Remove the HA_NO_BLOBS restriction from the HEAP engine, allowing
the optimizer to keep temporary tables with BLOB/TEXT columns in
memory when they fit within max_heap_table_size / tmp_memory_table_size
limits. Additionally, advertise HA_CAN_GEOMETRY so explicit
CREATE TABLE ... ENGINE=MEMORY with GEOMETRY columns works.
Unlike other HEAP blob implementations (e.g. Percona), this patch
provides full HASH index support on blob columns, enabling efficient
lookups, GROUP BY, and DISTINCT operations directly in HEAP without
falling back to disk.
Architecture
------------
BLOB data is stored using continuation records -- additional fixed-size
records allocated from the same HP_BLOCK that holds regular rows. This
reuses existing allocation, free list, and size accounting with minimal
structural change, and avoids per-blob my_malloc() calls.
The existing single-byte visibility flag is extended into a flags byte
with bits for HP_ROW_HAS_CONT, HP_ROW_IS_CONT, HP_ROW_CONT_ZEROCOPY,
HP_ROW_SINGLE_REC, and HP_ROW_MULTIPLE_REC. Continuation records are
grouped into variable-length runs -- contiguous sequences within a leaf
block. Only the first record of each run carries a 10-byte header
(next_cont pointer + run_rec_count); inner records are pure payload.
Three storage formats, detected by flag bits via inline predicates:
Case A (HP_ROW_SINGLE_REC): single record, no header, data at
offset 0. Zero-copy read.
Case B (HP_ROW_CONT_ZEROCOPY): single run, multiple records.
Header in rec 0, data contiguous in rec 1..N-1. Zero-copy read
via chain + recbuffer.
Case C (HP_ROW_MULTIPLE_REC): one or more runs linked via
next_cont. Reassembly into blob_buff required.
Run allocation uses a two-phase strategy: (1) peek-then-unlink walk
of the free list detecting contiguous groups, (2) tail allocation
from HP_BLOCK for remaining data. A Step 3 scavenge fallback
walks the entire free list when tail allocation fails.
HP_SHARE::total_records tracks all physical records (primary +
continuation), while HP_SHARE::records remains the logical count
used by hash bucket mapping.
Reassembly buffer (HP_INFO::blob_buff) follows the same pattern as
InnoDB's blob_heap -- allocated once, grown via my_realloc, freed
on heap_reset()/close. Zero-copy cases (A/B) return pointers
directly into HP_BLOCK with no copy.
Full HASH index key handling for BLOB columns: hp_rec_hashnr(),
hp_rec_key_cmp(), hp_key_cmp(), hp_make_key(), hp_hashnr() are
extended for HA_BLOB_PART segments. Hash pre-check optimization
skips expensive blob materialization when hashes differ. PAD SPACE
collation semantics are preserved for blob key comparisons.
Field_blob_key (Monty) produces HEAP-native key format (4-byte length
+ 8-byte data pointer) directly, eliminating key buffer translation
between the SQL layer and HEAP engine.
SQL layer changes
-----------------
pick_engine() (new, extracted from choose_engine()): replaced the
blob_fields check with a reclength > HA_MAX_REC_LENGTH guard.
choose_engine() calls pick_engine() with the real reclength.
pick_engine() is also called early in finalize() with reclength=0
to predict whether the engine will be HEAP, enabling blob-aware
GROUP BY key setup that avoids unnecessary m_using_unique_constraint.
finalize(): HEAP+blob uses fixed-width rows; GROUP BY key setup sets
key_part_flag from field, uses item max_length for blob key sizing.
store_length initialized for all GROUP BY key parts. key_type uses
field->binary() to determine FIELDFLAG_BINARY vs text collation.
DISTINCT key setup skips null-bits helper for HEAP.
remove_duplicates(): blob check moved before HEAP check to fall
through to remove_dup_with_compare().
Aggregator_distinct::add(): overflow-to-disk conversion via
create_internal_tmp_table_from_heap() for non-dup write errors.
Expression cache disabled for HEAP+blob (key format incompatibility).
FULLTEXT early detection in mysql_derived_prepare(): forces disk
engine via TMP_TABLE_FORCE_MYISAM when outer query uses MATCH.
Deferred blob chain free (MDEV-39732): heap_delete() saves chain
pointers to pending_blob_chains, flushed on next mutation or
heap_reset()/close. Prevents dangling zero-copy pointers during
binlog_log_row().
REPLACE safety (MDEV-39825): HP_SHARE::write_can_replace flag
forces copy mode in hp_read_blobs(), preventing blob data corruption
from freed-then-reused continuation records during REPLACE.
Geometry GROUP_CONCAT fix (MDEV-39761): downgrade Field_geom to
Field_blob for GROUP_CONCAT temp tables in both expression creation
paths. Type_handler_geometry::type_handler_for_tmp_table() added.
Geometry GROUP BY key fix (MDEV-39871): detect when new_key_field()
produced non-blob Field_varstring for a blob column, replace with
Field_blob_key.
Performance
-----------
Non-blob tables: zero regression. Every blob-specific code path is
guarded by if (share->blob_count). No new allocations, no format
changes, no hash function changes for non-blob keys.
Blob tables: eliminates file creation/deletion overhead and page cache
management. For single-run blobs (common case), the read path is
entirely zero-copy.
Limitations
-----------
1. No BTREE indexes on blob columns (HASH only)
2. No partial-key prefix indexing for blobs
3. 2x memory for Case C reads only (A/B are 1x)
4. No blob compression
5. 65,535 records per run (uint16 cap, auto-split)
6. max_heap_table_size applies to continuation records
7. Expression cache disabled for HEAP+blob
8. FULLTEXT forces disk engine
Linked bugs fixed:
- MDEV-39703: mroonga fulltext test ordering
- MDEV-39723: ER_DUP_ENTRY on GROUP BY with blob column
- MDEV-39724: crash in hp_is_single_rec with GROUP BY
- MDEV-39732: slave crash in hp_free_run_chain on blob replication
- MDEV-39761: Field_geom::store() assertion in GROUP_CONCAT
- MDEV-39782: RBR ER_KEY_NOT_FOUND on HEAP blob UPDATE
- MDEV-39825: blob data corruption on REPLACE into HEAP table
- MDEV-39871: crash in my_hash_sort_bin on GROUP BY with geometry
Reviewed by: Michael Widenius <monty@mariadb.org>
Monty reviewed the entire patch. Areas where he suggested changes
or contributed code:
- Field_blob_key class (HEAP-native blob key format, 4-byte length +
data pointer)
- Duplicate key fix on HEAP-to-Aria conversion
- hp_blob_key_length() uint32 fix
- hp_rec_hashnr_stored removal
- type_handler_for_tmp_table() param cleanup
- Type_handler_geometry::type_handler_for_tmp_table() virtual
- blob pointer bzero()
- find_unique_row() double-materialization fix
- Tail reclaim review
- Batch tail allocation review
- hp_update.c cleanup
- Field_blob_compressed temp table fix
- row_pack_length() dedup
- pack_length_no_ptr() removal
- Race condition fix in HEAP
- MDEV-39703 mroonga test fix
- MDEV-39825 write_can_replace optimization
- is_text_key_segment removal (field->binary() simplification)
- Documentation (Docs/internal-temporary-tables.txt)
Contribution by: Alexander Barkov <bar@mariadb.com>
Type_handler::make_and_init_table_field_ex() -- refactored temp table
field creation from inline code in sql_select.cc into type handler
virtual methods (sql_type.cc, sql_type_geom.cc), enabling clean
per-type-handler field creation for HEAP blob promotion.
The server_audit_timestamp_format %f specifier always rendered zero
microseconds. The server downcast the precise time to seconds before
passing it to audit plugins, and the plugin then re-fetched the time
itself at write time.
Pass the server's high-resolution time through the audit API instead:
- extend mysql_event_general with general_time_microseconds (added in
MYSQL_AUDIT_INTERFACE_VERSION 0x0304), keeping general_time in
seconds for backward compatibility
- the server_audit plugin uses event->general_time_microseconds for
query log entries instead of re-fetching the time at write time
Connection and table events carry no timestamp in the audit API, so the
plugin keeps taking the time at event time for those entries.
SET sql_mode=ORACLE;
DELIMITER $$
CREATE OR REPLACE PACKAGE pkg AS
-- Declare a package public data type
TYPE varchar_array IS TABLE OF VARCHAR(2000) INDEX BY INTEGER;
END;
$$
DELIMITER ;
DELIMITER $$
CREATE OR REPLACE PROCEDURE p1 AS
v pkg.varchar_array; -- Use the package public data type
BEGIN
v(0):='test';
SELECT v(0);
END;
$$
DELIMITER ;
Note, the change is done only for sql_mode=ORACLE, because the TYPE
declaration is not available for the default mode.
Where package-wide types are available
--------------------------------------
- Variabe list type:
DECLARE var pkg1.type1;
- RETURN type for a package routine:
CREATE FUNCTION .. RETURN pkg1.type1 ...
- Parameter type for a package routine:
PROCEDURE p1(param1 pkg1.type1);
- Assoc array element type:
TYPE assoc1_t IS TABLE OF pkg1.type1 ...
- REF CURSOR RETURN type:
TYPE cur1_t IS REF CURSOR RETURN pkg1.type1;
Change details
--------------
- Adding a member Lex_length_and_dec_st::m_foreign_module_type
It's set to true when the data type was initialized from a TYPE
in foreign routine (e.g. in PACKAGE spec).
It's needed to prevent use of qualified identifiers in public contexts,
i.e. in schema public routine parameter types and schema publuc function
RETURN types.
Adding a helper method sp_head::check_applicability() which prevents
use of qualified types in public context.
- Adding a helper method sp_head::raise_unknown_data_type().
- Adding methods LEX::set_field_type_typedef_package_spec() for
2-step and 3-step qualified indentifiers.
It's used in field_type_all_with_typedefs which covers cases:
- Variabe list type : DECLARE var pkg1.type1;
- RETURN type : CREATE FUNCTION .. RETURN pkg1.type1 ...
- Parameter type : PROCEDURE p1(param1 pkg1.type1);
- Assoc array element type : TYPE assoc1_t IS TABLE OF pkg1.type1 ...
- Adding a method LEX::declare_type_ref_cursor_return_typedef().
It handles cases when a new TYPE REF CURSOR RETURN is declared,
for both for qualified RETURN types and non-qualified RETURN types:
- TYPE cur0_t IS REF CURSOR RETURN rec1_t;
- TYPE cur0_t IS REF CURSOR RETURN pkg1.rec1_t;
- TYPE cur0_t IS REF CURSOR RETURN db1.pkg1.rec1_t;
The code was moved from LEX::declare_type_ref_cursor() into
LEX::declare_type_ref_cursor_return_typedef() and extended
to cover qualified RETURN types.
- Adding a method Sql_path::find_package_spec_type().
It iterates through all schemas specified in @@path and searches
for the given type in the given package.
- Adding a helper method sp_pcontext::type_defs_add_ref_cursor()
to reuse the code.
- Adding a new method sp_package::get_typedef() to search
for TYPE definitions in PACKAGE specifications.
- Adding a new method sp_head::get_typedef_package_spec()
to search for TYPE definitions used by a PROCEDURE or FUNCTION.
- Adding a helper method
Sp_handler::sp_cache_routine_reentrant_suppress_errors
Adding a method Sp_handler::find_package_spec().
plugins are not yet supported in replay mode.
So, disabling tests type_test.type_test_int8, and type_test.type_test_double
to be run in replay-server mode
qc_info cannot parse query cache key, this cannot be done
without changing query cache key format.
let's limit the (possibly incorrect) db_length to make sure
it does not extend beyond the db name end, this way there will be no
OOB reads and no information leak.
* reduced PARSEC_ITERATIONS_MAX because PKCS5_PBKDF2_HMAC() takes
a signed int
* instead of checking for '0'..'L' range (which includes punctuation,
check for the valid result of base62_to_uchar()
* add a connection test with non-default parsec_iterations
* make parsec_iterations a normal command line option
* update C/C to allow larger number of iterations in the client parsec
This patch adds a global and session level plugin variable
parsec_iterations to define the number of iterations to be
used when generating the key corresponding to the password.
It has a default value, lower and upper bounds.
THDVAR(thd, loc_info) can return NULL if the plugin was previously
uninstalled. get_loc_info() and its callers dereferenced it.
Guard against a NULL loc_info in get_loc_info(), auditing() and
log_current_query(). Such a session is not audited.
The plugin used clock() only to timestamp cache entries and measure
elapsed time against millisecond timeouts. As shown in MDEV-12345,
clock() is prohibitively expensive, and we do not need per-thread or
per-process CPU time here, only time differences.
Replace clock()/clock_t with the monotonic std::chrono::steady_clock.
Timeouts are kept as std::chrono::milliseconds
This statement:
OPEN c FOR 'dynamic sql'
was only allowed for SYS_REFCURSOR.
Additionally allow it for REF CURSOR with no RETURN clause, e.g.:
TYPE cur0_t IS REF CURSOR; -- No RETURN clause - OK for OPEN FOR
Note, REF CURSORs with RETURN clause are still not allowed for dynamic OPEN,
as expected.
Charset specified for XMLTYPE constructor.
Error returned if XML is in UTF16/UTF32 as we can't parse it yet.
XML check added so only well-formed allowed.
Field_xmltype::store has to do the my_error when it gets invalid XML,
just the warning with the -1 return results in the DBUG_ASSERT.
Also the UpdateXML function result reverted back to the STRING as
it can legally produce invalid XML, so we can't have columns built
from the UpdateXML to be XMLTYPE.
