    const char *member, uint64_t *rc)
{
    static int gotid;
    static mdb_ctf_id_t rc_id;
    ulong_t off;

    if (!gotid) {
        if (mdb_ctf_lookup_by_name("struct refcount", &rc_id) == -1) {
            mdb_warn("couldn't find struct refcount");
            return (DCMD_ERR);
        }
        gotid = TRUE;
    }

    if (mdb_ctf_offsetof(*id, member, &off) == -1) {
        char name[64];
        mdb_ctf_type_name(*id, name, sizeof (name));
        mdb_warn("couldn't find member %s of type %s\n", member, name);
        return (DCMD_ERR);
    }
    off /= 8;

    return (GETMEMBID(addr + off, &rc_id, rc_count, *rc));
}

static int verbose;

static int
freelist_walk_init(mdb_walk_state_t *wsp)
{
    if (wsp->walk_addr == NULL) {
        mdb_warn("must supply starting address\n");
        return (WALK_ERR);
    }

    wsp->walk_data = 0;  /* Index into the freelist */
    return (WALK_NEXT);
}

static int
freelist_walk_step(mdb_walk_state_t *wsp)
{
    uint64_t entry;
    uintptr_t number = (uintptr_t)wsp->walk_data;
    char *ddata[] = { "ALLOC", "FREE", "CONDENSE", "INVALID",
                "INVALID", "INVALID", "INVALID", "INVALID" };
    int mapshift = SPA_MINBLOCKSHIFT;

    if (mdb_vread(&entry, sizeof (entry), wsp->walk_addr) == -1) {
        mdb_warn("failed to read freelist entry %p", wsp->walk_addr);
        return (WALK_DONE);
    }
    wsp->walk_addr += sizeof (entry);
    wsp->walk_data = (void *)(number + 1);

    if (SM_DEBUG_DECODE(entry)) {
        mdb_printf("DEBUG: %3u  %10s: txg=%llu  pass=%llu\n",
            number,
            ddata[SM_DEBUG_ACTION_DECODE(entry)],
            SM_DEBUG_TXG_DECODE(entry),
            SM_DEBUG_SYNCPASS_DECODE(entry));
    } else {
        mdb_printf("Entry: %3u  offsets=%08llx-%08llx  type=%c  "
            "size=%06llx", number,
            SM_OFFSET_DECODE(entry) << mapshift,
            (SM_OFFSET_DECODE(entry) + SM_RUN_DECODE(entry)) <<
            mapshift,
            SM_TYPE_DECODE(entry) == SM_ALLOC ? 'A' : 'F',
            SM_RUN_DECODE(entry) << mapshift);
        if (verbose)
            mdb_printf("      (raw=%012llx)\n", entry);
        mdb_printf("\n");
    }
    return (WALK_NEXT);
}


static int
dataset_name(uintptr_t addr, char *buf)
{
    static int gotid;
    static mdb_ctf_id_t dd_id;
    uintptr_t dd_parent;
    char dd_myname[MAXNAMELEN];

    if (!gotid) {
        if (mdb_ctf_lookup_by_name("struct dsl_dir",
            &dd_id) == -1) {
            mdb_warn("couldn't find struct dsl_dir");
            return (DCMD_ERR);
        }
        gotid = TRUE;
    }
    if (GETMEMBID(addr, &dd_id, dd_parent, dd_parent) ||
        GETMEMBID(addr, &dd_id, dd_myname, dd_myname)) {
        return (DCMD_ERR);
    }

    if (dd_parent) {
        if (dataset_name(dd_parent, buf))
            return (DCMD_ERR);
        strcat(buf, "/");
    }

    if (dd_myname[0])
        strcat(buf, dd_myname);
    else
        strcat(buf, "???");

    return (0);
}

static int
objset_name(uintptr_t addr, char *buf)
{
    static int gotid;
    static mdb_ctf_id_t os_id, ds_id;
    uintptr_t os_dsl_dataset;
    char ds_snapname[MAXNAMELEN];
    uintptr_t ds_dir;

    buf[0] = '\0';

    if (!gotid) {
        if (mdb_ctf_lookup_by_name("struct objset",
            &os_id) == -1) {
            mdb_warn("couldn't find struct objset");
            return (DCMD_ERR);
        }
        if (mdb_ctf_lookup_by_name("struct dsl_dataset",
            &ds_id) == -1) {
            mdb_warn("couldn't find struct dsl_dataset");
            return (DCMD_ERR);
        }

        gotid = TRUE;
    }

    if (GETMEMBID(addr, &os_id, os_dsl_dataset, os_dsl_dataset))
        return (DCMD_ERR);

    if (os_dsl_dataset == 0) {
        strcat(buf, "mos");
        return (0);
    }

    if (GETMEMBID(os_dsl_dataset, &ds_id, ds_snapname, ds_snapname) ||
        GETMEMBID(os_dsl_dataset, &ds_id, ds_dir, ds_dir)) {
        return (DCMD_ERR);
    }

    if (ds_dir && dataset_name(ds_dir, buf))
        return (DCMD_ERR);

    if (ds_snapname[0]) {
        strcat(buf, "@");
        strcat(buf, ds_snapname);
    }
    return (0);
}

static void
enum_lookup(char *out, size_t size, mdb_ctf_id_t id, int val,
    const char *prefix)
{
    const char *cp;
    size_t len = strlen(prefix);

    if ((cp = mdb_ctf_enum_name(id, val)) != NULL) {
        if (strncmp(cp, prefix, len) == 0)
            cp += len;
        (void) strncpy(out, cp, size);
    } else {
        mdb_snprintf(out, size, "? (%d)", val);
    }
}

/* ARGSUSED */
static int
zfs_params(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    /*
     * This table can be approximately generated by running:
     * egrep "^[a-z0-9_]+ [a-z0-9_]+( =.*)?;" *.c | cut -d ' ' -f 2
     */
    static const char *params[] = {
        "arc_reduce_dnlc_percent",
        "zfs_arc_max",
        "zfs_arc_min",
        "arc_shrink_shift",
        "zfs_mdcomp_disable",
        "zfs_prefetch_disable",
        "zfetch_max_streams",
        "zfetch_min_sec_reap",
        "zfetch_block_cap",
        "zfetch_array_rd_sz",
        "zfs_default_bs",
        "zfs_default_ibs",
        "metaslab_aliquot",
        "reference_tracking_enable",
        "reference_history",
        "spa_max_replication_override",
        "spa_mode_global",
        "zfs_flags",
        "zfs_txg_synctime",
        "zfs_txg_timeout",
        "zfs_write_limit_min",
        "zfs_write_limit_max",
        "zfs_write_limit_shift",
        "zfs_write_limit_override",
        "zfs_no_write_throttle",
        "zfs_vdev_cache_max",
        "zfs_vdev_cache_size",
        "zfs_vdev_cache_bshift",
        "vdev_mirror_shift",
        "zfs_vdev_max_pending",
        "zfs_vdev_min_pending",
        "zfs_scrub_limit",
        "zfs_no_scrub_io",
        "zfs_no_scrub_prefetch",
        "zfs_vdev_time_shift",
        "zfs_vdev_ramp_rate",
        "zfs_vdev_aggregation_limit",
        "fzap_default_block_shift",
        "zfs_immediate_write_sz",
        "zfs_read_chunk_size",
        "zil_disable",
        "zfs_nocacheflush",
        "metaslab_gang_bang",
        "metaslab_df_alloc_threshold",
        "metaslab_df_free_pct",
        "zio_injection_enabled",
        "zvol_immediate_write_sz",
    };

    for (int i = 0; i < sizeof (params) / sizeof (params[0]); i++) {
        int sz;
        uint64_t val64;
        uint32_t *val32p = (uint32_t *)&val64;

        sz = mdb_readvar(&val64, params[i]);
        if (sz == 4) {
            mdb_printf("%s = 0x%x\n", params[i], *val32p);
        } else if (sz == 8) {
            mdb_printf("%s = 0x%llx\n", params[i], val64);
        } else {
            mdb_warn("variable %s not found", params[i]);
        }
    }

    return (DCMD_OK);
}

/* ARGSUSED */
static int
blkptr(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    mdb_ctf_id_t type_enum, checksum_enum, compress_enum;
    char type[80], checksum[80], compress[80];
    blkptr_t blk, *bp = &blk;
    char buf[BP_SPRINTF_LEN];

    if (mdb_vread(&blk, sizeof (blkptr_t), addr) == -1) {
        mdb_warn("failed to read blkptr_t");
        return (DCMD_ERR);
    }

    if (mdb_ctf_lookup_by_name("enum dmu_object_type", &type_enum) == -1 ||
        mdb_ctf_lookup_by_name("enum zio_checksum", &checksum_enum) == -1 ||
        mdb_ctf_lookup_by_name("enum zio_compress", &compress_enum) == -1) {
        mdb_warn("Could not find blkptr enumerated types");
        return (DCMD_ERR);
    }

    enum_lookup(type, sizeof (type), type_enum,
        BP_GET_TYPE(bp), "DMU_OT_");
    enum_lookup(checksum, sizeof (checksum), checksum_enum,
        BP_GET_CHECKSUM(bp), "ZIO_CHECKSUM_");
    enum_lookup(compress, sizeof (compress), compress_enum,
        BP_GET_COMPRESS(bp), "ZIO_COMPRESS_");

    SPRINTF_BLKPTR(mdb_snprintf, '\n', buf, bp, type, checksum, compress);

    mdb_printf("%s\n", buf);

    return (DCMD_OK);
}

/* ARGSUSED */
static int
dbuf(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    mdb_ctf_id_t id;
    dmu_buf_t db;
    uintptr_t objset;
    uint8_t level;
    uint64_t blkid;
    uint64_t holds;
    char objectname[32];
    char blkidname[32];
    char path[MAXNAMELEN];

    if (DCMD_HDRSPEC(flags)) {
        mdb_printf("        addr object lvl blkid holds os\n");
    }

    if (mdb_ctf_lookup_by_name("struct dmu_buf_impl", &id) == -1) {
        mdb_warn("couldn't find struct dmu_buf_impl_t");
        return (DCMD_ERR);
    }

    if (GETMEMBID(addr, &id, db_objset, objset) ||
        GETMEMBID(addr, &id, db, db) ||
        GETMEMBID(addr, &id, db_level, level) ||
        GETMEMBID(addr, &id, db_blkid, blkid)) {
        return (WALK_ERR);
    }

    if (getrefcount(addr, &id, "db_holds", &holds)) {
        return (WALK_ERR);
    }

    if (db.db_object == DMU_META_DNODE_OBJECT)
        (void) strcpy(objectname, "mdn");
    else
        (void) mdb_snprintf(objectname, sizeof (objectname), "%llx",
            (u_longlong_t)db.db_object);

    if (blkid == DB_BONUS_BLKID)
        (void) strcpy(blkidname, "bonus");
    else
        (void) mdb_snprintf(blkidname, sizeof (blkidname), "%llx",
            (u_longlong_t)blkid);

    if (objset_name(objset, path)) {
        return (WALK_ERR);
    }

    mdb_printf("%p %8s %1u %9s %2llu %s\n",
        addr, objectname, level, blkidname, holds, path);

    return (DCMD_OK);
}

/* ARGSUSED */
static int
dbuf_stats(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
#define HISTOSZ 32
    uintptr_t dbp;
    dmu_buf_impl_t db;
    dbuf_hash_table_t ht;
    uint64_t bucket, ndbufs;
    uint64_t histo[HISTOSZ];
    uint64_t histo2[HISTOSZ];
    int i, maxidx;

    if (mdb_readvar(&ht, "dbuf_hash_table") == -1) {
        mdb_warn("failed to read 'dbuf_hash_table'");
        return (DCMD_ERR);
    }

    for (i = 0; i < HISTOSZ; i++) {
        histo[i] = 0;
        histo2[i] = 0;
    }

    ndbufs = 0;
    for (bucket = 0; bucket < ht.hash_table_mask+1; bucket++) {
        int len;

        if (mdb_vread(&dbp, sizeof (void *),
            (uintptr_t)(ht.hash_table+bucket)) == -1) {
            mdb_warn("failed to read hash bucket %u at %p",
                bucket, ht.hash_table+bucket);
            return (DCMD_ERR);
        }

        len = 0;
        while (dbp != 0) {
            if (mdb_vread(&db, sizeof (dmu_buf_impl_t),
                dbp) == -1) {
                mdb_warn("failed to read dbuf at %p", dbp);
                return (DCMD_ERR);
            }
            dbp = (uintptr_t)db.db_hash_next;
            for (i = MIN(len, HISTOSZ - 1); i >= 0; i--)
                histo2[i]++;
            len++;
            ndbufs++;
        }

        if (len >= HISTOSZ)
            len = HISTOSZ-1;
        histo[len]++;
    }

    mdb_printf("hash table has %llu buckets, %llu dbufs "
        "(avg %llu buckets/dbuf)\n",
        ht.hash_table_mask+1, ndbufs,
        (ht.hash_table_mask+1)/ndbufs);

    mdb_printf("\n");
    maxidx = 0;
    for (i = 0; i < HISTOSZ; i++)
        if (histo[i] > 0)
            maxidx = i;
    mdb_printf("hash chain length   number of buckets\n");
    for (i = 0; i <= maxidx; i++)
        mdb_printf("%u          %llu\n", i, histo[i]);

    mdb_printf("\n");
    maxidx = 0;
    for (i = 0; i < HISTOSZ; i++)
        if (histo2[i] > 0)
            maxidx = i;
    mdb_printf("hash chain depth    number of dbufs\n");
    for (i = 0; i <= maxidx; i++)
        mdb_printf("%u or more      %llu    %llu%%\n",
            i, histo2[i], histo2[i]*100/ndbufs);


    return (DCMD_OK);
}

#define CHAIN_END 0xffff
/*
 * ::zap_leaf [-v]
 *
 * Print a zap_leaf_phys_t, assumed to be 16k
 */
/* ARGSUSED */
static int
zap_leaf(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    char buf[16*1024];
    int verbose = B_FALSE;
    int four = B_FALSE;
    zap_leaf_t l;
    zap_leaf_phys_t *zlp = (void *)buf;
    int i;

    if (mdb_getopts(argc, argv,
        'v', MDB_OPT_SETBITS, TRUE, &verbose,
        '4', MDB_OPT_SETBITS, TRUE, &four,
        NULL) != argc)
        return (DCMD_USAGE);

    l.l_phys = zlp;
    l.l_bs = 14; /* assume 16k blocks */
    if (four)
        l.l_bs = 12;

    if (!(flags & DCMD_ADDRSPEC)) {
        return (DCMD_USAGE);
    }

    if (mdb_vread(buf, sizeof (buf), addr) == -1) {
        mdb_warn("failed to read zap_leaf_phys_t at %p", addr);
        return (DCMD_ERR);
    }

    if (zlp->l_hdr.lh_block_type != ZBT_LEAF ||
        zlp->l_hdr.lh_magic != ZAP_LEAF_MAGIC) {
        mdb_warn("This does not appear to be a zap_leaf_phys_t");
        return (DCMD_ERR);
    }

    mdb_printf("zap_leaf_phys_t at %p:\n", addr);
    mdb_printf("    lh_prefix_len = %u\n", zlp->l_hdr.lh_prefix_len);
    mdb_printf("    lh_prefix = %llx\n", zlp->l_hdr.lh_prefix);
    mdb_printf("    lh_nentries = %u\n", zlp->l_hdr.lh_nentries);
    mdb_printf("    lh_nfree = %u\n", zlp->l_hdr.lh_nfree,
        zlp->l_hdr.lh_nfree * 100 / (ZAP_LEAF_NUMCHUNKS(&l)));
    mdb_printf("    lh_freelist = %u\n", zlp->l_hdr.lh_freelist);
    mdb_printf("    lh_flags = %x (%s)\n", zlp->l_hdr.lh_flags,
        zlp->l_hdr.lh_flags & ZLF_ENTRIES_CDSORTED ?
        "ENTRIES_CDSORTED" : "");

    if (verbose) {
        mdb_printf(" hash table:\n");
        for (i = 0; i < ZAP_LEAF_HASH_NUMENTRIES(&l); i++) {
            if (zlp->l_hash[i] != CHAIN_END)
                mdb_printf("    %u: %u\n", i, zlp->l_hash[i]);
        }
    }

    mdb_printf(" chunks:\n");
    for (i = 0; i < ZAP_LEAF_NUMCHUNKS(&l); i++) {
        /* LINTED: alignment */
        zap_leaf_chunk_t *zlc = &ZAP_LEAF_CHUNK(&l, i);
        switch (zlc->l_entry.le_type) {
        case ZAP_CHUNK_FREE:
            if (verbose) {
                mdb_printf("    %u: free; lf_next = %u\n",
                    i, zlc->l_free.lf_next);
            }
            break;
        case ZAP_CHUNK_ENTRY:
            mdb_printf("    %u: entry\n", i);
            if (verbose) {
                mdb_printf("        le_next = %u\n",
                    zlc->l_entry.le_next);
            }
            mdb_printf("        le_name_chunk = %u\n",
                zlc->l_entry.le_name_chunk);
            mdb_printf("        le_name_numints = %u\n",
                zlc->l_entry.le_name_numints);
            mdb_printf("        le_value_chunk = %u\n",
                zlc->l_entry.le_value_chunk);
            mdb_printf("        le_value_intlen = %u\n",
                zlc->l_entry.le_value_intlen);
            mdb_printf("        le_value_numints = %u\n",
                zlc->l_entry.le_value_numints);
            mdb_printf("        le_cd = %u\n",
                zlc->l_entry.le_cd);
            mdb_printf("        le_hash = %llx\n",
                zlc->l_entry.le_hash);
            break;
        case ZAP_CHUNK_ARRAY:
            mdb_printf("    %u: array \"%s\"\n",
                i, zlc->l_array.la_array);
            if (verbose) {
                int j;
                mdb_printf("        ");
                for (j = 0; j < ZAP_LEAF_ARRAY_BYTES; j++) {
                    mdb_printf("%02x ",
                        zlc->l_array.la_array[j]);
                }
                mdb_printf("\n");
            }
            if (zlc->l_array.la_next != CHAIN_END) {
                mdb_printf("        lf_next = %u\n",
                    zlc->l_array.la_next);
            }
            break;
        default:
            mdb_printf("    %u: undefined type %u\n",
                zlc->l_entry.le_type);
        }
    }

    return (DCMD_OK);
}

typedef struct dbufs_data {
    mdb_ctf_id_t id;
    uint64_t objset;
    uint64_t object;
    uint64_t level;
    uint64_t blkid;
    char *osname;
} dbufs_data_t;

#define DBUFS_UNSET (0xbaddcafedeadbeefULL)

/* ARGSUSED */
static int
dbufs_cb(uintptr_t addr, const void *unknown, void *arg)
{
    dbufs_data_t *data = arg;
    uintptr_t objset;
    dmu_buf_t db;
    uint8_t level;
    uint64_t blkid;
    char osname[MAXNAMELEN];

    if (GETMEMBID(addr, &data->id, db_objset, objset) ||
        GETMEMBID(addr, &data->id, db, db) ||
        GETMEMBID(addr, &data->id, db_level, level) ||
        GETMEMBID(addr, &data->id, db_blkid, blkid)) {
        return (WALK_ERR);
    }

    if ((data->objset == DBUFS_UNSET || data->objset == objset) &&
        (data->osname == NULL || (objset_name(objset, osname) == 0 &&
        strcmp(data->osname, osname) == 0)) &&
        (data->object == DBUFS_UNSET || data->object == db.db_object) &&
        (data->level == DBUFS_UNSET || data->level == level) &&
        (data->blkid == DBUFS_UNSET || data->blkid == blkid)) {
        mdb_printf("%#lr\n", addr);
    }
    return (WALK_NEXT);
}

/* ARGSUSED */
static int
dbufs(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    dbufs_data_t data;
    char *object = NULL;
    char *blkid = NULL;

    data.objset = data.object = data.level = data.blkid = DBUFS_UNSET;
    data.osname = NULL;

    if (mdb_getopts(argc, argv,
        'O', MDB_OPT_UINT64, &data.objset,
        'n', MDB_OPT_STR, &data.osname,
        'o', MDB_OPT_STR, &object,
        'l', MDB_OPT_UINT64, &data.level,
        'b', MDB_OPT_STR, &blkid) != argc) {
        return (DCMD_USAGE);
    }

    if (object) {
        if (strcmp(object, "mdn") == 0) {
            data.object = DMU_META_DNODE_OBJECT;
        } else {
            data.object = mdb_strtoull(object);
        }
    }

    if (blkid) {
        if (strcmp(blkid, "bonus") == 0) {
            data.blkid = DB_BONUS_BLKID;
        } else {
            data.blkid = mdb_strtoull(blkid);
        }
    }

    if (mdb_ctf_lookup_by_name("struct dmu_buf_impl", &data.id) == -1) {
        mdb_warn("couldn't find struct dmu_buf_impl_t");
        return (DCMD_ERR);
    }

    if (mdb_walk("dmu_buf_impl_t", dbufs_cb, &data) != 0) {
        mdb_warn("can't walk dbufs");
        return (DCMD_ERR);
    }

    return (DCMD_OK);
}

typedef struct abuf_find_data {
    dva_t dva;
    mdb_ctf_id_t id;
} abuf_find_data_t;

/* ARGSUSED */
static int
abuf_find_cb(uintptr_t addr, const void *unknown, void *arg)
{
    abuf_find_data_t *data = arg;
    dva_t dva;

    if (GETMEMBID(addr, &data->id, b_dva, dva)) {
        return (WALK_ERR);
    }

    if (dva.dva_word[0] == data->dva.dva_word[0] &&
        dva.dva_word[1] == data->dva.dva_word[1]) {
        mdb_printf("%#lr\n", addr);
    }
    return (WALK_NEXT);
}

/* ARGSUSED */
static int
abuf_find(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    abuf_find_data_t data;
    GElf_Sym sym;
    int i;
    const char *syms[] = {
        "ARC_mru",
        "ARC_mru_ghost",
        "ARC_mfu",
        "ARC_mfu_ghost",
    };

    if (argc != 2)
        return (DCMD_USAGE);

    for (i = 0; i < 2; i ++) {
        switch (argv[i].a_type) {
        case MDB_TYPE_STRING:
            data.dva.dva_word[i] = mdb_strtoull(argv[i].a_un.a_str);
            break;
        case MDB_TYPE_IMMEDIATE:
            data.dva.dva_word[i] = argv[i].a_un.a_val;
            break;
        default:
            return (DCMD_USAGE);
        }
    }

    if (mdb_ctf_lookup_by_name("struct arc_buf_hdr", &data.id) == -1) {
        mdb_warn("couldn't find struct arc_buf_hdr");
        return (DCMD_ERR);
    }

    for (i = 0; i < sizeof (syms) / sizeof (syms[0]); i++) {
        if (mdb_lookup_by_name(syms[i], &sym)) {
            mdb_warn("can't find symbol %s", syms[i]);
            return (DCMD_ERR);
        }

        if (mdb_pwalk("list", abuf_find_cb, &data, sym.st_value) != 0) {
            mdb_warn("can't walk %s", syms[i]);
            return (DCMD_ERR);
        }
    }

    return (DCMD_OK);
}

/*ARGSUSED*/
static int
arc_print(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    kstat_named_t *stats;
    GElf_Sym sym;
    int nstats, i;
    uint_t opt_a = FALSE;
    uint_t opt_b = FALSE;
    uint_t shift = 0;
    const char *suffix;

    static const char *bytestats[] = {
        "p", "c", "c_min", "c_max", "size", NULL
    };

    static const char *extras[] = {
        "arc_no_grow", "arc_tempreserve",
        "arc_meta_used", "arc_meta_limit", "arc_meta_max",
        NULL
    };

    if (mdb_lookup_by_name("arc_stats", &sym) == -1) {
        mdb_warn("failed to find 'arc_stats'");
        return (DCMD_ERR);
    }

    stats = mdb_zalloc(sym.st_size, UM_SLEEP | UM_GC);

    if (mdb_vread(stats, sym.st_size, sym.st_value) == -1) {
        mdb_warn("couldn't read 'arc_stats' at %p", sym.st_value);
        return (DCMD_ERR);
    }

    nstats = sym.st_size / sizeof (kstat_named_t);

    /* NB: -a / opt_a are ignored for backwards compatability */
    if (mdb_getopts(argc, argv,
        'a', MDB_OPT_SETBITS, TRUE, &opt_a,
        'b', MDB_OPT_SETBITS, TRUE, &opt_b,
        'k', MDB_OPT_SETBITS, 10, &shift,
        'm', MDB_OPT_SETBITS, 20, &shift,
        'g', MDB_OPT_SETBITS, 30, &shift,
        NULL) != argc)
        return (DCMD_USAGE);

    if (!opt_b && !shift)
        shift = 20;

    switch (shift) {
    case 0:
        suffix = "B";
        break;
    case 10:
        suffix = "KB";
        break;
    case 20:
        suffix = "MB";
        break;
    case 30:
        suffix = "GB";
        break;
    default:
        suffix = "XX";
    }

    for (i = 0; i < nstats; i++) {
        int j;
        boolean_t bytes = B_FALSE;

        for (j = 0; bytestats[j]; j++) {
            if (strcmp(stats[i].name, bytestats[j]) == 0) {
                bytes = B_TRUE;
                break;
            }
        }

        if (bytes) {
            mdb_printf("%-25s = %9llu %s\n", stats[i].name,
                stats[i].value.ui64 >> shift, suffix);
        } else {
            mdb_printf("%-25s = %9llu\n", stats[i].name,
                stats[i].value.ui64);
        }
    }

    for (i = 0; extras[i]; i++) {
        uint64_t buf;

        if (mdb_lookup_by_name(extras[i], &sym) == -1) {
            mdb_warn("failed to find '%s'", extras[i]);
            return (DCMD_ERR);
        }

        if (sym.st_size != sizeof (uint64_t) &&
            sym.st_size != sizeof (uint32_t)) {
            mdb_warn("expected scalar for variable '%s'\n",
                extras[i]);
            return (DCMD_ERR);
        }

        if (mdb_vread(&buf, sym.st_size, sym.st_value) == -1) {
            mdb_warn("couldn't read '%s'", extras[i]);
            return (DCMD_ERR);
        }

        mdb_printf("%-25s = ", extras[i]);

        /* NB: all the 64-bit extras happen to be byte counts */
        if (sym.st_size == sizeof (uint64_t))
            mdb_printf("%9llu %s\n", buf >> shift, suffix);

        if (sym.st_size == sizeof (uint32_t))
            mdb_printf("%9d\n", *((uint32_t *)&buf));
    }
    return (DCMD_OK);
}

/*
 * ::spa
 *
 *  -c  Print configuration information as well
 *  -v  Print vdev state
 *  -e  Print vdev error stats
 *
 * Print a summarized spa_t.  When given no arguments, prints out a table of all
 * active pools on the system.
 */
/* ARGSUSED */
static int
spa_print(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    spa_t spa;
    const char *statetab[] = { "ACTIVE", "EXPORTED", "DESTROYED",
        "SPARE", "L2CACHE", "UNINIT", "UNAVAIL", "POTENTIAL" };
    const char *state;
    int config = FALSE;
    int vdevs = FALSE;
    int errors = FALSE;

    if (mdb_getopts(argc, argv,
        'c', MDB_OPT_SETBITS, TRUE, &config,
        'v', MDB_OPT_SETBITS, TRUE, &vdevs,
        'e', MDB_OPT_SETBITS, TRUE, &errors,
        NULL) != argc)
        return (DCMD_USAGE);

    if (!(flags & DCMD_ADDRSPEC)) {
        if (mdb_walk_dcmd("spa", "spa", argc, argv) == -1) {
            mdb_warn("can't walk spa");
            return (DCMD_ERR);
        }

        return (DCMD_OK);
    }

    if (flags & DCMD_PIPE_OUT) {
        mdb_printf("%#lr\n", addr);
        return (DCMD_OK);
    }

    if (DCMD_HDRSPEC(flags))
        mdb_printf("%<u>%-?s %9s %-*s%</u>\n", "ADDR", "STATE",
            sizeof (uintptr_t) == 4 ? 60 : 52, "NAME");

    if (mdb_vread(&spa, sizeof (spa), addr) == -1) {
        mdb_warn("failed to read spa_t at %p", addr);
        return (DCMD_ERR);
    }

    if (spa.spa_state < 0 || spa.spa_state > POOL_STATE_UNAVAIL)
        state = "UNKNOWN";
    else
        state = statetab[spa.spa_state];

    mdb_printf("%0?p %9s %s\n", addr, state, spa.spa_name);

    if (config) {
        mdb_printf("\n");
        mdb_inc_indent(4);
        if (mdb_call_dcmd("spa_config", addr, flags, 0,
            NULL) != DCMD_OK)
            return (DCMD_ERR);
        mdb_dec_indent(4);
    }

    if (vdevs || errors) {
        mdb_arg_t v;

        v.a_type = MDB_TYPE_STRING;
        v.a_un.a_str = "-e";

        mdb_printf("\n");
        mdb_inc_indent(4);
        if (mdb_call_dcmd("spa_vdevs", addr, flags, errors ? 1 : 0,
            &v) != DCMD_OK)
            return (DCMD_ERR);
        mdb_dec_indent(4);
    }

    return (DCMD_OK);
}

/*
 * ::spa_config
 *
 * Given a spa_t, print the configuration information stored in spa_config.
 * Since it's just an nvlist, format it as an indented list of name=value pairs.
 * We simply read the value of spa_config and pass off to ::nvlist.
 */
/* ARGSUSED */
static int
spa_print_config(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    spa_t spa;

    if (argc != 0 || !(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (mdb_vread(&spa, sizeof (spa), addr) == -1) {
        mdb_warn("failed to read spa_t at %p", addr);
        return (DCMD_ERR);
    }

    if (spa.spa_config == NULL) {
        mdb_printf("(none)\n");
        return (DCMD_OK);
    }

    return (mdb_call_dcmd("nvlist", (uintptr_t)spa.spa_config, flags,
        0, NULL));
}

/*
 * ::vdev
 *
 * Print out a summarized vdev_t, in the following form:
 *
 * ADDR             STATE   AUX            DESC
 * fffffffbcde23df0 HEALTHY -              /dev/dsk/c0t0d0
 *
 * If '-r' is specified, recursively visit all children.
 *
 * With '-e', the statistics associated with the vdev are printed as well.
 */
static int
do_print_vdev(uintptr_t addr, int flags, int depth, int stats,
    int recursive)
{
    vdev_t vdev;
    char desc[MAXNAMELEN];
    int c, children;
    uintptr_t *child;
    const char *state, *aux;

    if (mdb_vread(&vdev, sizeof (vdev), (uintptr_t)addr) == -1) {
        mdb_warn("failed to read vdev_t at %p\n", (uintptr_t)addr);
        return (DCMD_ERR);
    }

    if (flags & DCMD_PIPE_OUT) {
        mdb_printf("%#lr", addr);
    } else {
        if (vdev.vdev_path != NULL) {
            if (mdb_readstr(desc, sizeof (desc),
                (uintptr_t)vdev.vdev_path) == -1) {
                mdb_warn("failed to read vdev_path at %p\n",
                    vdev.vdev_path);
                return (DCMD_ERR);
            }
        } else if (vdev.vdev_ops != NULL) {
            vdev_ops_t ops;
            if (mdb_vread(&ops, sizeof (ops),
                (uintptr_t)vdev.vdev_ops) == -1) {
                mdb_warn("failed to read vdev_ops at %p\n",
                    vdev.vdev_ops);
                return (DCMD_ERR);
            }
            (void) strcpy(desc, ops.vdev_op_type);
        } else {
            (void) strcpy(desc, "<unknown>");
        }

        if (depth == 0 && DCMD_HDRSPEC(flags))
            mdb_printf("%<u>%-?s %-9s %-12s %-*s%</u>\n",
                "ADDR", "STATE", "AUX",
                sizeof (uintptr_t) == 4 ? 43 : 35,
                "DESCRIPTION");

        mdb_printf("%0?p ", addr);

        switch (vdev.vdev_state) {
        case VDEV_STATE_CLOSED:
            state = "CLOSED";
            break;
        case VDEV_STATE_OFFLINE:
            state = "OFFLINE";
            break;
        case VDEV_STATE_CANT_OPEN:
            state = "CANT_OPEN";
            break;
        case VDEV_STATE_DEGRADED:
            state = "DEGRADED";
            break;
        case VDEV_STATE_HEALTHY:
            state = "HEALTHY";
            break;
        case VDEV_STATE_REMOVED:
            state = "REMOVED";
            break;
        case VDEV_STATE_FAULTED:
            state = "FAULTED";
            break;
        default:
            state = "UNKNOWN";
            break;
        }

        switch (vdev.vdev_stat.vs_aux) {
        case VDEV_AUX_NONE:
            aux = "-";
            break;
        case VDEV_AUX_OPEN_FAILED:
            aux = "OPEN_FAILED";
            break;
        case VDEV_AUX_CORRUPT_DATA:
            aux = "CORRUPT_DATA";
            break;
        case VDEV_AUX_NO_REPLICAS:
            aux = "NO_REPLICAS";
            break;
        case VDEV_AUX_BAD_GUID_SUM:
            aux = "BAD_GUID_SUM";
            break;
        case VDEV_AUX_TOO_SMALL:
            aux = "TOO_SMALL";
            break;
        case VDEV_AUX_BAD_LABEL:
            aux = "BAD_LABEL";
            break;
        case VDEV_AUX_VERSION_NEWER:
            aux = "VERS_NEWER";
            break;
        case VDEV_AUX_VERSION_OLDER:
            aux = "VERS_OLDER";
            break;
        case VDEV_AUX_SPARED:
            aux = "SPARED";
            break;
        case VDEV_AUX_ERR_EXCEEDED:
            aux = "ERR_EXCEEDED";
            break;
        case VDEV_AUX_IO_FAILURE:
            aux = "IO_FAILURE";
            break;
        case VDEV_AUX_BAD_LOG:
            aux = "BAD_LOG";
            break;
        case VDEV_AUX_EXTERNAL:
            aux = "EXTERNAL";
            break;
        case VDEV_AUX_SPLIT_POOL:
            aux = "SPLIT_POOL";
            break;
        default:
            aux = "UNKNOWN";
            break;
        }

        mdb_printf("%-9s %-12s %*s%s\n", state, aux, depth, "", desc);

        if (stats) {
            vdev_stat_t *vs = &vdev.vdev_stat;
            int i;

            mdb_inc_indent(4);
            mdb_printf("\n");
            mdb_printf("%<u>       %12s %12s %12s %12s "
                "%12s%</u>\n", "READ", "WRITE", "FREE", "CLAIM",
                "IOCTL");
            mdb_printf("OPS     ");
            for (i = 1; i < ZIO_TYPES; i++)
                mdb_printf("%11#llx%s", vs->vs_ops[i],
                    i == ZIO_TYPES - 1 ? "" : "  ");
            mdb_printf("\n");
            mdb_printf("BYTES   ");
            for (i = 1; i < ZIO_TYPES; i++)
                mdb_printf("%11#llx%s", vs->vs_bytes[i],
                    i == ZIO_TYPES - 1 ? "" : "  ");


            mdb_printf("\n");
            mdb_printf("EREAD    %10#llx\n", vs->vs_read_errors);
            mdb_printf("EWRITE   %10#llx\n", vs->vs_write_errors);
            mdb_printf("ECKSUM   %10#llx\n",
                vs->vs_checksum_errors);
            mdb_dec_indent(4);
        }

        if (stats)
            mdb_printf("\n");
    }

    children = vdev.vdev_children;

    if (children == 0 || !recursive)
        return (DCMD_OK);

    child = mdb_alloc(children * sizeof (void *), UM_SLEEP | UM_GC);
    if (mdb_vread(child, children * sizeof (void *),
        (uintptr_t)vdev.vdev_child) == -1) {
        mdb_warn("failed to read vdev children at %p", vdev.vdev_child);
        return (DCMD_ERR);
    }

    for (c = 0; c < children; c++) {
        if (do_print_vdev(child[c], flags, depth + 2, stats,
            recursive))
            return (DCMD_ERR);
    }

    return (DCMD_OK);
}

static int
vdev_print(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    int recursive = FALSE;
    int stats = FALSE;
    uint64_t depth = 0;

    if (mdb_getopts(argc, argv,
        'r', MDB_OPT_SETBITS, TRUE, &recursive,
        'e', MDB_OPT_SETBITS, TRUE, &stats,
        'd', MDB_OPT_UINT64, &depth,
        NULL) != argc)
        return (DCMD_USAGE);

    if (!(flags & DCMD_ADDRSPEC)) {
        mdb_warn("no vdev_t address given\n");
        return (DCMD_ERR);
    }

    return (do_print_vdev(addr, flags, (int)depth, stats, recursive));
}

typedef struct metaslab_walk_data {
    uint64_t mw_numvdevs;
    uintptr_t *mw_vdevs;
    int mw_curvdev;
    uint64_t mw_nummss;
    uintptr_t *mw_mss;
    int mw_curms;
} metaslab_walk_data_t;

static int
metaslab_walk_step(mdb_walk_state_t *wsp)
{
    metaslab_walk_data_t *mw = wsp->walk_data;
    metaslab_t ms;
    uintptr_t msp;

    if (mw->mw_curvdev >= mw->mw_numvdevs)
        return (WALK_DONE);

    if (mw->mw_mss == NULL) {
        uintptr_t mssp;
        uintptr_t vdevp;

        ASSERT(mw->mw_curms == 0);
        ASSERT(mw->mw_nummss == 0);

        vdevp = mw->mw_vdevs[mw->mw_curvdev];
        if (GETMEMB(vdevp, struct vdev, vdev_ms, mssp) ||
            GETMEMB(vdevp, struct vdev, vdev_ms_count, mw->mw_nummss)) {
            return (WALK_ERR);
        }

        mw->mw_mss = mdb_alloc(mw->mw_nummss * sizeof (void*),
            UM_SLEEP | UM_GC);
        if (mdb_vread(mw->mw_mss, mw->mw_nummss * sizeof (void*),
            mssp) == -1) {
            mdb_warn("failed to read vdev_ms at %p", mssp);
            return (WALK_ERR);
        }
    }

    if (mw->mw_curms >= mw->mw_nummss) {
        mw->mw_mss = NULL;
        mw->mw_curms = 0;
        mw->mw_nummss = 0;
        mw->mw_curvdev++;
        return (WALK_NEXT);
    }

    msp = mw->mw_mss[mw->mw_curms];
    if (mdb_vread(&ms, sizeof (metaslab_t), msp) == -1) {
        mdb_warn("failed to read metaslab_t at %p", msp);
        return (WALK_ERR);
    }

    mw->mw_curms++;

    return (wsp->walk_callback(msp, &ms, wsp->walk_cbdata));
}

/* ARGSUSED */
static int
metaslab_walk_init(mdb_walk_state_t *wsp)
{
    metaslab_walk_data_t *mw;
    uintptr_t root_vdevp;
    uintptr_t childp;

    if (wsp->walk_addr == NULL) {
        mdb_warn("must supply address of spa_t\n");
        return (WALK_ERR);
    }

    mw = mdb_zalloc(sizeof (metaslab_walk_data_t), UM_SLEEP | UM_GC);

    if (GETMEMB(wsp->walk_addr, struct spa, spa_root_vdev, root_vdevp) ||
        GETMEMB(root_vdevp, struct vdev, vdev_children, mw->mw_numvdevs) ||
        GETMEMB(root_vdevp, struct vdev, vdev_child, childp)) {
        return (DCMD_ERR);
    }

    mw->mw_vdevs = mdb_alloc(mw->mw_numvdevs * sizeof (void *),
        UM_SLEEP | UM_GC);
    if (mdb_vread(mw->mw_vdevs, mw->mw_numvdevs * sizeof (void *),
        childp) == -1) {
        mdb_warn("failed to read root vdev children at %p", childp);
        return (DCMD_ERR);
    }

    wsp->walk_data = mw;

    return (WALK_NEXT);
}

typedef struct mdb_spa {
    uintptr_t spa_dsl_pool;
    uintptr_t spa_root_vdev;
} mdb_spa_t;

typedef struct mdb_dsl_dir {
    uintptr_t dd_phys;
    int64_t dd_space_towrite[TXG_SIZE];
} mdb_dsl_dir_t;

typedef struct mdb_dsl_dir_phys {
    uint64_t dd_used_bytes;
    uint64_t dd_compressed_bytes;
    uint64_t dd_uncompressed_bytes;
} mdb_dsl_dir_phys_t;

typedef struct mdb_vdev {
    uintptr_t vdev_parent;
    uintptr_t vdev_ms;
    uint64_t vdev_ms_count;
    vdev_stat_t vdev_stat;
} mdb_vdev_t;

typedef struct mdb_metaslab {
    space_map_t ms_allocmap[TXG_SIZE];
    space_map_t ms_freemap[TXG_SIZE];
    space_map_t ms_map;
    space_map_obj_t ms_smo;
    space_map_obj_t ms_smo_syncing;
} mdb_metaslab_t;

typedef struct space_data {
    uint64_t ms_allocmap[TXG_SIZE];
    uint64_t ms_freemap[TXG_SIZE];
    uint64_t ms_map;
    uint64_t avail;
    uint64_t nowavail;
} space_data_t;

/* ARGSUSED */
static int
space_cb(uintptr_t addr, const void *unknown, void *arg)
{
    space_data_t *sd = arg;
    mdb_metaslab_t ms;

    if (GETMEMB(addr, struct metaslab, ms_allocmap, ms.ms_allocmap) ||
        GETMEMB(addr, struct metaslab, ms_freemap, ms.ms_freemap) ||
        GETMEMB(addr, struct metaslab, ms_map, ms.ms_map) ||
        GETMEMB(addr, struct metaslab, ms_smo, ms.ms_smo) ||
        GETMEMB(addr, struct metaslab, ms_smo_syncing, ms.ms_smo_syncing)) {
        return (WALK_ERR);
    }

    sd->ms_allocmap[0] += ms.ms_allocmap[0].sm_space;
    sd->ms_allocmap[1] += ms.ms_allocmap[1].sm_space;
    sd->ms_allocmap[2] += ms.ms_allocmap[2].sm_space;
    sd->ms_allocmap[3] += ms.ms_allocmap[3].sm_space;
    sd->ms_freemap[0] += ms.ms_freemap[0].sm_space;
    sd->ms_freemap[1] += ms.ms_freemap[1].sm_space;
    sd->ms_freemap[2] += ms.ms_freemap[2].sm_space;
    sd->ms_freemap[3] += ms.ms_freemap[3].sm_space;
    sd->ms_map += ms.ms_map.sm_space;
    sd->avail += ms.ms_map.sm_size - ms.ms_smo.smo_alloc;
    sd->nowavail += ms.ms_map.sm_size - ms.ms_smo_syncing.smo_alloc;

    return (WALK_NEXT);
}

/*
 * ::spa_space [-b]
 *
 * Given a spa_t, print out it's on-disk space usage and in-core
 * estimates of future usage.  If -b is given, print space in bytes.
 * Otherwise print in megabytes.
 */
/* ARGSUSED */
static int
spa_space(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    mdb_spa_t spa;
    uintptr_t dp_root_dir;
    mdb_dsl_dir_t dd;
    mdb_dsl_dir_phys_t dsp;
    uint64_t children;
    uintptr_t childaddr;
    space_data_t sd;
    int shift = 20;
    char *suffix = "M";
    int bits = FALSE;

    if (mdb_getopts(argc, argv, 'b', MDB_OPT_SETBITS, TRUE, &bits, NULL) !=
        argc)
        return (DCMD_USAGE);
    if (!(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (bits) {
        shift = 0;
        suffix = "";
    }

    if (GETMEMB(addr, struct spa, spa_dsl_pool, spa.spa_dsl_pool) ||
        GETMEMB(addr, struct spa, spa_root_vdev, spa.spa_root_vdev) ||
        GETMEMB(spa.spa_root_vdev, struct vdev, vdev_children, children) ||
        GETMEMB(spa.spa_root_vdev, struct vdev, vdev_child, childaddr) ||
        GETMEMB(spa.spa_dsl_pool, struct dsl_pool,
        dp_root_dir, dp_root_dir) ||
        GETMEMB(dp_root_dir, struct dsl_dir, dd_phys, dd.dd_phys) ||
        GETMEMB(dp_root_dir, struct dsl_dir,
        dd_space_towrite, dd.dd_space_towrite) ||
        GETMEMB(dd.dd_phys, struct dsl_dir_phys,
        dd_used_bytes, dsp.dd_used_bytes) ||
        GETMEMB(dd.dd_phys, struct dsl_dir_phys,
        dd_compressed_bytes, dsp.dd_compressed_bytes) ||
        GETMEMB(dd.dd_phys, struct dsl_dir_phys,
        dd_uncompressed_bytes, dsp.dd_uncompressed_bytes)) {
        return (DCMD_ERR);
    }

    mdb_printf("dd_space_towrite = %llu%s %llu%s %llu%s %llu%s\n",
        dd.dd_space_towrite[0] >> shift, suffix,
        dd.dd_space_towrite[1] >> shift, suffix,
        dd.dd_space_towrite[2] >> shift, suffix,
        dd.dd_space_towrite[3] >> shift, suffix);

    mdb_printf("dd_phys.dd_used_bytes = %llu%s\n",
        dsp.dd_used_bytes >> shift, suffix);
    mdb_printf("dd_phys.dd_compressed_bytes = %llu%s\n",
        dsp.dd_compressed_bytes >> shift, suffix);
    mdb_printf("dd_phys.dd_uncompressed_bytes = %llu%s\n",
        dsp.dd_uncompressed_bytes >> shift, suffix);

    bzero(&sd, sizeof (sd));
    if (mdb_pwalk("metaslab", space_cb, &sd, addr) != 0) {
        mdb_warn("can't walk metaslabs");
        return (DCMD_ERR);
    }

    mdb_printf("ms_allocmap = %llu%s %llu%s %llu%s %llu%s\n",
        sd.ms_allocmap[0] >> shift, suffix,
        sd.ms_allocmap[1] >> shift, suffix,
        sd.ms_allocmap[2] >> shift, suffix,
        sd.ms_allocmap[3] >> shift, suffix);
    mdb_printf("ms_freemap = %llu%s %llu%s %llu%s %llu%s\n",
        sd.ms_freemap[0] >> shift, suffix,
        sd.ms_freemap[1] >> shift, suffix,
        sd.ms_freemap[2] >> shift, suffix,
        sd.ms_freemap[3] >> shift, suffix);
    mdb_printf("ms_map = %llu%s\n", sd.ms_map >> shift, suffix);
    mdb_printf("last synced avail = %llu%s\n", sd.avail >> shift, suffix);
    mdb_printf("current syncing avail = %llu%s\n",
        sd.nowavail >> shift, suffix);

    return (DCMD_OK);
}

/*
 * ::spa_verify
 *
 * Given a spa_t, verify that that the pool is self-consistent.
 * Currently, it only checks to make sure that the vdev tree exists.
 */
/* ARGSUSED */
static int
spa_verify(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    spa_t spa;

    if (argc != 0 || !(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (mdb_vread(&spa, sizeof (spa), addr) == -1) {
        mdb_warn("failed to read spa_t at %p", addr);
        return (DCMD_ERR);
    }

    if (spa.spa_root_vdev == NULL) {
        mdb_printf("no vdev tree present\n");
        return (DCMD_OK);
    }

    return (DCMD_OK);
}

static int
spa_print_aux(spa_aux_vdev_t *sav, uint_t flags, mdb_arg_t *v,
    const char *name)
{
    uintptr_t *aux;
    size_t len;
    int ret, i;

    /*
     * Iterate over aux vdevs and print those out as well.  This is a
     * little annoying because we don't have a root vdev to pass to ::vdev.
     * Instead, we print a single line and then call it for each child
     * vdev.
     */
    if (sav->sav_count != 0) {
        v[1].a_type = MDB_TYPE_STRING;
        v[1].a_un.a_str = "-d";
        v[2].a_type = MDB_TYPE_IMMEDIATE;
        v[2].a_un.a_val = 2;

        len = sav->sav_count * sizeof (uintptr_t);
        aux = mdb_alloc(len, UM_SLEEP);
        if (mdb_vread(aux, len,
            (uintptr_t)sav->sav_vdevs) == -1) {
            mdb_free(aux, len);
            mdb_warn("failed to read l2cache vdevs at %p",
                sav->sav_vdevs);
            return (DCMD_ERR);
        }

        mdb_printf("%-?s %-9s %-12s %s\n", "-", "-", "-", name);

        for (i = 0; i < sav->sav_count; i++) {
            ret = mdb_call_dcmd("vdev", aux[i], flags, 3, v);
            if (ret != DCMD_OK) {
                mdb_free(aux, len);
                return (ret);
            }
        }

        mdb_free(aux, len);
    }

    return (0);
}

/*
 * ::spa_vdevs
 *
 *  -e  Include error stats
 *
 * Print out a summarized list of vdevs for the given spa_t.
 * This is accomplished by invoking "::vdev -re" on the root vdev, as well as
 * iterating over the cache devices.
 */
/* ARGSUSED */
static int
spa_vdevs(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    spa_t spa;
    mdb_arg_t v[3];
    int errors = FALSE;
    int ret;

    if (mdb_getopts(argc, argv,
        'e', MDB_OPT_SETBITS, TRUE, &errors,
        NULL) != argc)
        return (DCMD_USAGE);

    if (!(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (mdb_vread(&spa, sizeof (spa), addr) == -1) {
        mdb_warn("failed to read spa_t at %p", addr);
        return (DCMD_ERR);
    }

    /*
     * Unitialized spa_t structures can have a NULL root vdev.
     */
    if (spa.spa_root_vdev == NULL) {
        mdb_printf("no associated vdevs\n");
        return (DCMD_OK);
    }

    v[0].a_type = MDB_TYPE_STRING;
    v[0].a_un.a_str = errors ? "-re" : "-r";

    ret = mdb_call_dcmd("vdev", (uintptr_t)spa.spa_root_vdev,
        flags, 1, v);
    if (ret != DCMD_OK)
        return (ret);

    if (spa_print_aux(&spa.spa_l2cache, flags, v, "cache") != 0 ||
        spa_print_aux(&spa.spa_spares, flags, v, "spares") != 0)
        return (DCMD_ERR);

    return (DCMD_OK);
}

/*
 * ::zio
 *
 * Print a summary of zio_t and all its children.  This is intended to display a
 * zio tree, and hence we only pick the most important pieces of information for
 * the main summary.  More detailed information can always be found by doing a
 * '::print zio' on the underlying zio_t.  The columns we display are:
 *
 *  ADDRESS     TYPE    STAGE       WAITER
 *
 * The 'address' column is indented by one space for each depth level as we
 * descend down the tree.
 */

#define ZIO_MAXINDENT   24
#define ZIO_MAXWIDTH    (sizeof (uintptr_t) * 2 + ZIO_MAXINDENT)
#define ZIO_WALK_SELF   0
#define ZIO_WALK_CHILD  1
#define ZIO_WALK_PARENT 2

typedef struct zio_print_args {
    int zpa_current_depth;
    int zpa_min_depth;
    int zpa_max_depth;
    int zpa_type;
    uint_t  zpa_flags;
} zio_print_args_t;

static int zio_child_cb(uintptr_t addr, const void *unknown, void *arg);

static int
zio_print_cb(uintptr_t addr, const void *data, void *priv)
{
    const zio_t *zio = data;
    zio_print_args_t *zpa = priv;
    mdb_ctf_id_t type_enum, stage_enum;
    int indent = zpa->zpa_current_depth;
    const char *type, *stage;
    uintptr_t laddr;

    if (indent > ZIO_MAXINDENT)
        indent = ZIO_MAXINDENT;

    if (mdb_ctf_lookup_by_name("enum zio_type", &type_enum) == -1 ||
        mdb_ctf_lookup_by_name("enum zio_stage", &stage_enum) == -1) {
        mdb_warn("failed to lookup zio enums");
        return (WALK_ERR);
    }

    if ((type = mdb_ctf_enum_name(type_enum, zio->io_type)) != NULL)
        type += sizeof ("ZIO_TYPE_") - 1;
    else
        type = "?";

    if ((stage = mdb_ctf_enum_name(stage_enum, zio->io_stage)) != NULL)
        stage += sizeof ("ZIO_STAGE_") - 1;
    else
        stage = "?";

    if (zpa->zpa_current_depth >= zpa->zpa_min_depth) {
        if (zpa->zpa_flags & DCMD_PIPE_OUT) {
            mdb_printf("%?p\n", addr);
        } else {
            mdb_printf("%*s%-*p %-5s %-16s ", indent, "",
                ZIO_MAXWIDTH - indent, addr, type, stage);
            if (zio->io_waiter)
                mdb_printf("%?p\n", zio->io_waiter);
            else
                mdb_printf("-\n");
        }
    }

    if (zpa->zpa_current_depth >= zpa->zpa_max_depth)
        return (WALK_NEXT);

    if (zpa->zpa_type == ZIO_WALK_PARENT)
        laddr = addr + OFFSETOF(zio_t, io_parent_list);
    else
        laddr = addr + OFFSETOF(zio_t, io_child_list);

    zpa->zpa_current_depth++;
    if (mdb_pwalk("list", zio_child_cb, zpa, laddr) != 0) {
        mdb_warn("failed to walk zio_t children at %p\n", laddr);
        return (WALK_ERR);
    }
    zpa->zpa_current_depth--;

    return (WALK_NEXT);
}

/* ARGSUSED */
static int
zio_child_cb(uintptr_t addr, const void *unknown, void *arg)
{
    zio_link_t zl;
    zio_t zio;
    uintptr_t ziop;
    zio_print_args_t *zpa = arg;

    if (mdb_vread(&zl, sizeof (zl), addr) == -1) {
        mdb_warn("failed to read zio_link_t at %p", addr);
        return (WALK_ERR);
    }

    if (zpa->zpa_type == ZIO_WALK_PARENT)
        ziop = (uintptr_t)zl.zl_parent;
    else
        ziop = (uintptr_t)zl.zl_child;

    if (mdb_vread(&zio, sizeof (zio_t), ziop) == -1) {
        mdb_warn("failed to read zio_t at %p", ziop);
        return (WALK_ERR);
    }

    return (zio_print_cb(ziop, &zio, arg));
}

/* ARGSUSED */
static int
zio_print(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    zio_t zio;
    zio_print_args_t zpa = { 0 };

    if (!(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (mdb_getopts(argc, argv,
        'r', MDB_OPT_SETBITS, INT_MAX, &zpa.zpa_max_depth,
        'c', MDB_OPT_SETBITS, ZIO_WALK_CHILD, &zpa.zpa_type,
        'p', MDB_OPT_SETBITS, ZIO_WALK_PARENT, &zpa.zpa_type,
        NULL) != argc)
        return (DCMD_USAGE);

    zpa.zpa_flags = flags;
    if (zpa.zpa_max_depth != 0) {
        if (zpa.zpa_type == ZIO_WALK_SELF)
            zpa.zpa_type = ZIO_WALK_CHILD;
    } else if (zpa.zpa_type != ZIO_WALK_SELF) {
        zpa.zpa_min_depth = 1;
        zpa.zpa_max_depth = 1;
    }

    if (mdb_vread(&zio, sizeof (zio_t), addr) == -1) {
        mdb_warn("failed to read zio_t at %p", addr);
        return (DCMD_ERR);
    }

    if (!(flags & DCMD_PIPE_OUT) && DCMD_HDRSPEC(flags))
        mdb_printf("%<u>%-*s %-5s %-16s %-?s%</u>\n", ZIO_MAXWIDTH,
            "ADDRESS", "TYPE", "STAGE", "WAITER");

    if (zio_print_cb(addr, &zio, &zpa) != WALK_NEXT)
        return (DCMD_ERR);

    return (DCMD_OK);
}

/*
 * [addr]::zio_state
 *
 * Print a summary of all zio_t structures on the system, or for a particular
 * pool.  This is equivalent to '::walk zio_root | ::zio'.
 */
/*ARGSUSED*/
static int
zio_state(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    /*
     * MDB will remember the last address of the pipeline, so if we don't
     * zero this we'll end up trying to walk zio structures for a
     * non-existent spa_t.
     */
    if (!(flags & DCMD_ADDRSPEC))
        addr = 0;

    return (mdb_pwalk_dcmd("zio_root", "zio", argc, argv, addr));
}

typedef struct txg_list_walk_data {
    uintptr_t lw_head[TXG_SIZE];
    int lw_txgoff;
    int lw_maxoff;
    size_t  lw_offset;
    void    *lw_obj;
} txg_list_walk_data_t;

static int
txg_list_walk_init_common(mdb_walk_state_t *wsp, int txg, int maxoff)
{
    txg_list_walk_data_t *lwd;
    txg_list_t list;
    int i;

    lwd = mdb_alloc(sizeof (txg_list_walk_data_t), UM_SLEEP | UM_GC);
    if (mdb_vread(&list, sizeof (txg_list_t), wsp->walk_addr) == -1) {
        mdb_warn("failed to read txg_list_t at %#lx", wsp->walk_addr);
        return (WALK_ERR);
    }

    for (i = 0; i < TXG_SIZE; i++)
        lwd->lw_head[i] = (uintptr_t)list.tl_head[i];
    lwd->lw_offset = list.tl_offset;
    lwd->lw_obj = mdb_alloc(lwd->lw_offset + sizeof (txg_node_t),
        UM_SLEEP | UM_GC);
    lwd->lw_txgoff = txg;
    lwd->lw_maxoff = maxoff;

    wsp->walk_addr = lwd->lw_head[lwd->lw_txgoff];
    wsp->walk_data = lwd;

    return (WALK_NEXT);
}

static int
txg_list_walk_init(mdb_walk_state_t *wsp)
{
    return (txg_list_walk_init_common(wsp, 0, TXG_SIZE-1));
}

static int
txg_list0_walk_init(mdb_walk_state_t *wsp)
{
    return (txg_list_walk_init_common(wsp, 0, 0));
}

static int
txg_list1_walk_init(mdb_walk_state_t *wsp)
{
    return (txg_list_walk_init_common(wsp, 1, 1));
}

static int
txg_list2_walk_init(mdb_walk_state_t *wsp)
{
    return (txg_list_walk_init_common(wsp, 2, 2));
}

static int
txg_list3_walk_init(mdb_walk_state_t *wsp)
{
    return (txg_list_walk_init_common(wsp, 3, 3));
}

static int
txg_list_walk_step(mdb_walk_state_t *wsp)
{
    txg_list_walk_data_t *lwd = wsp->walk_data;
    uintptr_t addr;
    txg_node_t *node;
    int status;

    while (wsp->walk_addr == NULL && lwd->lw_txgoff < lwd->lw_maxoff) {
        lwd->lw_txgoff++;
        wsp->walk_addr = lwd->lw_head[lwd->lw_txgoff];
    }

    if (wsp->walk_addr == NULL)
        return (WALK_DONE);

    addr = wsp->walk_addr - lwd->lw_offset;

    if (mdb_vread(lwd->lw_obj,
        lwd->lw_offset + sizeof (txg_node_t), addr) == -1) {
        mdb_warn("failed to read list element at %#lx", addr);
        return (WALK_ERR);
    }

    status = wsp->walk_callback(addr, lwd->lw_obj, wsp->walk_cbdata);
    node = (txg_node_t *)((uintptr_t)lwd->lw_obj + lwd->lw_offset);
    wsp->walk_addr = (uintptr_t)node->tn_next[lwd->lw_txgoff];

    return (status);
}

/*
 * ::walk spa
 *
 * Walk all named spa_t structures in the namespace.  This is nothing more than
 * a layered avl walk.
 */
static int
spa_walk_init(mdb_walk_state_t *wsp)
{
    GElf_Sym sym;

    if (wsp->walk_addr != NULL) {
        mdb_warn("spa walk only supports global walks\n");
        return (WALK_ERR);
    }

    if (mdb_lookup_by_obj(ZFS_OBJ_NAME, "spa_namespace_avl", &sym) == -1) {
        mdb_warn("failed to find symbol 'spa_namespace_avl'");
        return (WALK_ERR);
    }

    wsp->walk_addr = (uintptr_t)sym.st_value;

    if (mdb_layered_walk("avl", wsp) == -1) {
        mdb_warn("failed to walk 'avl'\n");
        return (WALK_ERR);
    }

    return (WALK_NEXT);
}

static int
spa_walk_step(mdb_walk_state_t *wsp)
{
    spa_t   spa;

    if (mdb_vread(&spa, sizeof (spa), wsp->walk_addr) == -1) {
        mdb_warn("failed to read spa_t at %p", wsp->walk_addr);
        return (WALK_ERR);
    }

    return (wsp->walk_callback(wsp->walk_addr, &spa, wsp->walk_cbdata));
}

/*
 * [addr]::walk zio
 *
 * Walk all active zio_t structures on the system.  This is simply a layered
 * walk on top of ::walk zio_cache, with the optional ability to limit the
 * structures to a particular pool.
 */
static int
zio_walk_init(mdb_walk_state_t *wsp)
{
    wsp->walk_data = (void *)wsp->walk_addr;

    if (mdb_layered_walk("zio_cache", wsp) == -1) {
        mdb_warn("failed to walk 'zio_cache'\n");
        return (WALK_ERR);
    }

    return (WALK_NEXT);
}

static int
zio_walk_step(mdb_walk_state_t *wsp)
{
    zio_t zio;

    if (mdb_vread(&zio, sizeof (zio), wsp->walk_addr) == -1) {
        mdb_warn("failed to read zio_t at %p", wsp->walk_addr);
        return (WALK_ERR);
    }

    if (wsp->walk_data != NULL && wsp->walk_data != zio.io_spa)
        return (WALK_NEXT);

    return (wsp->walk_callback(wsp->walk_addr, &zio, wsp->walk_cbdata));
}

/*
 * [addr]::walk zio_root
 *
 * Walk only root zio_t structures, optionally for a particular spa_t.
 */
static int
zio_walk_root_step(mdb_walk_state_t *wsp)
{
    zio_t zio;

    if (mdb_vread(&zio, sizeof (zio), wsp->walk_addr) == -1) {
        mdb_warn("failed to read zio_t at %p", wsp->walk_addr);
        return (WALK_ERR);
    }

    if (wsp->walk_data != NULL && wsp->walk_data != zio.io_spa)
        return (WALK_NEXT);

    /* If the parent list is not empty, ignore */
    if (zio.io_parent_list.list_head.list_next !=
        &((zio_t *)wsp->walk_addr)->io_parent_list.list_head)
        return (WALK_NEXT);

    return (wsp->walk_callback(wsp->walk_addr, &zio, wsp->walk_cbdata));
}

#define NICENUM_BUFLEN 6

static int
snprintfrac(char *buf, int len,
    uint64_t numerator, uint64_t denom, int frac_digits)
{
    int mul = 1;
    int whole, frac, i;

    for (i = frac_digits; i; i--)
        mul *= 10;
    whole = numerator / denom;
    frac = mul * numerator / denom - mul * whole;
    return (mdb_snprintf(buf, len, "%u.%0*u", whole, frac_digits, frac));
}

static void
mdb_nicenum(uint64_t num, char *buf)
{
    uint64_t n = num;
    int index = 0;
    char *u;

    while (n >= 1024) {
        n = (n + (1024 / 2)) / 1024; /* Round up or down */
        index++;
    }

    u = &" \0K\0M\0G\0T\0P\0E\0"[index*2];

    if (index == 0) {
        (void) mdb_snprintf(buf, NICENUM_BUFLEN, "%llu",
            (u_longlong_t)n);
    } else if (n < 10 && (num & (num - 1)) != 0) {
        (void) snprintfrac(buf, NICENUM_BUFLEN,
            num, 1ULL << 10 * index, 2);
        strcat(buf, u);
    } else if (n < 100 && (num & (num - 1)) != 0) {
        (void) snprintfrac(buf, NICENUM_BUFLEN,
            num, 1ULL << 10 * index, 1);
        strcat(buf, u);
    } else {
        (void) mdb_snprintf(buf, NICENUM_BUFLEN, "%llu%s",
            (u_longlong_t)n, u);
    }
}

/*
 * ::zfs_blkstats
 *
 *  -v  print verbose per-level information
 *
 */
static int
zfs_blkstats(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    boolean_t verbose = B_FALSE;
    zfs_all_blkstats_t stats;
    dmu_object_type_t t;
    zfs_blkstat_t *tzb;
    uint64_t ditto;
    dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES + 10];
    /* +10 in case it grew */

    if (mdb_readvar(&dmu_ot, "dmu_ot") == -1) {
        mdb_warn("failed to read 'dmu_ot'");
        return (DCMD_ERR);
    }

    if (mdb_getopts(argc, argv,
        'v', MDB_OPT_SETBITS, TRUE, &verbose,
        NULL) != argc)
        return (DCMD_USAGE);

    if (!(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (GETMEMB(addr, struct spa, spa_dsl_pool, addr) ||
        GETMEMB(addr, struct dsl_pool, dp_blkstats, addr) ||
        mdb_vread(&stats, sizeof (zfs_all_blkstats_t), addr) == -1) {
        mdb_warn("failed to read data at %p;", addr);
        mdb_printf("maybe no stats? run \"zpool scrub\" first.");
        return (DCMD_ERR);
    }

    tzb = &stats.zab_type[DN_MAX_LEVELS][DMU_OT_NUMTYPES];
    if (tzb->zb_gangs != 0) {
        mdb_printf("Ganged blocks: %llu\n",
            (longlong_t)tzb->zb_gangs);
    }

    ditto = tzb->zb_ditto_2_of_2_samevdev + tzb->zb_ditto_2_of_3_samevdev +
        tzb->zb_ditto_3_of_3_samevdev;
    if (ditto != 0) {
        mdb_printf("Dittoed blocks on same vdev: %llu\n",
            (longlong_t)ditto);
    }

    mdb_printf("\nBlocks\tLSIZE\tPSIZE\tASIZE"
        "\t  avg\t comp\t%%Total\tType\n");

    for (t = 0; t <= DMU_OT_NUMTYPES; t++) {
        char csize[NICENUM_BUFLEN], lsize[NICENUM_BUFLEN];
        char psize[NICENUM_BUFLEN], asize[NICENUM_BUFLEN];
        char avg[NICENUM_BUFLEN];
        char comp[NICENUM_BUFLEN], pct[NICENUM_BUFLEN];
        char typename[64];
        int l;


        if (t == DMU_OT_DEFERRED)
            strcpy(typename, "deferred free");
        else if (t == DMU_OT_TOTAL)
            strcpy(typename, "Total");
        else if (mdb_readstr(typename, sizeof (typename),
            (uintptr_t)dmu_ot[t].ot_name) == -1) {
            mdb_warn("failed to read type name");
            return (DCMD_ERR);
        }

        if (stats.zab_type[DN_MAX_LEVELS][t].zb_asize == 0)
            continue;

        for (l = -1; l < DN_MAX_LEVELS; l++) {
            int level = (l == -1 ? DN_MAX_LEVELS : l);
            zfs_blkstat_t *zb = &stats.zab_type[level][t];

            if (zb->zb_asize == 0)
                continue;

            /*
             * Don't print each level unless requested.
             */
            if (!verbose && level != DN_MAX_LEVELS)
                continue;

            /*
             * If all the space is level 0, don't print the
             * level 0 separately.
             */
            if (level == 0 && zb->zb_asize ==
                stats.zab_type[DN_MAX_LEVELS][t].zb_asize)
                continue;

            mdb_nicenum(zb->zb_count, csize);
            mdb_nicenum(zb->zb_lsize, lsize);
            mdb_nicenum(zb->zb_psize, psize);
            mdb_nicenum(zb->zb_asize, asize);
            mdb_nicenum(zb->zb_asize / zb->zb_count, avg);
            (void) snprintfrac(comp, NICENUM_BUFLEN,
                zb->zb_lsize, zb->zb_psize, 2);
            (void) snprintfrac(pct, NICENUM_BUFLEN,
                100 * zb->zb_asize, tzb->zb_asize, 2);

            mdb_printf("%6s\t%5s\t%5s\t%5s\t%5s"
                "\t%5s\t%6s\t",
                csize, lsize, psize, asize, avg, comp, pct);

            if (level == DN_MAX_LEVELS)
                mdb_printf("%s\n", typename);
            else
                mdb_printf("  L%d %s\n",
                    level, typename);
        }
    }

    return (DCMD_OK);
}

/* ARGSUSED */
static int
reference_cb(uintptr_t addr, const void *ignored, void *arg)
{
    static int gotid;
    static mdb_ctf_id_t ref_id;
    uintptr_t ref_holder;
    uintptr_t ref_removed;
    uint64_t ref_number;
    boolean_t holder_is_str;
    char holder_str[128];
    boolean_t removed = (boolean_t)arg;

    if (!gotid) {
        if (mdb_ctf_lookup_by_name("struct reference", &ref_id) == -1) {
            mdb_warn("couldn't find struct reference");
            return (WALK_ERR);
        }
        gotid = TRUE;
    }

    if (GETMEMBID(addr, &ref_id, ref_holder, ref_holder) ||
        GETMEMBID(addr, &ref_id, ref_removed, ref_removed) ||
        GETMEMBID(addr, &ref_id, ref_number, ref_number))
        return (WALK_ERR);

    if (mdb_readstr(holder_str, sizeof (holder_str), ref_holder) != -1) {
        char *cp;
        holder_is_str = B_TRUE;
        for (cp = holder_str; *cp; cp++) {
            if (!isprint(*cp)) {
                holder_is_str = B_FALSE;
                break;
            }
        }
    } else {
        holder_is_str = B_FALSE;
    }

    if (removed)
        mdb_printf("removed ");
    mdb_printf("reference ");
    if (ref_number != 1)
        mdb_printf("with count=%llu ", ref_number);
    mdb_printf("with tag %p", (void*)ref_holder);
    if (holder_is_str)
        mdb_printf(" \"%s\"", holder_str);
    mdb_printf(", held at:\n");

    (void) mdb_call_dcmd("whatis", addr, DCMD_ADDRSPEC, 0, NULL);

    if (removed) {
        mdb_printf("removed at:\n");
        (void) mdb_call_dcmd("whatis", ref_removed,
            DCMD_ADDRSPEC, 0, NULL);
    }

    mdb_printf("\n");

    return (WALK_NEXT);
}

/* ARGSUSED */
static int
refcount(uintptr_t addr, uint_t flags, int argc, const mdb_arg_t *argv)
{
    uint64_t rc_count, rc_removed_count;
    uintptr_t rc_list, rc_removed;
    static int gotid;
    static mdb_ctf_id_t rc_id;
    ulong_t off;

    if (!(flags & DCMD_ADDRSPEC))
        return (DCMD_USAGE);

    if (!gotid) {
        if (mdb_ctf_lookup_by_name("struct refcount", &rc_id) == -1) {
            mdb_warn("couldn't find struct refcount");
            return (DCMD_ERR);
        }
        gotid = TRUE;
    }

    if (GETMEMBID(addr, &rc_id, rc_count, rc_count) ||
        GETMEMBID(addr, &rc_id, rc_removed_count, rc_removed_count))
        return (DCMD_ERR);

    mdb_printf("refcount_t at %p has %llu current holds, "
        "%llu recently released holds\n",
        addr, (longlong_t)rc_count, (longlong_t)rc_removed_count);

    if (rc_count > 0)
        mdb_printf("current holds:\n");
    if (mdb_ctf_offsetof(rc_id, "rc_list", &off) == -1)
        return (DCMD_ERR);
    rc_list = addr + off/NBBY;
    mdb_pwalk("list", reference_cb, (void*)B_FALSE, rc_list);

    if (rc_removed_count > 0)
        mdb_printf("released holds:\n");
    if (mdb_ctf_offsetof(rc_id, "rc_removed", &off) == -1)
        return (DCMD_ERR);
    rc_removed = addr + off/NBBY;
    mdb_pwalk("list", reference_cb, (void*)B_TRUE, rc_removed);

    return (DCMD_OK);
}

/*
 * MDB module linkage information:
 *
 * We declare a list of structures describing our dcmds, and a function
 * named _mdb_init to return a pointer to our module information.
 */

static const mdb_dcmd_t dcmds[] = {
    { "arc", "[-bkmg]", "print ARC variables", arc_print },
    { "blkptr", ":", "print blkptr_t", blkptr },
    { "dbuf", ":", "print dmu_buf_impl_t", dbuf },
    { "dbuf_stats", ":", "dbuf stats", dbuf_stats },
    { "dbufs",
        "\t[-O objset_t*] [-n objset_name | \"mos\"] "
        "[-o object | \"mdn\"] \n"
        "\t[-l level] [-b blkid | \"bonus\"]",
        "find dmu_buf_impl_t's that match specified criteria", dbufs },
    { "abuf_find", "dva_word[0] dva_word[1]",
        "find arc_buf_hdr_t of a specified DVA",
        abuf_find },
    { "spa", "?[-cv]", "spa_t summary", spa_print },
    { "spa_config", ":", "print spa_t configuration", spa_print_config },
    { "spa_verify", ":", "verify spa_t consistency", spa_verify },
    { "spa_space", ":[-b]", "print spa_t on-disk space usage", spa_space },
    { "spa_vdevs", ":", "given a spa_t, print vdev summary", spa_vdevs },
    { "vdev", ":[-re]\n"
        "\t-r display recursively\n"
        "\t-e print statistics",
        "vdev_t summary", vdev_print },
    { "zio", ":[cpr]\n"
        "\t-c display children\n"
        "\t-p display parents\n"
        "\t-r display recursively",
        "zio_t summary", zio_print },
    { "zio_state", "?", "print out all zio_t structures on system or "
        "for a particular pool", zio_state },
    { "zfs_blkstats", ":[-v]",
        "given a spa_t, print block type stats from last scrub",
        zfs_blkstats },
    { "zfs_params", "", "print zfs tunable parameters", zfs_params },
    { "refcount", "", "print refcount_t holders", refcount },
    { "zap_leaf", "", "print zap_leaf_phys_t", zap_leaf },
    { NULL }
};

static const mdb_walker_t walkers[] = {
    /*
     * In userland, there is no generic provider of list_t walkers, so we
     * need to add it.
     */
#ifndef _KERNEL
    { LIST_WALK_NAME, LIST_WALK_DESC,
        list_walk_init, list_walk_step, list_walk_fini },
#endif
    { "zms_freelist", "walk ZFS metaslab freelist",
        freelist_walk_init, freelist_walk_step, NULL },
    { "txg_list", "given any txg_list_t *, walk all entries in all txgs",
        txg_list_walk_init, txg_list_walk_step, NULL },
    { "txg_list0", "given any txg_list_t *, walk all entries in txg 0",
        txg_list0_walk_init, txg_list_walk_step, NULL },
    { "txg_list1", "given any txg_list_t *, walk all entries in txg 1",
        txg_list1_walk_init, txg_list_walk_step, NULL },
    { "txg_list2", "given any txg_list_t *, walk all entries in txg 2",
        txg_list2_walk_init, txg_list_walk_step, NULL },
    { "txg_list3", "given any txg_list_t *, walk all entries in txg 3",
        txg_list3_walk_init, txg_list_walk_step, NULL },
    { "zio", "walk all zio structures, optionally for a particular spa_t",
        zio_walk_init, zio_walk_step, NULL },
    { "zio_root", "walk all root zio_t structures, optionally for a "
        "particular spa_t",
        zio_walk_init, zio_walk_root_step, NULL },
    { "spa", "walk all spa_t entries in the namespace",
        spa_walk_init, spa_walk_step, NULL },
    { "metaslab", "given a spa_t *, walk all metaslab_t structures",
        metaslab_walk_init, metaslab_walk_step, NULL },
    { NULL }
};

static const mdb_modinfo_t modinfo = {
    MDB_API_VERSION, dcmds, walkers
};

const mdb_modinfo_t *
_mdb_init(void)
{
    return (&modinfo);
}