mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-20 22:12:30 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/storage/innobase/include/btr0btr.h
Marko Mäkelä a4948dafcd MDEV-11369 Instant ADD COLUMN for InnoDB 
For InnoDB tables, adding, dropping and reordering columns has
required a rebuild of the table and all its indexes. Since MySQL 5.6
(and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing
concurrent modification of the tables.

This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT
and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously,
with only minor changes performed to the table structure. The counter
innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS
is incremented whenever a table rebuild operation is converted into
an instant ADD COLUMN operation.

ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN.

Some usability limitations will be addressed in subsequent work:

MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY
and ALGORITHM=INSTANT
MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE

The format of the clustered index (PRIMARY KEY) is changed as follows:

(1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT,
and a new field PAGE_INSTANT will contain the original number of fields
in the clustered index ('core' fields).
If instant ADD COLUMN has not been used or the table becomes empty,
or the very first instant ADD COLUMN operation is rolled back,
the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset
to 0 and FIL_PAGE_INDEX.

(2) A special 'default row' record is inserted into the leftmost leaf,
between the page infimum and the first user record. This record is
distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the
same format as records that contain values for the instantly added
columns. This 'default row' always has the same number of fields as
the clustered index according to the table definition. The values of
'core' fields are to be ignored. For other fields, the 'default row'
will contain the default values as they were during the ALTER TABLE
statement. (If the column default values are changed later, those
values will only be stored in the .frm file. The 'default row' will
contain the original evaluated values, which must be the same for
every row.) The 'default row' must be completely hidden from
higher-level access routines. Assertions have been added to ensure
that no 'default row' is ever present in the adaptive hash index
or in locked records. The 'default row' is never delete-marked.

(3) In clustered index leaf page records, the number of fields must
reside between the number of 'core' fields (dict_index_t::n_core_fields
introduced in this work) and dict_index_t::n_fields. If the number
of fields is less than dict_index_t::n_fields, the missing fields
are replaced with the column value of the 'default row'.
Note: The number of fields in the record may shrink if some of the
last instantly added columns are updated to the value that is
in the 'default row'. The function btr_cur_trim() implements this
'compression' on update and rollback; dtuple::trim() implements it
on insert.

(4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new
status value REC_STATUS_COLUMNS_ADDED will indicate the presence of
a new record header that will encode n_fields-n_core_fields-1 in
1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header
always explicitly encodes the number of fields.)

We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for
covering the insert of the 'default row' record when instant ADD COLUMN
is used for the first time. Subsequent instant ADD COLUMN can use
TRX_UNDO_UPD_EXIST_REC.

This is joint work with Vin Chen (陈福荣) from Tencent. The design
that was discussed in April 2017 would not have allowed import or
export of data files, because instead of the 'default row' it would
have introduced a data dictionary table. The test
rpl.rpl_alter_instant is exactly as contributed in pull request #408.
The test innodb.instant_alter is based on a contributed test.

The redo log record format changes for ROW_FORMAT=DYNAMIC and
ROW_FORMAT=COMPACT are as contributed. (With this change present,
crash recovery from MariaDB 10.3.1 will fail in spectacular ways!)
Also the semantics of higher-level redo log records that modify the
PAGE_INSTANT field is changed. The redo log format version identifier
was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1.

Everything else has been rewritten by me. Thanks to Elena Stepanova,
the code has been tested extensively.

When rolling back an instant ADD COLUMN operation, we must empty the
PAGE_FREE list after deleting or shortening the 'default row' record,
by calling either btr_page_empty() or btr_page_reorganize(). We must
know the size of each entry in the PAGE_FREE list. If rollback left a
freed copy of the 'default row' in the PAGE_FREE list, we would be
unable to determine its size (if it is in ROW_FORMAT=COMPACT or
ROW_FORMAT=DYNAMIC) because it would contain more fields than the
rolled-back definition of the clustered index.

UNIV_SQL_DEFAULT: A new special constant that designates an instantly
added column that is not present in the clustered index record.

len_is_stored(): Check if a length is an actual length. There are
two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL.

dict_col_t::def_val: The 'default row' value of the column.  If the
column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT.

dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(),
instant_value().

dict_col_t::remove_instant(): Remove the 'instant ADD' status of
a column.

dict_col_t::name(const dict_table_t& table): Replaces
dict_table_get_col_name().

dict_index_t::n_core_fields: The original number of fields.
For secondary indexes and if instant ADD COLUMN has not been used,
this will be equal to dict_index_t::n_fields.

dict_index_t::n_core_null_bytes: Number of bytes needed to
represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable).

dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that
n_core_null_bytes was not initialized yet from the clustered index
root page.

dict_index_t: Add the accessors is_instant(), is_clust(),
get_n_nullable(), instant_field_value().

dict_index_t::instant_add_field(): Adjust clustered index metadata
for instant ADD COLUMN.

dict_index_t::remove_instant(): Remove the 'instant ADD' status
of a clustered index when the table becomes empty, or the very first
instant ADD COLUMN operation is rolled back.

dict_table_t: Add the accessors is_instant(), is_temporary(),
supports_instant().

dict_table_t::instant_add_column(): Adjust metadata for
instant ADD COLUMN.

dict_table_t::rollback_instant(): Adjust metadata on the rollback
of instant ADD COLUMN.

prepare_inplace_alter_table_dict(): First create the ctx->new_table,
and only then decide if the table really needs to be rebuilt.
We must split the creation of table or index metadata from the
creation of the dictionary table records and the creation of
the data. In this way, we can transform a table-rebuilding operation
into an instant ADD COLUMN operation. Dictionary objects will only
be added to cache when table rebuilding or index creation is needed.
The ctx->instant_table will never be added to cache.

dict_table_t::add_to_cache(): Modified and renamed from
dict_table_add_to_cache(). Do not modify the table metadata.
Let the callers invoke dict_table_add_system_columns() and if needed,
set can_be_evicted.

dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the
system columns (which will now exist in the dict_table_t object
already at this point).

dict_create_table_step(): Expect the callers to invoke
dict_table_add_system_columns().

pars_create_table(): Before creating the table creation execution
graph, invoke dict_table_add_system_columns().

row_create_table_for_mysql(): Expect all callers to invoke
dict_table_add_system_columns().

create_index_dict(): Replaces row_merge_create_index_graph().

innodb_update_n_cols(): Renamed from innobase_update_n_virtual().
Call my_error() if an error occurs.

btr_cur_instant_init(), btr_cur_instant_init_low(),
btr_cur_instant_root_init():
Load additional metadata from the clustered index and set
dict_index_t::n_core_null_bytes. This is invoked
when table metadata is first loaded into the data dictionary.

dict_boot(): Initialize n_core_null_bytes for the four hard-coded
dictionary tables.

dict_create_index_step(): Initialize n_core_null_bytes. This is
executed as part of CREATE TABLE.

dict_index_build_internal_clust(): Initialize n_core_null_bytes to
NO_CORE_NULL_BYTES if table->supports_instant().

row_create_index_for_mysql(): Initialize n_core_null_bytes for
CREATE TEMPORARY TABLE.

commit_cache_norebuild(): Call the code to rename or enlarge columns
in the cache only if instant ADD COLUMN is not being used.
(Instant ADD COLUMN would copy all column metadata from
instant_table to old_table, including the names and lengths.)

PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields.
This is repurposing the 16-bit field PAGE_DIRECTION, of which only the
least significant 3 bits were used. The original byte containing
PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B.

page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT.

page_ptr_get_direction(), page_get_direction(),
page_ptr_set_direction(): Accessors for PAGE_DIRECTION.

page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION.

page_direction_increment(): Increment PAGE_N_DIRECTION
and set PAGE_DIRECTION.

rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes,
and assume that heap_no is always set.
Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records,
even if the record contains fewer fields.

rec_offs_make_valid(): Add the parameter 'leaf'.

rec_copy_prefix_to_dtuple(): Assert that the tuple is only built
on the core fields. Instant ADD COLUMN only applies to the
clustered index, and we should never build a search key that has
more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR.
All these columns are always present.

dict_index_build_data_tuple(): Remove assertions that would be
duplicated in rec_copy_prefix_to_dtuple().

rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose
number of fields is between n_core_fields and n_fields.

cmp_rec_rec_with_match(): Implement the comparison between two
MIN_REC_FLAG records.

trx_t::in_rollback: Make the field available in non-debug builds.

trx_start_for_ddl_low(): Remove dangerous error-tolerance.
A dictionary transaction must be flagged as such before it has generated
any undo log records. This is because trx_undo_assign_undo() will mark
the transaction as a dictionary transaction in the undo log header
right before the very first undo log record is being written.

btr_index_rec_validate(): Account for instant ADD COLUMN

row_undo_ins_remove_clust_rec(): On the rollback of an insert into
SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the
last column from the table and the clustered index.

row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(),
trx_undo_update_rec_get_update(): Handle the 'default row'
as a special case.

dtuple_t::trim(index): Omit a redundant suffix of an index tuple right
before insert or update. After instant ADD COLUMN, if the last fields
of a clustered index tuple match the 'default row', there is no
need to store them. While trimming the entry, we must hold a page latch,
so that the table cannot be emptied and the 'default row' be deleted.

btr_cur_optimistic_update(), btr_cur_pessimistic_update(),
row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low():
Invoke dtuple_t::trim() if needed.

row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling
row_ins_clust_index_entry_low().

rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number
of fields to be between n_core_fields and n_fields. Do not support
infimum,supremum. They are never supposed to be stored in dtuple_t,
because page creation nowadays uses a lower-level method for initializing
them.

rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the
number of fields.

btr_cur_trim(): In an update, trim the index entry as needed. For the
'default row', handle rollback specially. For user records, omit
fields that match the 'default row'.

btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete():
Skip locking and adaptive hash index for the 'default row'.

row_log_table_apply_convert_mrec(): Replace 'default row' values if needed.
In the temporary file that is applied by row_log_table_apply(),
we must identify whether the records contain the extra header for
instantly added columns. For now, we will allocate an additional byte
for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table
has been subject to instant ADD COLUMN. The ROW_T_DELETE records are
fine, as they will be converted and will only contain 'core' columns
(PRIMARY KEY and some system columns) that are converted from dtuple_t.

rec_get_converted_size_temp(), rec_init_offsets_temp(),
rec_convert_dtuple_to_temp(): Add the parameter 'status'.

REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED:
An info_bits constant for distinguishing the 'default row' record.

rec_comp_status_t: An enum of the status bit values.

rec_leaf_format: An enum that replaces the bool parameter of
rec_init_offsets_comp_ordinary().
2017-10-06 09:50:10 +03:00

846 lines
31 KiB
C
Raw Blame History

/*****************************************************************************
Copyright (c) 1994, 2016, Oracle and/or its affiliates. All Rights Reserved.
Copyright (c) 2012, Facebook Inc.
Copyright (c) 2014, 2017, MariaDB Corporation.
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
*****************************************************************************/
/**************************************************//**
@file include/btr0btr.h
The B-tree
Created 6/2/1994 Heikki Tuuri
*******************************************************/
#ifndef btr0btr_h
#define btr0btr_h
#include "univ.i"
#include "dict0dict.h"
#include "data0data.h"
#include "page0cur.h"
#include "mtr0mtr.h"
#include "btr0types.h"
#include "gis0type.h"
/** Maximum record size which can be stored on a page, without using the
special big record storage structure */
#define BTR_PAGE_MAX_REC_SIZE (UNIV_PAGE_SIZE / 2 - 200)
/** @brief Maximum depth of a B-tree in InnoDB.
Note that this isn't a maximum as such; none of the tree operations
avoid producing trees bigger than this. It is instead a "max depth
that other code must work with", useful for e.g. fixed-size arrays
that must store some information about each level in a tree. In other
words: if a B-tree with bigger depth than this is encountered, it is
not acceptable for it to lead to mysterious memory corruption, but it
is acceptable for the program to die with a clear assert failure. */
#define BTR_MAX_LEVELS 100
/** Latching modes for btr_cur_search_to_nth_level(). */
enum btr_latch_mode {
/** Search a record on a leaf page and S-latch it. */
BTR_SEARCH_LEAF = RW_S_LATCH,
/** (Prepare to) modify a record on a leaf page and X-latch it. */
BTR_MODIFY_LEAF = RW_X_LATCH,
/** Obtain no latches. */
BTR_NO_LATCHES = RW_NO_LATCH,
/** Start modifying the entire B-tree. */
BTR_MODIFY_TREE = 33,
/** Continue modifying the entire B-tree. */
BTR_CONT_MODIFY_TREE = 34,
/** Search the previous record. */
BTR_SEARCH_PREV = 35,
/** Modify the previous record. */
BTR_MODIFY_PREV = 36,
/** Start searching the entire B-tree. */
BTR_SEARCH_TREE = 37,
/** Continue searching the entire B-tree. */
BTR_CONT_SEARCH_TREE = 38,
/* BTR_INSERT, BTR_DELETE and BTR_DELETE_MARK are mutually
exclusive. */
/** The search tuple will be inserted to the secondary index
at the searched position. When the leaf page is not in the
buffer pool, try to use the change buffer. */
BTR_INSERT = 512,
/** Try to delete mark a secondary index leaf page record at
the searched position using the change buffer when the page is
not in the buffer pool. */
BTR_DELETE_MARK = 4096,
/** Try to purge the record using the change buffer when the
secondary index leaf page is not in the buffer pool. */
BTR_DELETE = 8192,
/** The caller is already holding dict_index_t::lock S-latch. */
BTR_ALREADY_S_LATCHED = 16384,
/** Search and S-latch a leaf page, assuming that the
dict_index_t::lock S-latch is being held. */
BTR_SEARCH_LEAF_ALREADY_S_LATCHED = BTR_SEARCH_LEAF
| BTR_ALREADY_S_LATCHED,
/** Search the entire index tree, assuming that the
dict_index_t::lock S-latch is being held. */
BTR_SEARCH_TREE_ALREADY_S_LATCHED = BTR_SEARCH_TREE
| BTR_ALREADY_S_LATCHED,
/** Search and X-latch a leaf page, assuming that the
dict_index_t::lock S-latch is being held. */
BTR_MODIFY_LEAF_ALREADY_S_LATCHED = BTR_MODIFY_LEAF
| BTR_ALREADY_S_LATCHED,
/** Attempt to delete-mark a secondary index record. */
BTR_DELETE_MARK_LEAF = BTR_MODIFY_LEAF | BTR_DELETE_MARK,
/** Attempt to delete-mark a secondary index record
while holding the dict_index_t::lock S-latch. */
BTR_DELETE_MARK_LEAF_ALREADY_S_LATCHED = BTR_DELETE_MARK_LEAF
| BTR_ALREADY_S_LATCHED,
/** Attempt to purge a secondary index record. */
BTR_PURGE_LEAF = BTR_MODIFY_LEAF | BTR_DELETE,
/** Attempt to purge a secondary index record
while holding the dict_index_t::lock S-latch. */
BTR_PURGE_LEAF_ALREADY_S_LATCHED = BTR_PURGE_LEAF
| BTR_ALREADY_S_LATCHED
};
/** This flag ORed to btr_latch_mode says that we do the search in query
optimization */
#define BTR_ESTIMATE 1024U
/** This flag ORed to BTR_INSERT says that we can ignore possible
UNIQUE definition on secondary indexes when we decide if we can use
the insert buffer to speed up inserts */
#define BTR_IGNORE_SEC_UNIQUE 2048U
/** In the case of BTR_MODIFY_TREE, the caller specifies the intention
to insert record only. It is used to optimize block->lock range.*/
#define BTR_LATCH_FOR_INSERT 32768U
/** In the case of BTR_MODIFY_TREE, the caller specifies the intention
to delete record only. It is used to optimize block->lock range.*/
#define BTR_LATCH_FOR_DELETE 65536U
/** This flag is for undo insert of rtree. For rtree, we need this flag
to find proper rec to undo insert.*/
#define BTR_RTREE_UNDO_INS 131072U
/** In the case of BTR_MODIFY_LEAF, the caller intends to allocate or
free the pages of externally stored fields. */
#define BTR_MODIFY_EXTERNAL 262144U
/** Try to delete mark the record at the searched position when the
record is in spatial index */
#define BTR_RTREE_DELETE_MARK 524288U
#define BTR_LATCH_MODE_WITHOUT_FLAGS(latch_mode) \
((latch_mode) & btr_latch_mode(~(BTR_INSERT \
| BTR_DELETE_MARK \
| BTR_RTREE_UNDO_INS \
| BTR_RTREE_DELETE_MARK \
| BTR_DELETE \
| BTR_ESTIMATE \
| BTR_IGNORE_SEC_UNIQUE \
| BTR_ALREADY_S_LATCHED \
| BTR_LATCH_FOR_INSERT \
| BTR_LATCH_FOR_DELETE \
| BTR_MODIFY_EXTERNAL)))
#define BTR_LATCH_MODE_WITHOUT_INTENTION(latch_mode) \
((latch_mode) & btr_latch_mode(~(BTR_LATCH_FOR_INSERT \
| BTR_LATCH_FOR_DELETE \
| BTR_MODIFY_EXTERNAL)))
/**************************************************************//**
Report that an index page is corrupted. */
void
btr_corruption_report(
/*==================*/
const buf_block_t* block, /*!< in: corrupted block */
const dict_index_t* index) /*!< in: index tree */
ATTRIBUTE_COLD __attribute__((nonnull));
/** Assert that a B-tree page is not corrupted.
@param block buffer block containing a B-tree page
@param index the B-tree index */
#define btr_assert_not_corrupted(block, index) \
if ((ibool) !!page_is_comp(buf_block_get_frame(block)) \
!= dict_table_is_comp((index)->table)) { \
btr_corruption_report(block, index); \
ut_error; \
}
/**************************************************************//**
Gets the root node of a tree and sx-latches it for segment access.
@return root page, sx-latched */
page_t*
btr_root_get(
/*=========*/
const dict_index_t* index, /*!< in: index tree */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull));
/**************************************************************//**
Checks and adjusts the root node of a tree during IMPORT TABLESPACE.
@return error code, or DB_SUCCESS */
dberr_t
btr_root_adjust_on_import(
/*======================*/
const dict_index_t* index) /*!< in: index tree */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Gets the height of the B-tree (the level of the root, when the leaf
level is assumed to be 0). The caller must hold an S or X latch on
the index.
@return tree height (level of the root) */
ulint
btr_height_get(
/*===========*/
dict_index_t* index, /*!< in: index tree */
mtr_t* mtr) /*!< in/out: mini-transaction */
MY_ATTRIBUTE((warn_unused_result));
/** Gets a buffer page and declares its latching order level.
@param[in] page_id page id
@param[in] mode latch mode
@param[in] file file name
@param[in] line line where called
@param[in] index index tree, may be NULL if it is not an insert buffer
tree
@param[in,out] mtr mini-transaction
@return block */
UNIV_INLINE
buf_block_t*
btr_block_get_func(
const page_id_t& page_id,
const page_size_t& page_size,
ulint mode,
const char* file,
unsigned line,
dict_index_t* index,
mtr_t* mtr);
# ifdef UNIV_DEBUG
/** Gets a buffer page and declares its latching order level.
@param page_id tablespace/page identifier
@param page_size page size
@param mode latch mode
@param index index tree, may be NULL if not the insert buffer tree
@param mtr mini-transaction handle
@return the block descriptor */
# define btr_block_get(page_id, page_size, mode, index, mtr) \
btr_block_get_func(page_id, page_size, mode, \
__FILE__, __LINE__, (dict_index_t*)index, mtr)
# else /* UNIV_DEBUG */
/** Gets a buffer page and declares its latching order level.
@param page_id tablespace/page identifier
@param page_size page size
@param mode latch mode
@param index index tree, may be NULL if not the insert buffer tree
@param mtr mini-transaction handle
@return the block descriptor */
# define btr_block_get(page_id, page_size, mode, index, mtr) \
btr_block_get_func(page_id, page_size, mode, __FILE__, __LINE__, (dict_index_t*)index, mtr)
# endif /* UNIV_DEBUG */
/** Gets a buffer page and declares its latching order level.
@param page_id tablespace/page identifier
@param page_size page size
@param mode latch mode
@param index index tree, may be NULL if not the insert buffer tree
@param mtr mini-transaction handle
@return the uncompressed page frame */
UNIV_INLINE
page_t*
btr_page_get(
/*=========*/
const page_id_t& page_id,
const page_size_t& page_size,
ulint mode,
dict_index_t* index,
mtr_t* mtr)
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Gets the index id field of a page.
@return index id */
UNIV_INLINE
index_id_t
btr_page_get_index_id(
/*==================*/
const page_t* page) /*!< in: index page */
MY_ATTRIBUTE((warn_unused_result));
/********************************************************//**
Gets the node level field in an index page.
@return level, leaf level == 0 */
UNIV_INLINE
ulint
btr_page_get_level_low(
/*===================*/
const page_t* page) /*!< in: index page */
MY_ATTRIBUTE((warn_unused_result));
#define btr_page_get_level(page, mtr) btr_page_get_level_low(page)
/********************************************************//**
Gets the next index page number.
@return next page number */
UNIV_INLINE
ulint
btr_page_get_next(
/*==============*/
const page_t* page, /*!< in: index page */
mtr_t* mtr) /*!< in: mini-transaction handle */
MY_ATTRIBUTE((warn_unused_result));
/********************************************************//**
Gets the previous index page number.
@return prev page number */
UNIV_INLINE
ulint
btr_page_get_prev(
/*==============*/
const page_t* page, /*!< in: index page */
mtr_t* mtr) /*!< in: mini-transaction handle */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Releases the latch on a leaf page and bufferunfixes it. */
UNIV_INLINE
void
btr_leaf_page_release(
/*==================*/
buf_block_t* block, /*!< in: buffer block */
ulint latch_mode, /*!< in: BTR_SEARCH_LEAF or
BTR_MODIFY_LEAF */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull));
/**************************************************************//**
Gets the child node file address in a node pointer.
NOTE: the offsets array must contain all offsets for the record since
we read the last field according to offsets and assume that it contains
the child page number. In other words offsets must have been retrieved
with rec_get_offsets(n_fields=ULINT_UNDEFINED).
@return child node address */
UNIV_INLINE
ulint
btr_node_ptr_get_child_page_no(
/*===========================*/
const rec_t* rec, /*!< in: node pointer record */
const ulint* offsets)/*!< in: array returned by rec_get_offsets() */
MY_ATTRIBUTE((warn_unused_result));
/** Create the root node for a new index tree.
@param[in] type type of the index
@param[in] space space where created
@param[in] page_size page size
@param[in] index_id index id
@param[in] index index, or NULL when applying TRUNCATE
log record during recovery
@param[in] btr_redo_create_info used for applying TRUNCATE log
@param[in] mtr mini-transaction handle
record during recovery
@return page number of the created root, FIL_NULL if did not succeed */
ulint
btr_create(
ulint type,
ulint space,
const page_size_t& page_size,
index_id_t index_id,
dict_index_t* index,
const btr_create_t* btr_redo_create_info,
mtr_t* mtr);
/** Free a persistent index tree if it exists.
@param[in] page_id root page id
@param[in] page_size page size
@param[in] index_id PAGE_INDEX_ID contents
@param[in,out] mtr mini-transaction */
void
btr_free_if_exists(
const page_id_t& page_id,
const page_size_t& page_size,
index_id_t index_id,
mtr_t* mtr);
/** Free an index tree in a temporary tablespace or during TRUNCATE TABLE.
@param[in] page_id root page id
@param[in] page_size page size */
void
btr_free(
const page_id_t& page_id,
const page_size_t& page_size);
/** Read the last used AUTO_INCREMENT value from PAGE_ROOT_AUTO_INC.
@param[in,out] index clustered index
@return the last used AUTO_INCREMENT value
@retval 0 on error or if no AUTO_INCREMENT value was used yet */
ib_uint64_t
btr_read_autoinc(dict_index_t* index)
MY_ATTRIBUTE((nonnull, warn_unused_result));
/** Read the last used AUTO_INCREMENT value from PAGE_ROOT_AUTO_INC,
or fall back to MAX(auto_increment_column).
@param[in] table table containing an AUTO_INCREMENT column
@param[in] col_no index of the AUTO_INCREMENT column
@return the AUTO_INCREMENT value
@retval 0 on error or if no AUTO_INCREMENT value was used yet */
ib_uint64_t
btr_read_autoinc_with_fallback(const dict_table_t* table, unsigned col_no)
MY_ATTRIBUTE((nonnull, warn_unused_result));
/** Write the next available AUTO_INCREMENT value to PAGE_ROOT_AUTO_INC.
@param[in,out] index clustered index
@param[in] autoinc the AUTO_INCREMENT value
@param[in] reset whether to reset the AUTO_INCREMENT
to a possibly smaller value than currently
exists in the page */
void
btr_write_autoinc(dict_index_t* index, ib_uint64_t autoinc, bool reset = false)
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Makes tree one level higher by splitting the root, and inserts
the tuple. It is assumed that mtr contains an x-latch on the tree.
NOTE that the operation of this function must always succeed,
we cannot reverse it: therefore enough free disk space must be
guaranteed to be available before this function is called.
@return inserted record */
rec_t*
btr_root_raise_and_insert(
/*======================*/
ulint flags, /*!< in: undo logging and locking flags */
btr_cur_t* cursor, /*!< in: cursor at which to insert: must be
on the root page; when the function returns,
the cursor is positioned on the predecessor
of the inserted record */
ulint** offsets,/*!< out: offsets on inserted record */
mem_heap_t** heap, /*!< in/out: pointer to memory heap
that can be emptied, or NULL */
const dtuple_t* tuple, /*!< in: tuple to insert */
ulint n_ext, /*!< in: number of externally stored columns */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Reorganizes an index page.
IMPORTANT: On success, the caller will have to update IBUF_BITMAP_FREE
if this is a compressed leaf page in a secondary index. This has to
be done either within the same mini-transaction, or by invoking
ibuf_reset_free_bits() before mtr_commit(). On uncompressed pages,
IBUF_BITMAP_FREE is unaffected by reorganization.
@retval true if the operation was successful
@retval false if it is a compressed page, and recompression failed */
bool
btr_page_reorganize_low(
/*====================*/
bool recovery,/*!< in: true if called in recovery:
locks should not be updated, i.e.,
there cannot exist locks on the
page, and a hash index should not be
dropped: it cannot exist */
ulint z_level,/*!< in: compression level to be used
if dealing with compressed page */
page_cur_t* cursor, /*!< in/out: page cursor */
dict_index_t* index, /*!< in: the index tree of the page */
mtr_t* mtr) /*!< in/out: mini-transaction */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Reorganizes an index page.
IMPORTANT: On success, the caller will have to update IBUF_BITMAP_FREE
if this is a compressed leaf page in a secondary index. This has to
be done either within the same mini-transaction, or by invoking
ibuf_reset_free_bits() before mtr_commit(). On uncompressed pages,
IBUF_BITMAP_FREE is unaffected by reorganization.
@retval true if the operation was successful
@retval false if it is a compressed page, and recompression failed */
bool
btr_page_reorganize(
/*================*/
page_cur_t* cursor, /*!< in/out: page cursor */
dict_index_t* index, /*!< in: the index tree of the page */
mtr_t* mtr) /*!< in/out: mini-transaction */
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Decides if the page should be split at the convergence point of
inserts converging to left.
@return TRUE if split recommended */
ibool
btr_page_get_split_rec_to_left(
/*===========================*/
btr_cur_t* cursor, /*!< in: cursor at which to insert */
rec_t** split_rec)/*!< out: if split recommended,
the first record on upper half page,
or NULL if tuple should be first */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Decides if the page should be split at the convergence point of
inserts converging to right.
@return TRUE if split recommended */
ibool
btr_page_get_split_rec_to_right(
/*============================*/
btr_cur_t* cursor, /*!< in: cursor at which to insert */
rec_t** split_rec)/*!< out: if split recommended,
the first record on upper half page,
or NULL if tuple should be first */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Splits an index page to halves and inserts the tuple. It is assumed
that mtr holds an x-latch to the index tree. NOTE: the tree x-latch is
released within this function! NOTE that the operation of this
function must always succeed, we cannot reverse it: therefore enough
free disk space (2 pages) must be guaranteed to be available before
this function is called.
@return inserted record */
rec_t*
btr_page_split_and_insert(
/*======================*/
ulint flags, /*!< in: undo logging and locking flags */
btr_cur_t* cursor, /*!< in: cursor at which to insert; when the
function returns, the cursor is positioned
on the predecessor of the inserted record */
ulint** offsets,/*!< out: offsets on inserted record */
mem_heap_t** heap, /*!< in/out: pointer to memory heap
that can be emptied, or NULL */
const dtuple_t* tuple, /*!< in: tuple to insert */
ulint n_ext, /*!< in: number of externally stored columns */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((warn_unused_result));
/*******************************************************//**
Inserts a data tuple to a tree on a non-leaf level. It is assumed
that mtr holds an x-latch on the tree. */
void
btr_insert_on_non_leaf_level_func(
/*==============================*/
ulint flags, /*!< in: undo logging and locking flags */
dict_index_t* index, /*!< in: index */
ulint level, /*!< in: level, must be > 0 */
dtuple_t* tuple, /*!< in: the record to be inserted */
const char* file, /*!< in: file name */
unsigned line, /*!< in: line where called */
mtr_t* mtr); /*!< in: mtr */
#define btr_insert_on_non_leaf_level(f,i,l,t,m) \
btr_insert_on_non_leaf_level_func(f,i,l,t,__FILE__,__LINE__,m)
/****************************************************************//**
Sets a record as the predefined minimum record. */
void
btr_set_min_rec_mark(
/*=================*/
rec_t* rec, /*!< in/out: record */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Deletes on the upper level the node pointer to a page. */
void
btr_node_ptr_delete(
/*================*/
dict_index_t* index, /*!< in: index tree */
buf_block_t* block, /*!< in: page whose node pointer is deleted */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull));
#ifdef UNIV_DEBUG
/************************************************************//**
Checks that the node pointer to a page is appropriate.
@return TRUE */
ibool
btr_check_node_ptr(
/*===============*/
dict_index_t* index, /*!< in: index tree */
buf_block_t* block, /*!< in: index page */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((warn_unused_result));
#endif /* UNIV_DEBUG */
/*************************************************************//**
Tries to merge the page first to the left immediate brother if such a
brother exists, and the node pointers to the current page and to the
brother reside on the same page. If the left brother does not satisfy these
conditions, looks at the right brother. If the page is the only one on that
level lifts the records of the page to the father page, thus reducing the
tree height. It is assumed that mtr holds an x-latch on the tree and on the
page. If cursor is on the leaf level, mtr must also hold x-latches to
the brothers, if they exist.
@return TRUE on success */
ibool
btr_compress(
/*=========*/
btr_cur_t* cursor, /*!< in/out: cursor on the page to merge
or lift; the page must not be empty:
when deleting records, use btr_discard_page()
if the page would become empty */
ibool adjust, /*!< in: TRUE if should adjust the
cursor position even if compression occurs */
mtr_t* mtr) /*!< in/out: mini-transaction */
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Discards a page from a B-tree. This is used to remove the last record from
a B-tree page: the whole page must be removed at the same time. This cannot
be used for the root page, which is allowed to be empty. */
void
btr_discard_page(
/*=============*/
btr_cur_t* cursor, /*!< in: cursor on the page to discard: not on
the root page */
mtr_t* mtr); /*!< in: mtr */
/****************************************************************//**
Parses the redo log record for setting an index record as the predefined
minimum record.
@return end of log record or NULL */
byte*
btr_parse_set_min_rec_mark(
/*=======================*/
byte* ptr, /*!< in: buffer */
byte* end_ptr,/*!< in: buffer end */
ulint comp, /*!< in: nonzero=compact page format */
page_t* page, /*!< in: page or NULL */
mtr_t* mtr) /*!< in: mtr or NULL */
MY_ATTRIBUTE((nonnull(1,2), warn_unused_result));
/***********************************************************//**
Parses a redo log record of reorganizing a page.
@return end of log record or NULL */
byte*
btr_parse_page_reorganize(
/*======================*/
byte* ptr, /*!< in: buffer */
byte* end_ptr,/*!< in: buffer end */
dict_index_t* index, /*!< in: record descriptor */
bool compressed,/*!< in: true if compressed page */
buf_block_t* block, /*!< in: page to be reorganized, or NULL */
mtr_t* mtr) /*!< in: mtr or NULL */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Gets the number of pages in a B-tree.
@return number of pages, or ULINT_UNDEFINED if the index is unavailable */
ulint
btr_get_size(
/*=========*/
dict_index_t* index, /*!< in: index */
ulint flag, /*!< in: BTR_N_LEAF_PAGES or BTR_TOTAL_SIZE */
mtr_t* mtr) /*!< in/out: mini-transaction where index
is s-latched */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Gets the number of reserved and used pages in a B-tree.
@return number of pages reserved, or ULINT_UNDEFINED if the index
is unavailable */
UNIV_INTERN
ulint
btr_get_size_and_reserved(
/*======================*/
dict_index_t* index, /*!< in: index */
ulint flag, /*!< in: BTR_N_LEAF_PAGES or BTR_TOTAL_SIZE */
ulint* used, /*!< out: number of pages used (<= reserved) */
mtr_t* mtr) /*!< in/out: mini-transaction where index
is s-latched */
__attribute__((nonnull));
/**************************************************************//**
Allocates a new file page to be used in an index tree. NOTE: we assume
that the caller has made the reservation for free extents!
@retval NULL if no page could be allocated
@retval block, rw_lock_x_lock_count(&block->lock) == 1 if allocation succeeded
(init_mtr == mtr, or the page was not previously freed in mtr)
@retval block (not allocated or initialized) otherwise */
buf_block_t*
btr_page_alloc(
/*===========*/
dict_index_t* index, /*!< in: index tree */
ulint hint_page_no, /*!< in: hint of a good page */
byte file_direction, /*!< in: direction where a possible
page split is made */
ulint level, /*!< in: level where the page is placed
in the tree */
mtr_t* mtr, /*!< in/out: mini-transaction
for the allocation */
mtr_t* init_mtr) /*!< in/out: mini-transaction
for x-latching and initializing
the page */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Frees a file page used in an index tree. NOTE: cannot free field external
storage pages because the page must contain info on its level. */
void
btr_page_free(
/*==========*/
dict_index_t* index, /*!< in: index tree */
buf_block_t* block, /*!< in: block to be freed, x-latched */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull));
/** Empty an index page (possibly the root page). @see btr_page_create().
@param[in,out] block page to be emptied
@param[in,out] page_zip compressed page frame, or NULL
@param[in] index index of the page
@param[in] level B-tree level of the page (0=leaf)
@param[in,out] mtr mini-transaction */
void
btr_page_empty(
buf_block_t* block,
page_zip_des_t* page_zip,
dict_index_t* index,
ulint level,
mtr_t* mtr)
MY_ATTRIBUTE((nonnull(1, 3, 5)));
/**************************************************************//**
Creates a new index page (not the root, and also not
used in page reorganization). @see btr_page_empty(). */
void
btr_page_create(
/*============*/
buf_block_t* block, /*!< in/out: page to be created */
page_zip_des_t* page_zip,/*!< in/out: compressed page, or NULL */
dict_index_t* index, /*!< in: index */
ulint level, /*!< in: the B-tree level of the page */
mtr_t* mtr); /*!< in: mtr */
/**************************************************************//**
Frees a file page used in an index tree. Can be used also to BLOB
external storage pages. */
void
btr_page_free_low(
/*==============*/
dict_index_t* index, /*!< in: index tree */
buf_block_t* block, /*!< in: block to be freed, x-latched */
ulint level, /*!< in: page level (ULINT_UNDEFINED=BLOB) */
bool blob, /*!< in: blob page */
mtr_t* mtr) /*!< in: mtr */
MY_ATTRIBUTE((nonnull(1,2)));
/**************************************************************//**
Gets the root node of a tree and x- or s-latches it.
@return root page, x- or s-latched */
buf_block_t*
btr_root_block_get(
/*===============*/
const dict_index_t* index, /*!< in: index tree */
ulint mode, /*!< in: either RW_S_LATCH
or RW_X_LATCH */
mtr_t* mtr); /*!< in: mtr */
/*************************************************************//**
Reorganizes an index page.
IMPORTANT: On success, the caller will have to update IBUF_BITMAP_FREE
if this is a compressed leaf page in a secondary index. This has to
be done either within the same mini-transaction, or by invoking
ibuf_reset_free_bits() before mtr_commit(). On uncompressed pages,
IBUF_BITMAP_FREE is unaffected by reorganization.
@retval true if the operation was successful
@retval false if it is a compressed page, and recompression failed */
UNIV_INTERN
bool
btr_page_reorganize_block(
/*======================*/
bool recovery,/*!< in: true if called in recovery:
locks should not be updated, i.e.,
there cannot exist locks on the
page, and a hash index should not be
dropped: it cannot exist */
ulint z_level,/*!< in: compression level to be used
if dealing with compressed page */
buf_block_t* block, /*!< in/out: B-tree page */
dict_index_t* index, /*!< in: the index tree of the page */
mtr_t* mtr) /*!< in/out: mini-transaction */
__attribute__((nonnull));
#ifdef UNIV_BTR_PRINT
/*************************************************************//**
Prints size info of a B-tree. */
void
btr_print_size(
/*===========*/
dict_index_t* index) /*!< in: index tree */
MY_ATTRIBUTE((nonnull));
/**************************************************************//**
Prints directories and other info of all nodes in the index. */
void
btr_print_index(
/*============*/
dict_index_t* index, /*!< in: index */
ulint width) /*!< in: print this many entries from start
and end */
MY_ATTRIBUTE((nonnull));
#endif /* UNIV_BTR_PRINT */
/************************************************************//**
Checks the size and number of fields in a record based on the definition of
the index.
@return TRUE if ok */
ibool
btr_index_rec_validate(
/*===================*/
const rec_t* rec, /*!< in: index record */
const dict_index_t* index, /*!< in: index */
ibool dump_on_error) /*!< in: TRUE if the function
should print hex dump of record
and page on error */
MY_ATTRIBUTE((warn_unused_result));
/**************************************************************//**
Checks the consistency of an index tree.
@return DB_SUCCESS if ok, error code if not */
dberr_t
btr_validate_index(
/*===============*/
dict_index_t* index, /*!< in: index */
const trx_t* trx, /*!< in: transaction or 0 */
bool lockout)/*!< in: true if X-latch index is intended */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Removes a page from the level list of pages. */
UNIV_INTERN
void
btr_level_list_remove_func(
/*=======================*/
ulint space, /*!< in: space where removed */
const page_size_t& page_size,/*!< in: page size */
page_t* page, /*!< in/out: page to remove */
dict_index_t* index, /*!< in: index tree */
mtr_t* mtr); /*!< in/out: mini-transaction */
/*************************************************************//**
Removes a page from the level list of pages.
@param space in: space where removed
@param zip_size in: compressed page size in bytes, or 0 for uncompressed
@param page in/out: page to remove
@param index in: index tree
@param mtr in/out: mini-transaction */
# define btr_level_list_remove(space,zip_size,page,index,mtr) \
btr_level_list_remove_func(space,zip_size,page,index,mtr)
/*************************************************************//**
If page is the only on its level, this function moves its records to the
father page, thus reducing the tree height.
@return father block */
UNIV_INTERN
buf_block_t*
btr_lift_page_up(
/*=============*/
dict_index_t* index, /*!< in: index tree */
buf_block_t* block, /*!< in: page which is the only on its level;
must not be empty: use
btr_discard_only_page_on_level if the last
record from the page should be removed */
mtr_t* mtr) /*!< in: mtr */
__attribute__((nonnull));
#define BTR_N_LEAF_PAGES 1
#define BTR_TOTAL_SIZE 2
#include "btr0btr.ic"
/****************************************************************
Global variable controlling if scrubbing should be performed */
extern my_bool srv_immediate_scrub_data_uncompressed;
#endif
Reference in a new issue View git blame Copy permalink