* 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.
+ * or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
/*
* Copyright (c) 2009, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2013 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
+ * Copyright (c) 2022 by Pawel Jakub Dawidek
*/
#include <sys/zfs_context.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/ddt.h>
+#include <sys/ddt_impl.h>
#include <sys/zap.h>
#include <sys/dmu_tx.h>
#include <sys/arc.h>
#include <sys/dsl_pool.h>
#include <sys/zio_checksum.h>
-#include <sys/zio_compress.h>
#include <sys/dsl_scan.h>
+#include <sys/abd.h>
+
+/*
+ * These are the only checksums valid for dedup. They must match the list
+ * from dedup_table in zfs_prop.c
+ */
+#define DDT_CHECKSUM_VALID(c) \
+ (c == ZIO_CHECKSUM_SHA256 || c == ZIO_CHECKSUM_SHA512 || \
+ c == ZIO_CHECKSUM_SKEIN || c == ZIO_CHECKSUM_EDONR || \
+ c == ZIO_CHECKSUM_BLAKE3)
+
+static kmem_cache_t *ddt_cache;
+static kmem_cache_t *ddt_entry_cache;
/*
* Enable/disable prefetching of dedup-ed blocks which are going to be freed.
*/
-int zfs_dedup_prefetch = 1;
+int zfs_dedup_prefetch = 0;
-static const ddt_ops_t *ddt_ops[DDT_TYPES] = {
+static const ddt_ops_t *const ddt_ops[DDT_TYPES] = {
&ddt_zap_ops,
};
-static const char *ddt_class_name[DDT_CLASSES] = {
+static const char *const ddt_class_name[DDT_CLASSES] = {
"ditto",
"duplicate",
"unique",
};
static void
-ddt_object_create(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_create(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
dmu_tx_t *tx)
{
spa_t *spa = ddt->ddt_spa;
objset_t *os = ddt->ddt_os;
uint64_t *objectp = &ddt->ddt_object[type][class];
- boolean_t prehash = zio_checksum_table[ddt->ddt_checksum].ci_dedup;
+ boolean_t prehash = zio_checksum_table[ddt->ddt_checksum].ci_flags &
+ ZCHECKSUM_FLAG_DEDUP;
char name[DDT_NAMELEN];
ddt_object_name(ddt, type, class, name);
- ASSERT(*objectp == 0);
- VERIFY(ddt_ops[type]->ddt_op_create(os, objectp, tx, prehash) == 0);
- ASSERT(*objectp != 0);
+ ASSERT3U(*objectp, ==, 0);
+ VERIFY0(ddt_ops[type]->ddt_op_create(os, objectp, tx, prehash));
+ ASSERT3U(*objectp, !=, 0);
- VERIFY(zap_add(os, DMU_POOL_DIRECTORY_OBJECT, name,
- sizeof (uint64_t), 1, objectp, tx) == 0);
+ VERIFY0(zap_add(os, DMU_POOL_DIRECTORY_OBJECT, name,
+ sizeof (uint64_t), 1, objectp, tx));
- VERIFY(zap_add(os, spa->spa_ddt_stat_object, name,
+ VERIFY0(zap_add(os, spa->spa_ddt_stat_object, name,
sizeof (uint64_t), sizeof (ddt_histogram_t) / sizeof (uint64_t),
- &ddt->ddt_histogram[type][class], tx) == 0);
+ &ddt->ddt_histogram[type][class], tx));
}
static void
-ddt_object_destroy(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_destroy(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
dmu_tx_t *tx)
{
spa_t *spa = ddt->ddt_spa;
ddt_object_name(ddt, type, class, name);
- ASSERT(*objectp != 0);
+ ASSERT3U(*objectp, !=, 0);
ASSERT(ddt_histogram_empty(&ddt->ddt_histogram[type][class]));
- VERIFY(ddt_object_count(ddt, type, class, &count) == 0 && count == 0);
- VERIFY(zap_remove(os, DMU_POOL_DIRECTORY_OBJECT, name, tx) == 0);
- VERIFY(zap_remove(os, spa->spa_ddt_stat_object, name, tx) == 0);
- VERIFY(ddt_ops[type]->ddt_op_destroy(os, *objectp, tx) == 0);
- bzero(&ddt->ddt_object_stats[type][class], sizeof (ddt_object_t));
+ VERIFY0(ddt_object_count(ddt, type, class, &count));
+ VERIFY0(count);
+ VERIFY0(zap_remove(os, DMU_POOL_DIRECTORY_OBJECT, name, tx));
+ VERIFY0(zap_remove(os, spa->spa_ddt_stat_object, name, tx));
+ VERIFY0(ddt_ops[type]->ddt_op_destroy(os, *objectp, tx));
+ memset(&ddt->ddt_object_stats[type][class], 0, sizeof (ddt_object_t));
*objectp = 0;
}
static int
-ddt_object_load(ddt_t *ddt, enum ddt_type type, enum ddt_class class)
+ddt_object_load(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
{
ddt_object_t *ddo = &ddt->ddt_object_stats[type][class];
dmu_object_info_t doi;
error = zap_lookup(ddt->ddt_os, DMU_POOL_DIRECTORY_OBJECT, name,
sizeof (uint64_t), 1, &ddt->ddt_object[type][class]);
-
- if (error)
+ if (error != 0)
return (error);
error = zap_lookup(ddt->ddt_os, ddt->ddt_spa->spa_ddt_stat_object, name,
sizeof (uint64_t), sizeof (ddt_histogram_t) / sizeof (uint64_t),
&ddt->ddt_histogram[type][class]);
+ if (error != 0)
+ return (error);
/*
* Seed the cached statistics.
ddo->ddo_dspace = doi.doi_physical_blocks_512 << 9;
ddo->ddo_mspace = doi.doi_fill_count * doi.doi_data_block_size;
- ASSERT(error == 0);
- return (error);
+ return (0);
}
static void
-ddt_object_sync(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_sync(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
dmu_tx_t *tx)
{
ddt_object_t *ddo = &ddt->ddt_object_stats[type][class];
ddt_object_name(ddt, type, class, name);
- VERIFY(zap_update(ddt->ddt_os, ddt->ddt_spa->spa_ddt_stat_object, name,
+ VERIFY0(zap_update(ddt->ddt_os, ddt->ddt_spa->spa_ddt_stat_object, name,
sizeof (uint64_t), sizeof (ddt_histogram_t) / sizeof (uint64_t),
- &ddt->ddt_histogram[type][class], tx) == 0);
+ &ddt->ddt_histogram[type][class], tx));
/*
* Cache DDT statistics; this is the only time they'll change.
*/
- VERIFY(ddt_object_info(ddt, type, class, &doi) == 0);
- VERIFY(ddt_object_count(ddt, type, class, &count) == 0);
+ VERIFY0(ddt_object_info(ddt, type, class, &doi));
+ VERIFY0(ddt_object_count(ddt, type, class, &count));
ddo->ddo_count = count;
ddo->ddo_dspace = doi.doi_physical_blocks_512 << 9;
ddo->ddo_mspace = doi.doi_fill_count * doi.doi_data_block_size;
}
+static boolean_t
+ddt_object_exists(ddt_t *ddt, ddt_type_t type, ddt_class_t class)
+{
+ return (!!ddt->ddt_object[type][class]);
+}
+
static int
-ddt_object_lookup(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_lookup(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
ddt_entry_t *dde)
{
if (!ddt_object_exists(ddt, type, class))
return (SET_ERROR(ENOENT));
return (ddt_ops[type]->ddt_op_lookup(ddt->ddt_os,
- ddt->ddt_object[type][class], dde));
+ ddt->ddt_object[type][class], &dde->dde_key,
+ dde->dde_phys, sizeof (dde->dde_phys)));
+}
+
+static int
+ddt_object_contains(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
+ const ddt_key_t *ddk)
+{
+ if (!ddt_object_exists(ddt, type, class))
+ return (SET_ERROR(ENOENT));
+
+ return (ddt_ops[type]->ddt_op_contains(ddt->ddt_os,
+ ddt->ddt_object[type][class], ddk));
}
static void
-ddt_object_prefetch(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
- ddt_entry_t *dde)
+ddt_object_prefetch(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
+ const ddt_key_t *ddk)
{
if (!ddt_object_exists(ddt, type, class))
return;
ddt_ops[type]->ddt_op_prefetch(ddt->ddt_os,
- ddt->ddt_object[type][class], dde);
+ ddt->ddt_object[type][class], ddk);
}
-int
-ddt_object_update(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+static int
+ddt_object_update(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
ddt_entry_t *dde, dmu_tx_t *tx)
{
ASSERT(ddt_object_exists(ddt, type, class));
return (ddt_ops[type]->ddt_op_update(ddt->ddt_os,
- ddt->ddt_object[type][class], dde, tx));
+ ddt->ddt_object[type][class], &dde->dde_key, dde->dde_phys,
+ sizeof (dde->dde_phys), tx));
}
static int
-ddt_object_remove(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
- ddt_entry_t *dde, dmu_tx_t *tx)
+ddt_object_remove(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
+ const ddt_key_t *ddk, dmu_tx_t *tx)
{
ASSERT(ddt_object_exists(ddt, type, class));
return (ddt_ops[type]->ddt_op_remove(ddt->ddt_os,
- ddt->ddt_object[type][class], dde, tx));
+ ddt->ddt_object[type][class], ddk, tx));
}
int
-ddt_object_walk(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_walk(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
uint64_t *walk, ddt_entry_t *dde)
{
ASSERT(ddt_object_exists(ddt, type, class));
return (ddt_ops[type]->ddt_op_walk(ddt->ddt_os,
- ddt->ddt_object[type][class], dde, walk));
+ ddt->ddt_object[type][class], walk, &dde->dde_key,
+ dde->dde_phys, sizeof (dde->dde_phys)));
}
int
-ddt_object_count(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_count(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
uint64_t *count)
{
ASSERT(ddt_object_exists(ddt, type, class));
}
int
-ddt_object_info(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_info(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
dmu_object_info_t *doi)
{
if (!ddt_object_exists(ddt, type, class))
doi));
}
-boolean_t
-ddt_object_exists(ddt_t *ddt, enum ddt_type type, enum ddt_class class)
-{
- return (!!ddt->ddt_object[type][class]);
-}
-
void
-ddt_object_name(ddt_t *ddt, enum ddt_type type, enum ddt_class class,
+ddt_object_name(ddt_t *ddt, ddt_type_t type, ddt_class_t class,
char *name)
{
- (void) sprintf(name, DMU_POOL_DDT,
+ (void) snprintf(name, DDT_NAMELEN, DMU_POOL_DDT,
zio_checksum_table[ddt->ddt_checksum].ci_name,
ddt_ops[type]->ddt_op_name, ddt_class_name[class]);
}
void
ddt_bp_fill(const ddt_phys_t *ddp, blkptr_t *bp, uint64_t txg)
{
- int d;
- ASSERT(txg != 0);
+ ASSERT3U(txg, !=, 0);
- for (d = 0; d < SPA_DVAS_PER_BP; d++)
+ for (int d = 0; d < SPA_DVAS_PER_BP; d++)
bp->blk_dva[d] = ddp->ddp_dva[d];
BP_SET_BIRTH(bp, txg, ddp->ddp_phys_birth);
}
+/*
+ * The bp created via this function may be used for repairs and scrub, but it
+ * will be missing the salt / IV required to do a full decrypting read.
+ */
void
ddt_bp_create(enum zio_checksum checksum,
const ddt_key_t *ddk, const ddt_phys_t *ddp, blkptr_t *bp)
ddt_bp_fill(ddp, bp, ddp->ddp_phys_birth);
bp->blk_cksum = ddk->ddk_cksum;
- bp->blk_fill = 1;
BP_SET_LSIZE(bp, DDK_GET_LSIZE(ddk));
BP_SET_PSIZE(bp, DDK_GET_PSIZE(ddk));
BP_SET_COMPRESS(bp, DDK_GET_COMPRESS(ddk));
+ BP_SET_CRYPT(bp, DDK_GET_CRYPT(ddk));
+ BP_SET_FILL(bp, 1);
BP_SET_CHECKSUM(bp, checksum);
BP_SET_TYPE(bp, DMU_OT_DEDUP);
BP_SET_LEVEL(bp, 0);
- BP_SET_DEDUP(bp, 0);
+ BP_SET_DEDUP(bp, 1);
BP_SET_BYTEORDER(bp, ZFS_HOST_BYTEORDER);
}
ddk->ddk_cksum = bp->blk_cksum;
ddk->ddk_prop = 0;
+ ASSERT(BP_IS_ENCRYPTED(bp) || !BP_USES_CRYPT(bp));
+
DDK_SET_LSIZE(ddk, BP_GET_LSIZE(bp));
DDK_SET_PSIZE(ddk, BP_GET_PSIZE(bp));
DDK_SET_COMPRESS(ddk, BP_GET_COMPRESS(bp));
+ DDK_SET_CRYPT(ddk, BP_USES_CRYPT(bp));
}
void
ddt_phys_fill(ddt_phys_t *ddp, const blkptr_t *bp)
{
- int d;
- ASSERT(ddp->ddp_phys_birth == 0);
+ ASSERT0(ddp->ddp_phys_birth);
- for (d = 0; d < SPA_DVAS_PER_BP; d++)
+ for (int d = 0; d < SPA_DVAS_PER_BP; d++)
ddp->ddp_dva[d] = bp->blk_dva[d];
ddp->ddp_phys_birth = BP_PHYSICAL_BIRTH(bp);
}
void
ddt_phys_clear(ddt_phys_t *ddp)
{
- bzero(ddp, sizeof (*ddp));
+ memset(ddp, 0, sizeof (*ddp));
}
void
ddt_phys_decref(ddt_phys_t *ddp)
{
if (ddp) {
- ASSERT(ddp->ddp_refcnt > 0);
+ ASSERT3U(ddp->ddp_refcnt, >, 0);
ddp->ddp_refcnt--;
}
}
-void
+static void
ddt_phys_free(ddt_t *ddt, ddt_key_t *ddk, ddt_phys_t *ddp, uint64_t txg)
{
blkptr_t blk;
ddt_bp_create(ddt->ddt_checksum, ddk, ddp, &blk);
+
+ /*
+ * We clear the dedup bit so that zio_free() will actually free the
+ * space, rather than just decrementing the refcount in the DDT.
+ */
+ BP_SET_DEDUP(&blk, 0);
+
ddt_phys_clear(ddp);
zio_free(ddt->ddt_spa, txg, &blk);
}
ddt_phys_select(const ddt_entry_t *dde, const blkptr_t *bp)
{
ddt_phys_t *ddp = (ddt_phys_t *)dde->dde_phys;
- int p;
- for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
+ for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
if (DVA_EQUAL(BP_IDENTITY(bp), &ddp->ddp_dva[0]) &&
BP_PHYSICAL_BIRTH(bp) == ddp->ddp_phys_birth)
return (ddp);
ddt_phys_total_refcnt(const ddt_entry_t *dde)
{
uint64_t refcnt = 0;
- int p;
- for (p = DDT_PHYS_SINGLE; p <= DDT_PHYS_TRIPLE; p++)
+ for (int p = DDT_PHYS_SINGLE; p <= DDT_PHYS_TRIPLE; p++)
refcnt += dde->dde_phys[p].ddp_refcnt;
return (refcnt);
}
-static void
-ddt_stat_generate(ddt_t *ddt, ddt_entry_t *dde, ddt_stat_t *dds)
-{
- spa_t *spa = ddt->ddt_spa;
- ddt_phys_t *ddp = dde->dde_phys;
- ddt_key_t *ddk = &dde->dde_key;
- uint64_t lsize = DDK_GET_LSIZE(ddk);
- uint64_t psize = DDK_GET_PSIZE(ddk);
- int p, d;
-
- bzero(dds, sizeof (*dds));
-
- for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
- uint64_t dsize = 0;
- uint64_t refcnt = ddp->ddp_refcnt;
-
- if (ddp->ddp_phys_birth == 0)
- continue;
-
- for (d = 0; d < SPA_DVAS_PER_BP; d++)
- dsize += dva_get_dsize_sync(spa, &ddp->ddp_dva[d]);
-
- dds->dds_blocks += 1;
- dds->dds_lsize += lsize;
- dds->dds_psize += psize;
- dds->dds_dsize += dsize;
-
- dds->dds_ref_blocks += refcnt;
- dds->dds_ref_lsize += lsize * refcnt;
- dds->dds_ref_psize += psize * refcnt;
- dds->dds_ref_dsize += dsize * refcnt;
- }
-}
-
-void
-ddt_stat_add(ddt_stat_t *dst, const ddt_stat_t *src, uint64_t neg)
-{
- const uint64_t *s = (const uint64_t *)src;
- uint64_t *d = (uint64_t *)dst;
- uint64_t *d_end = (uint64_t *)(dst + 1);
-
- ASSERT(neg == 0 || neg == -1ULL); /* add or subtract */
-
- while (d < d_end)
- *d++ += (*s++ ^ neg) - neg;
-}
-
-static void
-ddt_stat_update(ddt_t *ddt, ddt_entry_t *dde, uint64_t neg)
-{
- ddt_stat_t dds;
- ddt_histogram_t *ddh;
- int bucket;
-
- ddt_stat_generate(ddt, dde, &dds);
-
- bucket = highbit(dds.dds_ref_blocks) - 1;
- ASSERT(bucket >= 0);
-
- ddh = &ddt->ddt_histogram[dde->dde_type][dde->dde_class];
-
- ddt_stat_add(&ddh->ddh_stat[bucket], &dds, neg);
-}
-
-void
-ddt_histogram_add(ddt_histogram_t *dst, const ddt_histogram_t *src)
-{
- int h;
-
- for (h = 0; h < 64; h++)
- ddt_stat_add(&dst->ddh_stat[h], &src->ddh_stat[h], 0);
-}
-
-void
-ddt_histogram_stat(ddt_stat_t *dds, const ddt_histogram_t *ddh)
-{
- int h;
-
- bzero(dds, sizeof (*dds));
-
- for (h = 0; h < 64; h++)
- ddt_stat_add(dds, &ddh->ddh_stat[h], 0);
-}
-
-boolean_t
-ddt_histogram_empty(const ddt_histogram_t *ddh)
-{
- const uint64_t *s = (const uint64_t *)ddh;
- const uint64_t *s_end = (const uint64_t *)(ddh + 1);
-
- while (s < s_end)
- if (*s++ != 0)
- return (B_FALSE);
-
- return (B_TRUE);
-}
-
-void
-ddt_get_dedup_object_stats(spa_t *spa, ddt_object_t *ddo_total)
-{
- enum zio_checksum c;
- enum ddt_type type;
- enum ddt_class class;
-
- /* Sum the statistics we cached in ddt_object_sync(). */
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
- ddt_t *ddt = spa->spa_ddt[c];
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class < DDT_CLASSES;
- class++) {
- ddt_object_t *ddo =
- &ddt->ddt_object_stats[type][class];
- ddo_total->ddo_count += ddo->ddo_count;
- ddo_total->ddo_dspace += ddo->ddo_dspace;
- ddo_total->ddo_mspace += ddo->ddo_mspace;
- }
- }
- }
-
- /* ... and compute the averages. */
- if (ddo_total->ddo_count != 0) {
- ddo_total->ddo_dspace /= ddo_total->ddo_count;
- ddo_total->ddo_mspace /= ddo_total->ddo_count;
- }
-}
-
-void
-ddt_get_dedup_histogram(spa_t *spa, ddt_histogram_t *ddh)
+ddt_t *
+ddt_select(spa_t *spa, const blkptr_t *bp)
{
- enum zio_checksum c;
- enum ddt_type type;
- enum ddt_class class;
-
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
- ddt_t *ddt = spa->spa_ddt[c];
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class < DDT_CLASSES;
- class++) {
- ddt_histogram_add(ddh,
- &ddt->ddt_histogram_cache[type][class]);
- }
- }
- }
+ ASSERT(DDT_CHECKSUM_VALID(BP_GET_CHECKSUM(bp)));
+ return (spa->spa_ddt[BP_GET_CHECKSUM(bp)]);
}
void
-ddt_get_dedup_stats(spa_t *spa, ddt_stat_t *dds_total)
-{
- ddt_histogram_t *ddh_total;
-
- /* XXX: Move to a slab */
- ddh_total = kmem_zalloc(sizeof (ddt_histogram_t), KM_PUSHPAGE);
- ddt_get_dedup_histogram(spa, ddh_total);
- ddt_histogram_stat(dds_total, ddh_total);
- kmem_free(ddh_total, sizeof (ddt_histogram_t));
-}
-
-uint64_t
-ddt_get_dedup_dspace(spa_t *spa)
-{
- ddt_stat_t dds_total = { 0 };
-
- ddt_get_dedup_stats(spa, &dds_total);
- return (dds_total.dds_ref_dsize - dds_total.dds_dsize);
-}
-
-uint64_t
-ddt_get_pool_dedup_ratio(spa_t *spa)
-{
- ddt_stat_t dds_total = { 0 };
-
- ddt_get_dedup_stats(spa, &dds_total);
- if (dds_total.dds_dsize == 0)
- return (100);
-
- return (dds_total.dds_ref_dsize * 100 / dds_total.dds_dsize);
-}
-
-int
-ddt_ditto_copies_needed(ddt_t *ddt, ddt_entry_t *dde, ddt_phys_t *ddp_willref)
-{
- spa_t *spa = ddt->ddt_spa;
- uint64_t total_refcnt = 0;
- uint64_t ditto = spa->spa_dedup_ditto;
- int total_copies = 0;
- int desired_copies = 0;
- int p;
-
- for (p = DDT_PHYS_SINGLE; p <= DDT_PHYS_TRIPLE; p++) {
- ddt_phys_t *ddp = &dde->dde_phys[p];
- zio_t *zio = dde->dde_lead_zio[p];
- uint64_t refcnt = ddp->ddp_refcnt; /* committed refs */
- if (zio != NULL)
- refcnt += zio->io_parent_count; /* pending refs */
- if (ddp == ddp_willref)
- refcnt++; /* caller's ref */
- if (refcnt != 0) {
- total_refcnt += refcnt;
- total_copies += p;
- }
- }
-
- if (ditto == 0 || ditto > UINT32_MAX)
- ditto = UINT32_MAX;
-
- if (total_refcnt >= 1)
- desired_copies++;
- if (total_refcnt >= ditto)
- desired_copies++;
- if (total_refcnt >= ditto * ditto)
- desired_copies++;
-
- return (MAX(desired_copies, total_copies) - total_copies);
-}
-
-int
-ddt_ditto_copies_present(ddt_entry_t *dde)
-{
- ddt_phys_t *ddp = &dde->dde_phys[DDT_PHYS_DITTO];
- dva_t *dva = ddp->ddp_dva;
- int copies = 0 - DVA_GET_GANG(dva);
- int d;
-
- for (d = 0; d < SPA_DVAS_PER_BP; d++, dva++)
- if (DVA_IS_VALID(dva))
- copies++;
-
- ASSERT(copies >= 0 && copies < SPA_DVAS_PER_BP);
-
- return (copies);
-}
-
-size_t
-ddt_compress(void *src, uchar_t *dst, size_t s_len, size_t d_len)
+ddt_enter(ddt_t *ddt)
{
- uchar_t *version = dst++;
- int cpfunc = ZIO_COMPRESS_ZLE;
- zio_compress_info_t *ci = &zio_compress_table[cpfunc];
- size_t c_len;
-
- ASSERT(d_len >= s_len + 1); /* no compression plus version byte */
-
- c_len = ci->ci_compress(src, dst, s_len, d_len - 1, ci->ci_level);
-
- if (c_len == s_len) {
- cpfunc = ZIO_COMPRESS_OFF;
- bcopy(src, dst, s_len);
- }
-
- *version = (ZFS_HOST_BYTEORDER & DDT_COMPRESS_BYTEORDER_MASK) | cpfunc;
-
- return (c_len + 1);
+ mutex_enter(&ddt->ddt_lock);
}
void
-ddt_decompress(uchar_t *src, void *dst, size_t s_len, size_t d_len)
-{
- uchar_t version = *src++;
- int cpfunc = version & DDT_COMPRESS_FUNCTION_MASK;
- zio_compress_info_t *ci = &zio_compress_table[cpfunc];
-
- if (ci->ci_decompress != NULL)
- (void) ci->ci_decompress(src, dst, s_len, d_len, ci->ci_level);
- else
- bcopy(src, dst, d_len);
-
- if ((version ^ ZFS_HOST_BYTEORDER) & DDT_COMPRESS_BYTEORDER_MASK)
- byteswap_uint64_array(dst, d_len);
-}
-
-ddt_t *
-ddt_select_by_checksum(spa_t *spa, enum zio_checksum c)
-{
- return (spa->spa_ddt[c]);
-}
-
-ddt_t *
-ddt_select(spa_t *spa, const blkptr_t *bp)
+ddt_exit(ddt_t *ddt)
{
- return (spa->spa_ddt[BP_GET_CHECKSUM(bp)]);
+ mutex_exit(&ddt->ddt_lock);
}
void
-ddt_enter(ddt_t *ddt)
+ddt_init(void)
{
- mutex_enter(&ddt->ddt_lock);
+ ddt_cache = kmem_cache_create("ddt_cache",
+ sizeof (ddt_t), 0, NULL, NULL, NULL, NULL, NULL, 0);
+ ddt_entry_cache = kmem_cache_create("ddt_entry_cache",
+ sizeof (ddt_entry_t), 0, NULL, NULL, NULL, NULL, NULL, 0);
}
void
-ddt_exit(ddt_t *ddt)
+ddt_fini(void)
{
- mutex_exit(&ddt->ddt_lock);
+ kmem_cache_destroy(ddt_entry_cache);
+ kmem_cache_destroy(ddt_cache);
}
static ddt_entry_t *
{
ddt_entry_t *dde;
- /* XXX: Move to a slab */
- dde = kmem_zalloc(sizeof (ddt_entry_t), KM_PUSHPAGE);
+ dde = kmem_cache_alloc(ddt_entry_cache, KM_SLEEP);
+ memset(dde, 0, sizeof (ddt_entry_t));
cv_init(&dde->dde_cv, NULL, CV_DEFAULT, NULL);
dde->dde_key = *ddk;
static void
ddt_free(ddt_entry_t *dde)
{
- int p;
-
- ASSERT(!dde->dde_loading);
+ ASSERT(dde->dde_flags & DDE_FLAG_LOADED);
- for (p = 0; p < DDT_PHYS_TYPES; p++)
- ASSERT(dde->dde_lead_zio[p] == NULL);
+ for (int p = 0; p < DDT_PHYS_TYPES; p++)
+ ASSERT3P(dde->dde_lead_zio[p], ==, NULL);
- if (dde->dde_repair_data != NULL)
- zio_buf_free(dde->dde_repair_data,
- DDK_GET_PSIZE(&dde->dde_key));
+ if (dde->dde_repair_abd != NULL)
+ abd_free(dde->dde_repair_abd);
cv_destroy(&dde->dde_cv);
- kmem_free(dde, sizeof (*dde));
+ kmem_cache_free(ddt_entry_cache, dde);
}
void
ddt_entry_t *
ddt_lookup(ddt_t *ddt, const blkptr_t *bp, boolean_t add)
{
- ddt_entry_t *dde, dde_search;
- enum ddt_type type;
- enum ddt_class class;
+ ddt_key_t search;
+ ddt_entry_t *dde;
+ ddt_type_t type;
+ ddt_class_t class;
avl_index_t where;
int error;
ASSERT(MUTEX_HELD(&ddt->ddt_lock));
- ddt_key_fill(&dde_search.dde_key, bp);
+ ddt_key_fill(&search, bp);
- dde = avl_find(&ddt->ddt_tree, &dde_search, &where);
- if (dde == NULL) {
- if (!add)
- return (NULL);
- dde = ddt_alloc(&dde_search.dde_key);
- avl_insert(&ddt->ddt_tree, dde, where);
- }
+ /* Find an existing live entry */
+ dde = avl_find(&ddt->ddt_tree, &search, &where);
+ if (dde != NULL) {
+ /* Found it. If it's already loaded, we can just return it. */
+ if (dde->dde_flags & DDE_FLAG_LOADED)
+ return (dde);
- while (dde->dde_loading)
- cv_wait(&dde->dde_cv, &ddt->ddt_lock);
+ /* Someone else is loading it, wait for it. */
+ while (!(dde->dde_flags & DDE_FLAG_LOADED))
+ cv_wait(&dde->dde_cv, &ddt->ddt_lock);
- if (dde->dde_loaded)
return (dde);
+ }
- dde->dde_loading = B_TRUE;
+ /* Not found. */
+ if (!add)
+ return (NULL);
+ /* Time to make a new entry. */
+ dde = ddt_alloc(&search);
+ avl_insert(&ddt->ddt_tree, dde, where);
+
+ /*
+ * ddt_tree is now stable, so unlock and let everyone else keep moving.
+ * Anyone landing on this entry will find it without DDE_FLAG_LOADED,
+ * and go to sleep waiting for it above.
+ */
ddt_exit(ddt);
+ /* Search all store objects for the entry. */
error = ENOENT;
-
for (type = 0; type < DDT_TYPES; type++) {
for (class = 0; class < DDT_CLASSES; class++) {
error = ddt_object_lookup(ddt, type, class, dde);
- if (error != ENOENT)
+ if (error != ENOENT) {
+ ASSERT0(error);
break;
+ }
}
if (error != ENOENT)
break;
}
- ASSERT(error == 0 || error == ENOENT);
-
ddt_enter(ddt);
- ASSERT(dde->dde_loaded == B_FALSE);
- ASSERT(dde->dde_loading == B_TRUE);
+ ASSERT(!(dde->dde_flags & DDE_FLAG_LOADED));
dde->dde_type = type; /* will be DDT_TYPES if no entry found */
dde->dde_class = class; /* will be DDT_CLASSES if no entry found */
- dde->dde_loaded = B_TRUE;
- dde->dde_loading = B_FALSE;
if (error == 0)
ddt_stat_update(ddt, dde, -1ULL);
+ /* Entry loaded, everyone can proceed now */
+ dde->dde_flags |= DDE_FLAG_LOADED;
cv_broadcast(&dde->dde_cv);
return (dde);
ddt_prefetch(spa_t *spa, const blkptr_t *bp)
{
ddt_t *ddt;
- ddt_entry_t dde;
- enum ddt_type type;
- enum ddt_class class;
+ ddt_key_t ddk;
if (!zfs_dedup_prefetch || bp == NULL || !BP_GET_DEDUP(bp))
return;
* Thus no locking is required as the DDT can't disappear on us.
*/
ddt = ddt_select(spa, bp);
- ddt_key_fill(&dde.dde_key, bp);
+ ddt_key_fill(&ddk, bp);
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class < DDT_CLASSES; class++) {
- ddt_object_prefetch(ddt, type, class, &dde);
+ for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
+ for (ddt_class_t class = 0; class < DDT_CLASSES; class++) {
+ ddt_object_prefetch(ddt, type, class, &ddk);
}
}
}
+/*
+ * Key comparison. Any struct wanting to make use of this function must have
+ * the key as the first element.
+ */
+#define DDT_KEY_CMP_LEN (sizeof (ddt_key_t) / sizeof (uint16_t))
+
+typedef struct ddt_key_cmp {
+ uint16_t u16[DDT_KEY_CMP_LEN];
+} ddt_key_cmp_t;
+
int
-ddt_entry_compare(const void *x1, const void *x2)
+ddt_key_compare(const void *x1, const void *x2)
{
- const ddt_entry_t *dde1 = x1;
- const ddt_entry_t *dde2 = x2;
- const uint64_t *u1 = (const uint64_t *)&dde1->dde_key;
- const uint64_t *u2 = (const uint64_t *)&dde2->dde_key;
- int i;
-
- for (i = 0; i < DDT_KEY_WORDS; i++) {
- if (u1[i] < u2[i])
- return (-1);
- if (u1[i] > u2[i])
- return (1);
+ const ddt_key_cmp_t *k1 = (const ddt_key_cmp_t *)x1;
+ const ddt_key_cmp_t *k2 = (const ddt_key_cmp_t *)x2;
+ int32_t cmp = 0;
+
+ for (int i = 0; i < DDT_KEY_CMP_LEN; i++) {
+ cmp = (int32_t)k1->u16[i] - (int32_t)k2->u16[i];
+ if (likely(cmp))
+ break;
}
- return (0);
+ return (TREE_ISIGN(cmp));
}
static ddt_t *
{
ddt_t *ddt;
- /* XXX: Move to a slab */
- ddt = kmem_zalloc(sizeof (*ddt), KM_PUSHPAGE | KM_NODEBUG);
+ ddt = kmem_cache_alloc(ddt_cache, KM_SLEEP);
+ memset(ddt, 0, sizeof (ddt_t));
mutex_init(&ddt->ddt_lock, NULL, MUTEX_DEFAULT, NULL);
- avl_create(&ddt->ddt_tree, ddt_entry_compare,
+ avl_create(&ddt->ddt_tree, ddt_key_compare,
sizeof (ddt_entry_t), offsetof(ddt_entry_t, dde_node));
- avl_create(&ddt->ddt_repair_tree, ddt_entry_compare,
+ avl_create(&ddt->ddt_repair_tree, ddt_key_compare,
sizeof (ddt_entry_t), offsetof(ddt_entry_t, dde_node));
ddt->ddt_checksum = c;
ddt->ddt_spa = spa;
static void
ddt_table_free(ddt_t *ddt)
{
- ASSERT(avl_numnodes(&ddt->ddt_tree) == 0);
- ASSERT(avl_numnodes(&ddt->ddt_repair_tree) == 0);
+ ASSERT0(avl_numnodes(&ddt->ddt_tree));
+ ASSERT0(avl_numnodes(&ddt->ddt_repair_tree));
avl_destroy(&ddt->ddt_tree);
avl_destroy(&ddt->ddt_repair_tree);
mutex_destroy(&ddt->ddt_lock);
- kmem_free(ddt, sizeof (*ddt));
+ kmem_cache_free(ddt_cache, ddt);
}
void
ddt_create(spa_t *spa)
{
- enum zio_checksum c;
-
spa->spa_dedup_checksum = ZIO_DEDUPCHECKSUM;
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++)
- spa->spa_ddt[c] = ddt_table_alloc(spa, c);
+ for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
+ if (DDT_CHECKSUM_VALID(c))
+ spa->spa_ddt[c] = ddt_table_alloc(spa, c);
+ }
}
int
ddt_load(spa_t *spa)
{
- enum zio_checksum c;
- enum ddt_type type;
- enum ddt_class class;
int error;
ddt_create(spa);
if (error)
return (error == ENOENT ? 0 : error);
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
+ for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
+ if (!DDT_CHECKSUM_VALID(c))
+ continue;
+
ddt_t *ddt = spa->spa_ddt[c];
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class < DDT_CLASSES;
+ for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
+ for (ddt_class_t class = 0; class < DDT_CLASSES;
class++) {
error = ddt_object_load(ddt, type, class);
if (error != 0 && error != ENOENT)
/*
* Seed the cached histograms.
*/
- bcopy(ddt->ddt_histogram, &ddt->ddt_histogram_cache,
+ memcpy(&ddt->ddt_histogram_cache, ddt->ddt_histogram,
sizeof (ddt->ddt_histogram));
+ spa->spa_dedup_dspace = ~0ULL;
}
return (0);
void
ddt_unload(spa_t *spa)
{
- enum zio_checksum c;
-
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
+ for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
if (spa->spa_ddt[c]) {
ddt_table_free(spa->spa_ddt[c]);
spa->spa_ddt[c] = NULL;
}
boolean_t
-ddt_class_contains(spa_t *spa, enum ddt_class max_class, const blkptr_t *bp)
+ddt_class_contains(spa_t *spa, ddt_class_t max_class, const blkptr_t *bp)
{
ddt_t *ddt;
- ddt_entry_t *dde;
- enum ddt_type type;
- enum ddt_class class;
+ ddt_key_t ddk;
if (!BP_GET_DEDUP(bp))
return (B_FALSE);
return (B_TRUE);
ddt = spa->spa_ddt[BP_GET_CHECKSUM(bp)];
- dde = kmem_alloc(sizeof (ddt_entry_t), KM_PUSHPAGE);
- ddt_key_fill(&(dde->dde_key), bp);
+ ddt_key_fill(&ddk, bp);
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class <= max_class; class++) {
- if (ddt_object_lookup(ddt, type, class, dde) == 0) {
- kmem_free(dde, sizeof (ddt_entry_t));
+ for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
+ for (ddt_class_t class = 0; class <= max_class; class++) {
+ if (ddt_object_contains(ddt, type, class, &ddk) == 0)
return (B_TRUE);
- }
}
}
- kmem_free(dde, sizeof (ddt_entry_t));
return (B_FALSE);
}
{
ddt_key_t ddk;
ddt_entry_t *dde;
- enum ddt_type type;
- enum ddt_class class;
ddt_key_fill(&ddk, bp);
dde = ddt_alloc(&ddk);
- for (type = 0; type < DDT_TYPES; type++) {
- for (class = 0; class < DDT_CLASSES; class++) {
+ for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
+ for (ddt_class_t class = 0; class < DDT_CLASSES; class++) {
/*
* We can only do repair if there are multiple copies
* of the block. For anything in the UNIQUE class,
}
}
- bzero(dde->dde_phys, sizeof (dde->dde_phys));
+ memset(dde->dde_phys, 0, sizeof (dde->dde_phys));
return (dde);
}
ddt_enter(ddt);
- if (dde->dde_repair_data != NULL && spa_writeable(ddt->ddt_spa) &&
+ if (dde->dde_repair_abd != NULL && spa_writeable(ddt->ddt_spa) &&
avl_find(&ddt->ddt_repair_tree, dde, &where) == NULL)
avl_insert(&ddt->ddt_repair_tree, dde, where);
else
ddt_key_t *rddk = &rdde->dde_key;
zio_t *zio;
blkptr_t blk;
- int p;
zio = zio_null(rio, rio->io_spa, NULL,
ddt_repair_entry_done, rdde, rio->io_flags);
- for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++, rddp++) {
+ for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++, rddp++) {
if (ddp->ddp_phys_birth == 0 ||
ddp->ddp_phys_birth != rddp->ddp_phys_birth ||
- bcmp(ddp->ddp_dva, rddp->ddp_dva, sizeof (ddp->ddp_dva)))
+ memcmp(ddp->ddp_dva, rddp->ddp_dva, sizeof (ddp->ddp_dva)))
continue;
ddt_bp_create(ddt->ddt_checksum, ddk, ddp, &blk);
zio_nowait(zio_rewrite(zio, zio->io_spa, 0, &blk,
- rdde->dde_repair_data, DDK_GET_PSIZE(rddk), NULL, NULL,
+ rdde->dde_repair_abd, DDK_GET_PSIZE(rddk), NULL, NULL,
ZIO_PRIORITY_SYNC_WRITE, ZIO_DDT_CHILD_FLAGS(zio), NULL));
}
dsl_pool_t *dp = ddt->ddt_spa->spa_dsl_pool;
ddt_phys_t *ddp = dde->dde_phys;
ddt_key_t *ddk = &dde->dde_key;
- enum ddt_type otype = dde->dde_type;
- enum ddt_type ntype = DDT_TYPE_CURRENT;
- enum ddt_class oclass = dde->dde_class;
- enum ddt_class nclass;
+ ddt_type_t otype = dde->dde_type;
+ ddt_type_t ntype = DDT_TYPE_DEFAULT;
+ ddt_class_t oclass = dde->dde_class;
+ ddt_class_t nclass;
uint64_t total_refcnt = 0;
- int p;
- ASSERT(dde->dde_loaded);
- ASSERT(!dde->dde_loading);
+ ASSERT(dde->dde_flags & DDE_FLAG_LOADED);
- for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
- ASSERT(dde->dde_lead_zio[p] == NULL);
+ for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
+ ASSERT3P(dde->dde_lead_zio[p], ==, NULL);
if (ddp->ddp_phys_birth == 0) {
- ASSERT(ddp->ddp_refcnt == 0);
+ ASSERT0(ddp->ddp_refcnt);
continue;
}
if (p == DDT_PHYS_DITTO) {
- if (ddt_ditto_copies_needed(ddt, dde, NULL) == 0)
- ddt_phys_free(ddt, ddk, ddp, txg);
+ /*
+ * Note, we no longer create DDT-DITTO blocks, but we
+ * don't want to leak any written by older software.
+ */
+ ddt_phys_free(ddt, ddk, ddp, txg);
continue;
}
if (ddp->ddp_refcnt == 0)
total_refcnt += ddp->ddp_refcnt;
}
- if (dde->dde_phys[DDT_PHYS_DITTO].ddp_phys_birth != 0)
- nclass = DDT_CLASS_DITTO;
- else if (total_refcnt > 1)
+ /* We do not create new DDT-DITTO blocks. */
+ ASSERT0(dde->dde_phys[DDT_PHYS_DITTO].ddp_phys_birth);
+ if (total_refcnt > 1)
nclass = DDT_CLASS_DUPLICATE;
else
nclass = DDT_CLASS_UNIQUE;
if (otype != DDT_TYPES &&
(otype != ntype || oclass != nclass || total_refcnt == 0)) {
- VERIFY(ddt_object_remove(ddt, otype, oclass, dde, tx) == 0);
- ASSERT(ddt_object_lookup(ddt, otype, oclass, dde) == ENOENT);
+ VERIFY0(ddt_object_remove(ddt, otype, oclass, ddk, tx));
+ ASSERT3U(
+ ddt_object_contains(ddt, otype, oclass, ddk), ==, ENOENT);
}
if (total_refcnt != 0) {
ddt_stat_update(ddt, dde, 0);
if (!ddt_object_exists(ddt, ntype, nclass))
ddt_object_create(ddt, ntype, nclass, tx);
- VERIFY(ddt_object_update(ddt, ntype, nclass, dde, tx) == 0);
+ VERIFY0(ddt_object_update(ddt, ntype, nclass, dde, tx));
/*
* If the class changes, the order that we scan this bp
spa_t *spa = ddt->ddt_spa;
ddt_entry_t *dde;
void *cookie = NULL;
- enum ddt_type type;
- enum ddt_class class;
if (avl_numnodes(&ddt->ddt_tree) == 0)
return;
- ASSERT(spa->spa_uberblock.ub_version >= SPA_VERSION_DEDUP);
+ ASSERT3U(spa->spa_uberblock.ub_version, >=, SPA_VERSION_DEDUP);
if (spa->spa_ddt_stat_object == 0) {
spa->spa_ddt_stat_object = zap_create_link(ddt->ddt_os,
ddt_free(dde);
}
- for (type = 0; type < DDT_TYPES; type++) {
+ for (ddt_type_t type = 0; type < DDT_TYPES; type++) {
uint64_t add, count = 0;
- for (class = 0; class < DDT_CLASSES; class++) {
+ for (ddt_class_t class = 0; class < DDT_CLASSES; class++) {
if (ddt_object_exists(ddt, type, class)) {
ddt_object_sync(ddt, type, class, tx);
- VERIFY(ddt_object_count(ddt, type, class,
- &add) == 0);
+ VERIFY0(ddt_object_count(ddt, type, class,
+ &add));
count += add;
}
}
- for (class = 0; class < DDT_CLASSES; class++) {
+ for (ddt_class_t class = 0; class < DDT_CLASSES; class++) {
if (count == 0 && ddt_object_exists(ddt, type, class))
ddt_object_destroy(ddt, type, class, tx);
}
}
- bcopy(ddt->ddt_histogram, &ddt->ddt_histogram_cache,
+ memcpy(&ddt->ddt_histogram_cache, ddt->ddt_histogram,
sizeof (ddt->ddt_histogram));
+ spa->spa_dedup_dspace = ~0ULL;
}
void
ddt_sync(spa_t *spa, uint64_t txg)
{
+ dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
dmu_tx_t *tx;
- zio_t *rio = zio_root(spa, NULL, NULL,
- ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE);
- enum zio_checksum c;
+ zio_t *rio;
- ASSERT(spa_syncing_txg(spa) == txg);
+ ASSERT3U(spa_syncing_txg(spa), ==, txg);
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
- for (c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
+ rio = zio_root(spa, NULL, NULL,
+ ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SELF_HEAL);
+
+ /*
+ * This function may cause an immediate scan of ddt blocks (see
+ * the comment above dsl_scan_ddt() for details). We set the
+ * scan's root zio here so that we can wait for any scan IOs in
+ * addition to the regular ddt IOs.
+ */
+ ASSERT3P(scn->scn_zio_root, ==, NULL);
+ scn->scn_zio_root = rio;
+
+ for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
ddt_t *ddt = spa->spa_ddt[c];
if (ddt == NULL)
continue;
}
(void) zio_wait(rio);
+ scn->scn_zio_root = NULL;
dmu_tx_commit(tx);
}
do {
do {
ddt_t *ddt = spa->spa_ddt[ddb->ddb_checksum];
+ if (ddt == NULL)
+ continue;
int error = ENOENT;
if (ddt_object_exists(ddt, ddb->ddb_type,
ddb->ddb_class)) {
return (SET_ERROR(ENOENT));
}
-#if defined(_KERNEL) && defined(HAVE_SPL)
-module_param(zfs_dedup_prefetch, int, 0644);
-MODULE_PARM_DESC(zfs_dedup_prefetch, "Enable prefetching dedup-ed blks");
-#endif
+/*
+ * This function is used by Block Cloning (brt.c) to increase reference
+ * counter for the DDT entry if the block is already in DDT.
+ *
+ * Return false if the block, despite having the D bit set, is not present
+ * in the DDT. Currently this is not possible but might be in the future.
+ * See the comment below.
+ */
+boolean_t
+ddt_addref(spa_t *spa, const blkptr_t *bp)
+{
+ ddt_t *ddt;
+ ddt_entry_t *dde;
+ boolean_t result;
+
+ spa_config_enter(spa, SCL_ZIO, FTAG, RW_READER);
+ ddt = ddt_select(spa, bp);
+ ddt_enter(ddt);
+
+ dde = ddt_lookup(ddt, bp, B_TRUE);
+ ASSERT3P(dde, !=, NULL);
+
+ if (dde->dde_type < DDT_TYPES) {
+ ddt_phys_t *ddp;
+
+ ASSERT3S(dde->dde_class, <, DDT_CLASSES);
+
+ ddp = &dde->dde_phys[BP_GET_NDVAS(bp)];
+
+ /*
+ * This entry already existed (dde_type is real), so it must
+ * have refcnt >0 at the start of this txg. We are called from
+ * brt_pending_apply(), before frees are issued, so the refcnt
+ * can't be lowered yet. Therefore, it must be >0. We assert
+ * this because if the order of BRT and DDT interactions were
+ * ever to change and the refcnt was ever zero here, then
+ * likely further action is required to fill out the DDT entry,
+ * and this is a place that is likely to be missed in testing.
+ */
+ ASSERT3U(ddp->ddp_refcnt, >, 0);
+
+ ddt_phys_addref(ddp);
+ result = B_TRUE;
+ } else {
+ /*
+ * At the time of implementating this if the block has the
+ * DEDUP flag set it must exist in the DEDUP table, but
+ * there are many advocates that want ability to remove
+ * entries from DDT with refcnt=1. If this will happen,
+ * we may have a block with the DEDUP set, but which doesn't
+ * have a corresponding entry in the DDT. Be ready.
+ */
+ ASSERT3S(dde->dde_class, ==, DDT_CLASSES);
+ ddt_remove(ddt, dde);
+ result = B_FALSE;
+ }
+
+ ddt_exit(ddt);
+ spa_config_exit(spa, SCL_ZIO, FTAG);
+
+ return (result);
+}
+
+ZFS_MODULE_PARAM(zfs_dedup, zfs_dedup_, prefetch, INT, ZMOD_RW,
+ "Enable prefetching dedup-ed blks");
projects / mirror_zfs.git / blobdiff