2008-11-20 23:01:55 +03:00
|
|
|
/*
|
|
|
|
* 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
|
2022-07-12 00:16:13 +03:00
|
|
|
* or https://opensource.org/licenses/CDDL-1.0.
|
2008-11-20 23:01:55 +03:00
|
|
|
* 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
|
|
|
|
*/
|
2017-02-03 01:13:41 +03:00
|
|
|
|
2008-11-20 23:01:55 +03:00
|
|
|
/*
|
2010-05-29 00:45:14 +04:00
|
|
|
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
|
2015-04-02 06:44:32 +03:00
|
|
|
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
|
2016-07-21 01:39:55 +03:00
|
|
|
* Copyright (c) 2013, 2016 by Delphix. All rights reserved.
|
2017-02-03 01:13:41 +03:00
|
|
|
* Copyright 2017 Nexenta Systems, Inc.
|
feature: large_microzap
In a4b21eadec we added the zap_micro_max_size tuneable to raise the size
at which "micro" (single-block) ZAPs are upgraded to "fat" (multi-block)
ZAPs. Before this, a microZAP was limited to 128KiB, which was the old
largest block size. The side effect of raising the max size past 128KiB
is that it be stored in a large block, requiring the large_blocks
feature.
Unfortunately, this means that a backup stream created without the
--large-block (-L) flag to zfs send would split the microZAP block into
smaller blocks and send those, as is normal behaviour for large blocks.
This would be received correctly, but since microZAPs are limited to the
first block in the object by definition, the entries in the later blocks
would be inaccessible. For directory ZAPs, this gives the appearance of
files being lost.
This commit adds a feature flag, large_microzap, that must be enabled
for microZAPs to grow beyond 128KiB, and which will be activated the
first time that occurs. This feature is later checked when generating
the stream and if active, the send operation will abort unless
--large-block has also been requested.
Changing the limit still requires zap_micro_max_size to be changed. The
state of this flag effectively sets the upper value for this tuneable,
that is, if the feature is disabled, the tuneable will be clamped to
128KiB.
A stream flag is also added to ensure that the receiver also activates
its own feature flag upon receiving the stream. This is not strictly
necessary to _use_ the received microZAP, since it doesn't care how
large its block is, but it is required to send the microZAP object on,
otherwise the original problem occurs again.
Because it's difficult to reliably distinguish a microZAP from a fatZAP
from outside the ZAP code, and because it seems unlikely that most
users are affected (a fairly niche tuneable combined with what should be
an uncommon use of send), and for the sake of expediency, this change
activates the feature the first time a microZAP grows to use a large
block, and is never deactivated after that. This can be improved in the
future.
This commit changes nothing for existing pools that already have large
microZAPs. The feature will not be retroactively applied, but will be
activated the next time a microZAP grows past the limit.
Don't use large_blocks feature for enable/disable tests. The
large_microzap depends on large_blocks, so it gets enabled as a
dependency, breaking the test. Instead use feature "longname", which has
the exact same feature characteristics.
Sponsored-by: Klara, Inc.
Sponsored-by: Wasabi Technology, Inc.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Rob Norris <rob.norris@klarasystems.com>
Closes #16593
2024-10-03 06:47:11 +03:00
|
|
|
* Copyright (c) 2024, Klara, Inc.
|
2008-11-20 23:01:55 +03:00
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef _SYS_ZAP_IMPL_H
|
|
|
|
#define _SYS_ZAP_IMPL_H
|
|
|
|
|
|
|
|
#include <sys/zap.h>
|
|
|
|
#include <sys/zfs_context.h>
|
|
|
|
#include <sys/avl.h>
|
|
|
|
|
|
|
|
#ifdef __cplusplus
|
|
|
|
extern "C" {
|
|
|
|
#endif
|
|
|
|
|
|
|
|
extern int fzap_default_block_shift;
|
|
|
|
|
|
|
|
#define ZAP_MAGIC 0x2F52AB2ABULL
|
|
|
|
|
|
|
|
#define FZAP_BLOCK_SHIFT(zap) ((zap)->zap_f.zap_block_shift)
|
|
|
|
|
|
|
|
#define MZAP_ENT_LEN 64
|
|
|
|
#define MZAP_NAME_LEN (MZAP_ENT_LEN - 8 - 4 - 2)
|
|
|
|
|
2010-05-29 00:45:14 +04:00
|
|
|
#define ZAP_NEED_CD (-1U)
|
|
|
|
|
2008-11-20 23:01:55 +03:00
|
|
|
typedef struct mzap_ent_phys {
|
|
|
|
uint64_t mze_value;
|
|
|
|
uint32_t mze_cd;
|
|
|
|
uint16_t mze_pad; /* in case we want to chain them someday */
|
|
|
|
char mze_name[MZAP_NAME_LEN];
|
|
|
|
} mzap_ent_phys_t;
|
|
|
|
|
|
|
|
typedef struct mzap_phys {
|
|
|
|
uint64_t mz_block_type; /* ZBT_MICRO */
|
|
|
|
uint64_t mz_salt;
|
|
|
|
uint64_t mz_normflags;
|
|
|
|
uint64_t mz_pad[5];
|
|
|
|
mzap_ent_phys_t mz_chunk[1];
|
|
|
|
/* actually variable size depending on block size */
|
|
|
|
} mzap_phys_t;
|
|
|
|
|
|
|
|
typedef struct mzap_ent {
|
2022-10-20 21:57:15 +03:00
|
|
|
uint32_t mze_hash;
|
|
|
|
uint16_t mze_cd; /* copy from mze_phys->mze_cd */
|
|
|
|
uint16_t mze_chunkid;
|
2008-11-20 23:01:55 +03:00
|
|
|
} mzap_ent_t;
|
|
|
|
|
2010-05-29 00:45:14 +04:00
|
|
|
#define MZE_PHYS(zap, mze) \
|
2015-04-01 18:14:34 +03:00
|
|
|
(&zap_m_phys(zap)->mz_chunk[(mze)->mze_chunkid])
|
2008-11-20 23:01:55 +03:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The (fat) zap is stored in one object. It is an array of
|
|
|
|
* 1<<FZAP_BLOCK_SHIFT byte blocks. The layout looks like one of:
|
|
|
|
*
|
|
|
|
* ptrtbl fits in first block:
|
|
|
|
* [zap_phys_t zap_ptrtbl_shift < 6] [zap_leaf_t] ...
|
|
|
|
*
|
|
|
|
* ptrtbl too big for first block:
|
|
|
|
* [zap_phys_t zap_ptrtbl_shift >= 6] [zap_leaf_t] [ptrtbl] ...
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
struct dmu_buf;
|
|
|
|
struct zap_leaf;
|
|
|
|
|
|
|
|
#define ZBT_LEAF ((1ULL << 63) + 0)
|
|
|
|
#define ZBT_HEADER ((1ULL << 63) + 1)
|
|
|
|
#define ZBT_MICRO ((1ULL << 63) + 3)
|
|
|
|
/* any other values are ptrtbl blocks */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* the embedded pointer table takes up half a block:
|
|
|
|
* block size / entry size (2^3) / 2
|
|
|
|
*/
|
|
|
|
#define ZAP_EMBEDDED_PTRTBL_SHIFT(zap) (FZAP_BLOCK_SHIFT(zap) - 3 - 1)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The embedded pointer table starts half-way through the block. Since
|
|
|
|
* the pointer table itself is half the block, it starts at (64-bit)
|
|
|
|
* word number (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap)).
|
|
|
|
*/
|
|
|
|
#define ZAP_EMBEDDED_PTRTBL_ENT(zap, idx) \
|
2015-04-01 18:14:34 +03:00
|
|
|
((uint64_t *)zap_f_phys(zap)) \
|
2008-11-20 23:01:55 +03:00
|
|
|
[(idx) + (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap))]
|
|
|
|
|
|
|
|
/*
|
|
|
|
* TAKE NOTE:
|
|
|
|
* If zap_phys_t is modified, zap_byteswap() must be modified.
|
|
|
|
*/
|
|
|
|
typedef struct zap_phys {
|
|
|
|
uint64_t zap_block_type; /* ZBT_HEADER */
|
|
|
|
uint64_t zap_magic; /* ZAP_MAGIC */
|
|
|
|
|
|
|
|
struct zap_table_phys {
|
|
|
|
uint64_t zt_blk; /* starting block number */
|
|
|
|
uint64_t zt_numblks; /* number of blocks */
|
|
|
|
uint64_t zt_shift; /* bits to index it */
|
|
|
|
uint64_t zt_nextblk; /* next (larger) copy start block */
|
|
|
|
uint64_t zt_blks_copied; /* number source blocks copied */
|
|
|
|
} zap_ptrtbl;
|
|
|
|
|
|
|
|
uint64_t zap_freeblk; /* the next free block */
|
|
|
|
uint64_t zap_num_leafs; /* number of leafs */
|
|
|
|
uint64_t zap_num_entries; /* number of entries */
|
|
|
|
uint64_t zap_salt; /* salt to stir into hash function */
|
|
|
|
uint64_t zap_normflags; /* flags for u8_textprep_str() */
|
2010-05-29 00:45:14 +04:00
|
|
|
uint64_t zap_flags; /* zap_flags_t */
|
2008-11-20 23:01:55 +03:00
|
|
|
/*
|
|
|
|
* This structure is followed by padding, and then the embedded
|
|
|
|
* pointer table. The embedded pointer table takes up second
|
|
|
|
* half of the block. It is accessed using the
|
|
|
|
* ZAP_EMBEDDED_PTRTBL_ENT() macro.
|
|
|
|
*/
|
|
|
|
} zap_phys_t;
|
|
|
|
|
|
|
|
typedef struct zap_table_phys zap_table_phys_t;
|
|
|
|
|
|
|
|
typedef struct zap {
|
2015-04-02 06:44:32 +03:00
|
|
|
dmu_buf_user_t zap_dbu;
|
2008-11-20 23:01:55 +03:00
|
|
|
objset_t *zap_objset;
|
|
|
|
uint64_t zap_object;
|
2024-03-26 00:58:50 +03:00
|
|
|
dnode_t *zap_dnode;
|
2008-11-20 23:01:55 +03:00
|
|
|
struct dmu_buf *zap_dbuf;
|
|
|
|
krwlock_t zap_rwlock;
|
|
|
|
boolean_t zap_ismicro;
|
|
|
|
int zap_normflags;
|
|
|
|
uint64_t zap_salt;
|
|
|
|
union {
|
|
|
|
struct {
|
|
|
|
/*
|
|
|
|
* zap_num_entries_mtx protects
|
|
|
|
* zap_num_entries
|
|
|
|
*/
|
|
|
|
kmutex_t zap_num_entries_mtx;
|
|
|
|
int zap_block_shift;
|
|
|
|
} zap_fat;
|
|
|
|
struct {
|
|
|
|
int16_t zap_num_entries;
|
|
|
|
int16_t zap_num_chunks;
|
|
|
|
int16_t zap_alloc_next;
|
2022-10-20 21:57:15 +03:00
|
|
|
zfs_btree_t zap_tree;
|
2008-11-20 23:01:55 +03:00
|
|
|
} zap_micro;
|
|
|
|
} zap_u;
|
|
|
|
} zap_t;
|
|
|
|
|
2015-04-01 18:14:34 +03:00
|
|
|
static inline zap_phys_t *
|
|
|
|
zap_f_phys(zap_t *zap)
|
|
|
|
{
|
|
|
|
return (zap->zap_dbuf->db_data);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline mzap_phys_t *
|
|
|
|
zap_m_phys(zap_t *zap)
|
|
|
|
{
|
|
|
|
return (zap->zap_dbuf->db_data);
|
|
|
|
}
|
|
|
|
|
2008-11-20 23:01:55 +03:00
|
|
|
typedef struct zap_name {
|
|
|
|
zap_t *zn_zap;
|
2010-05-29 00:45:14 +04:00
|
|
|
int zn_key_intlen;
|
|
|
|
const void *zn_key_orig;
|
|
|
|
int zn_key_orig_numints;
|
|
|
|
const void *zn_key_norm;
|
|
|
|
int zn_key_norm_numints;
|
2008-11-20 23:01:55 +03:00
|
|
|
uint64_t zn_hash;
|
|
|
|
matchtype_t zn_matchtype;
|
2017-02-03 01:13:41 +03:00
|
|
|
int zn_normflags;
|
2021-02-02 16:54:15 +03:00
|
|
|
int zn_normbuf_len;
|
|
|
|
char zn_normbuf[];
|
2008-11-20 23:01:55 +03:00
|
|
|
} zap_name_t;
|
|
|
|
|
|
|
|
#define zap_f zap_u.zap_fat
|
|
|
|
#define zap_m zap_u.zap_micro
|
|
|
|
|
|
|
|
boolean_t zap_match(zap_name_t *zn, const char *matchname);
|
|
|
|
int zap_lockdir(objset_t *os, uint64_t obj, dmu_tx_t *tx,
|
2022-04-19 21:49:30 +03:00
|
|
|
krw_t lti, boolean_t fatreader, boolean_t adding, const void *tag,
|
|
|
|
zap_t **zapp);
|
|
|
|
void zap_unlockdir(zap_t *zap, const void *tag);
|
2017-01-27 01:43:28 +03:00
|
|
|
void zap_evict_sync(void *dbu);
|
2022-10-20 21:57:15 +03:00
|
|
|
zap_name_t *zap_name_alloc_str(zap_t *zap, const char *key, matchtype_t mt);
|
2008-11-20 23:01:55 +03:00
|
|
|
void zap_name_free(zap_name_t *zn);
|
2010-05-29 00:45:14 +04:00
|
|
|
int zap_hashbits(zap_t *zap);
|
|
|
|
uint32_t zap_maxcd(zap_t *zap);
|
|
|
|
uint64_t zap_getflags(zap_t *zap);
|
2008-11-20 23:01:55 +03:00
|
|
|
|
feature: large_microzap
In a4b21eadec we added the zap_micro_max_size tuneable to raise the size
at which "micro" (single-block) ZAPs are upgraded to "fat" (multi-block)
ZAPs. Before this, a microZAP was limited to 128KiB, which was the old
largest block size. The side effect of raising the max size past 128KiB
is that it be stored in a large block, requiring the large_blocks
feature.
Unfortunately, this means that a backup stream created without the
--large-block (-L) flag to zfs send would split the microZAP block into
smaller blocks and send those, as is normal behaviour for large blocks.
This would be received correctly, but since microZAPs are limited to the
first block in the object by definition, the entries in the later blocks
would be inaccessible. For directory ZAPs, this gives the appearance of
files being lost.
This commit adds a feature flag, large_microzap, that must be enabled
for microZAPs to grow beyond 128KiB, and which will be activated the
first time that occurs. This feature is later checked when generating
the stream and if active, the send operation will abort unless
--large-block has also been requested.
Changing the limit still requires zap_micro_max_size to be changed. The
state of this flag effectively sets the upper value for this tuneable,
that is, if the feature is disabled, the tuneable will be clamped to
128KiB.
A stream flag is also added to ensure that the receiver also activates
its own feature flag upon receiving the stream. This is not strictly
necessary to _use_ the received microZAP, since it doesn't care how
large its block is, but it is required to send the microZAP object on,
otherwise the original problem occurs again.
Because it's difficult to reliably distinguish a microZAP from a fatZAP
from outside the ZAP code, and because it seems unlikely that most
users are affected (a fairly niche tuneable combined with what should be
an uncommon use of send), and for the sake of expediency, this change
activates the feature the first time a microZAP grows to use a large
block, and is never deactivated after that. This can be improved in the
future.
This commit changes nothing for existing pools that already have large
microZAPs. The feature will not be retroactively applied, but will be
activated the next time a microZAP grows past the limit.
Don't use large_blocks feature for enable/disable tests. The
large_microzap depends on large_blocks, so it gets enabled as a
dependency, breaking the test. Instead use feature "longname", which has
the exact same feature characteristics.
Sponsored-by: Klara, Inc.
Sponsored-by: Wasabi Technology, Inc.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Rob Norris <rob.norris@klarasystems.com>
Closes #16593
2024-10-03 06:47:11 +03:00
|
|
|
uint64_t zap_get_micro_max_size(spa_t *spa);
|
|
|
|
|
2008-11-20 23:01:55 +03:00
|
|
|
#define ZAP_HASH_IDX(hash, n) (((n) == 0) ? 0 : ((hash) >> (64 - (n))))
|
|
|
|
|
|
|
|
void fzap_byteswap(void *buf, size_t size);
|
|
|
|
int fzap_count(zap_t *zap, uint64_t *count);
|
|
|
|
int fzap_lookup(zap_name_t *zn,
|
|
|
|
uint64_t integer_size, uint64_t num_integers, void *buf,
|
|
|
|
char *realname, int rn_len, boolean_t *normalization_conflictp);
|
2010-05-29 00:45:14 +04:00
|
|
|
void fzap_prefetch(zap_name_t *zn);
|
2008-11-20 23:01:55 +03:00
|
|
|
int fzap_add(zap_name_t *zn, uint64_t integer_size, uint64_t num_integers,
|
2022-04-19 21:49:30 +03:00
|
|
|
const void *val, const void *tag, dmu_tx_t *tx);
|
2008-11-20 23:01:55 +03:00
|
|
|
int fzap_update(zap_name_t *zn,
|
2016-07-21 01:39:55 +03:00
|
|
|
int integer_size, uint64_t num_integers, const void *val,
|
2022-04-19 21:49:30 +03:00
|
|
|
const void *tag, dmu_tx_t *tx);
|
2008-11-20 23:01:55 +03:00
|
|
|
int fzap_length(zap_name_t *zn,
|
|
|
|
uint64_t *integer_size, uint64_t *num_integers);
|
|
|
|
int fzap_remove(zap_name_t *zn, dmu_tx_t *tx);
|
|
|
|
int fzap_cursor_retrieve(zap_t *zap, zap_cursor_t *zc, zap_attribute_t *za);
|
|
|
|
void fzap_get_stats(zap_t *zap, zap_stats_t *zs);
|
|
|
|
void zap_put_leaf(struct zap_leaf *l);
|
|
|
|
|
|
|
|
int fzap_add_cd(zap_name_t *zn,
|
|
|
|
uint64_t integer_size, uint64_t num_integers,
|
2022-04-19 21:49:30 +03:00
|
|
|
const void *val, uint32_t cd, const void *tag, dmu_tx_t *tx);
|
2010-05-29 00:45:14 +04:00
|
|
|
void fzap_upgrade(zap_t *zap, dmu_tx_t *tx, zap_flags_t flags);
|
2008-11-20 23:01:55 +03:00
|
|
|
|
|
|
|
#ifdef __cplusplus
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#endif /* _SYS_ZAP_IMPL_H */
|