diff options
author | unknown <knielsen@knielsen-hq.org> | 2010-01-15 16:58:25 +0100 |
---|---|---|
committer | unknown <knielsen@knielsen-hq.org> | 2010-01-15 16:58:25 +0100 |
commit | 13029651b5535d637aa6f652f17c8c4417a803bc (patch) | |
tree | b65c4d4bf3227b2e2b86c13f4d098b57e9a60410 | |
parent | d121e6630519a66eec7c953ee0eae623f592ce87 (diff) | |
parent | 5b0ab60a4bfb4d5b603cd929b7ceb0d9fd1bac22 (diff) | |
download | mariadb-git-13029651b5535d637aa6f652f17c8c4417a803bc.tar.gz |
Merge XtraDB 9 into MariaDB.
155 files changed, 5364 insertions, 3086 deletions
diff --git a/mysql-test/r/information_schema.result b/mysql-test/r/information_schema.result index 65378b62521..d8d6f461e90 100644 --- a/mysql-test/r/information_schema.result +++ b/mysql-test/r/information_schema.result @@ -85,6 +85,7 @@ TABLE_PRIVILEGES TRIGGERS USER_PRIVILEGES VIEWS +XTRADB_ADMIN_COMMAND XTRADB_ENHANCEMENTS columns_priv db @@ -865,8 +866,8 @@ TABLE_CONSTRAINTS TABLE_NAME select TABLE_PRIVILEGES TABLE_NAME select VIEWS TABLE_NAME select INNODB_BUFFER_POOL_PAGES_INDEX table_name select -INNODB_INDEX_STATS table_name select INNODB_TABLE_STATS table_name select +INNODB_INDEX_STATS table_name select delete from mysql.user where user='mysqltest_4'; delete from mysql.db where user='mysqltest_4'; flush privileges; diff --git a/mysql-test/r/information_schema_all_engines.result b/mysql-test/r/information_schema_all_engines.result index d7f5c30e9a1..189efedec60 100644 --- a/mysql-test/r/information_schema_all_engines.result +++ b/mysql-test/r/information_schema_all_engines.result @@ -35,7 +35,7 @@ INNODB_CMP INNODB_RSEG XTRADB_ENHANCEMENTS INNODB_BUFFER_POOL_PAGES_INDEX -INNODB_INDEX_STATS +XTRADB_ADMIN_COMMAND INNODB_TRX INNODB_CMP_RESET INNODB_LOCK_WAITS @@ -44,6 +44,7 @@ INNODB_LOCKS INNODB_CMPMEM INNODB_TABLE_STATS INNODB_BUFFER_POOL_PAGES_BLOB +INNODB_INDEX_STATS SELECT t.table_name, c1.column_name FROM information_schema.tables t INNER JOIN @@ -93,7 +94,7 @@ INNODB_CMP page_size INNODB_RSEG rseg_id XTRADB_ENHANCEMENTS name INNODB_BUFFER_POOL_PAGES_INDEX schema_name -INNODB_INDEX_STATS table_name +XTRADB_ADMIN_COMMAND result_message INNODB_TRX trx_id INNODB_CMP_RESET page_size INNODB_LOCK_WAITS requesting_trx_id @@ -102,6 +103,7 @@ INNODB_LOCKS lock_id INNODB_CMPMEM page_size INNODB_TABLE_STATS table_name INNODB_BUFFER_POOL_PAGES_BLOB space_id +INNODB_INDEX_STATS table_name SELECT t.table_name, c1.column_name FROM information_schema.tables t INNER JOIN @@ -151,7 +153,7 @@ INNODB_CMP page_size INNODB_RSEG rseg_id XTRADB_ENHANCEMENTS name INNODB_BUFFER_POOL_PAGES_INDEX schema_name -INNODB_INDEX_STATS table_name +XTRADB_ADMIN_COMMAND result_message INNODB_TRX trx_id INNODB_CMP_RESET page_size INNODB_LOCK_WAITS requesting_trx_id @@ -160,6 +162,7 @@ INNODB_LOCKS lock_id INNODB_CMPMEM page_size INNODB_TABLE_STATS table_name INNODB_BUFFER_POOL_PAGES_BLOB space_id +INNODB_INDEX_STATS table_name select 1 as f1 from information_schema.tables where "CHARACTER_SETS"= (select cast(table_name as char) from information_schema.tables order by table_name limit 1) limit 1; @@ -262,7 +265,7 @@ Database: information_schema | INNODB_RSEG | | XTRADB_ENHANCEMENTS | | INNODB_BUFFER_POOL_PAGES_INDEX | -| INNODB_INDEX_STATS | +| XTRADB_ADMIN_COMMAND | | INNODB_TRX | | INNODB_CMP_RESET | | INNODB_LOCK_WAITS | @@ -271,6 +274,7 @@ Database: information_schema | INNODB_CMPMEM | | INNODB_TABLE_STATS | | INNODB_BUFFER_POOL_PAGES_BLOB | +| INNODB_INDEX_STATS | +---------------------------------------+ Database: INFORMATION_SCHEMA +---------------------------------------+ @@ -310,7 +314,7 @@ Database: INFORMATION_SCHEMA | INNODB_RSEG | | XTRADB_ENHANCEMENTS | | INNODB_BUFFER_POOL_PAGES_INDEX | -| INNODB_INDEX_STATS | +| XTRADB_ADMIN_COMMAND | | INNODB_TRX | | INNODB_CMP_RESET | | INNODB_LOCK_WAITS | @@ -319,6 +323,7 @@ Database: INFORMATION_SCHEMA | INNODB_CMPMEM | | INNODB_TABLE_STATS | | INNODB_BUFFER_POOL_PAGES_BLOB | +| INNODB_INDEX_STATS | +---------------------------------------+ Wildcard: inf_rmation_schema +--------------------+ @@ -328,5 +333,5 @@ Wildcard: inf_rmation_schema +--------------------+ SELECT table_schema, count(*) FROM information_schema.TABLES WHERE table_schema IN ('mysql', 'INFORMATION_SCHEMA', 'test', 'mysqltest') AND table_name<>'ndb_binlog_index' AND table_name<>'ndb_apply_status' GROUP BY TABLE_SCHEMA; table_schema count(*) -information_schema 43 +information_schema 44 mysql 22 diff --git a/mysql-test/r/innodb-autoinc.result b/mysql-test/r/innodb-autoinc.result index a40a13dbe9f..abb8f3da072 100644 --- a/mysql-test/r/innodb-autoinc.result +++ b/mysql-test/r/innodb-autoinc.result @@ -875,11 +875,11 @@ ALTER TABLE t1 CHANGE c1 d1 INT NOT NULL AUTO_INCREMENT; SELECT * FROM t1; d1 1 -3 +2 SELECT * FROM t1; d1 1 -3 +2 INSERT INTO t1 VALUES(null); Got one of the listed errors ALTER TABLE t1 AUTO_INCREMENT = 3; @@ -888,13 +888,13 @@ Table Create Table t1 CREATE TABLE `t1` ( `d1` int(11) NOT NULL AUTO_INCREMENT, PRIMARY KEY (`d1`) -) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=latin1 +) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1 INSERT INTO t1 VALUES(null); SELECT * FROM t1; d1 1 +2 3 -4 DROP TABLE t1; SET @@SESSION.AUTO_INCREMENT_INCREMENT=1, @@SESSION.AUTO_INCREMENT_OFFSET=1; SHOW VARIABLES LIKE "%auto_inc%"; diff --git a/mysql-test/r/innodb-consistent.result b/mysql-test/r/innodb-consistent.result new file mode 100644 index 00000000000..9115791b99c --- /dev/null +++ b/mysql-test/r/innodb-consistent.result @@ -0,0 +1,35 @@ +drop table if exists t1; +set session transaction isolation level read committed; +create table t1(a int not null) engine=innodb DEFAULT CHARSET=latin1; +create table t2 like t1; +insert into t2 values (1),(2),(3),(4),(5),(6),(7); +set autocommit=0; +begin; +replace into t1 select * from t2; +set session transaction isolation level read committed; +set autocommit=0; +delete from t2 where a=5; +commit; +delete from t2; +commit; +commit; +begin; +insert into t1 select * from t2; +set session transaction isolation level read committed; +set autocommit=0; +delete from t2 where a=5; +commit; +delete from t2; +commit; +commit; +select * from t1; +a +1 +2 +3 +4 +5 +6 +7 +drop table t1; +drop table t2; diff --git a/mysql-test/r/innodb-index.result b/mysql-test/r/innodb-index.result index 7c4627ae8b2..9d199bb7260 100644 --- a/mysql-test/r/innodb-index.result +++ b/mysql-test/r/innodb-index.result @@ -1,4 +1,3 @@ -SET @save_innodb_file_format_check=@@global.innodb_file_format_check; create table t1(a int not null, b int, c char(10) not null, d varchar(20)) engine = innodb; insert into t1 values (5,5,'oo','oo'),(4,4,'tr','tr'),(3,4,'ad','ad'),(2,3,'ak','ak'); commit; @@ -629,7 +628,7 @@ drop table t1; create table t1(a int not null, b int, c char(10), d varchar(20), primary key (a)) engine = innodb; insert into t1 values (1,1,'ab','ab'),(2,2,'ac','ac'),(3,3,'ac','ac'),(4,4,'afe','afe'),(5,4,'affe','affe'); alter table t1 add unique index (b), add unique index (c), add unique index (d); -ERROR 23000: Duplicate entry '4' for key 'b' +ERROR 23000: Duplicate entry 'ac' for key 'c' alter table t1 add unique index (c), add unique index (b), add index (d); ERROR 23000: Duplicate entry 'ac' for key 'c' show create table t1; @@ -970,6 +969,7 @@ create index t1u on t1 (u(1)); drop table t1; set global innodb_file_per_table=0; set global innodb_file_format=Antelope; +set global innodb_file_format_check=Antelope; SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0; SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0; CREATE TABLE t1( @@ -1171,4 +1171,3 @@ a b 3 a 3 b DROP TABLE t1; -SET GLOBAL innodb_file_format_check=@save_innodb_file_format_check; diff --git a/mysql-test/r/innodb-zip.result b/mysql-test/r/innodb-zip.result index b26c4112826..21396d81ba8 100644 --- a/mysql-test/r/innodb-zip.result +++ b/mysql-test/r/innodb-zip.result @@ -196,15 +196,15 @@ drop table t1; set innodb_strict_mode = on; create table t1 (id int primary key) engine = innodb key_block_size = 0; ERROR HY000: Can't create table 'test.t1' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: invalid KEY_BLOCK_SIZE = 0. Valid values are [1, 2, 4, 8, 16] +Warning 1478 InnoDB: invalid KEY_BLOCK_SIZE = 0. Valid values are [1, 2, 4, 8, 16] Error 1005 Can't create table 'test.t1' (errno: 1478) create table t2 (id int primary key) engine = innodb key_block_size = 9; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] +Warning 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] Error 1005 Can't create table 'test.t2' (errno: 1478) create table t3 (id int primary key) engine = innodb key_block_size = 1; create table t4 (id int primary key) engine = innodb key_block_size = 2; @@ -233,30 +233,30 @@ key_block_size = 8 row_format = compressed; create table t2 (id int primary key) engine = innodb key_block_size = 8 row_format = redundant; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: cannot specify ROW_FORMAT = REDUNDANT with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: cannot specify ROW_FORMAT = REDUNDANT with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t2' (errno: 1478) create table t3 (id int primary key) engine = innodb key_block_size = 8 row_format = compact; ERROR HY000: Can't create table 'test.t3' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t3' (errno: 1478) create table t4 (id int primary key) engine = innodb key_block_size = 8 row_format = dynamic; ERROR HY000: Can't create table 'test.t4' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: cannot specify ROW_FORMAT = DYNAMIC with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: cannot specify ROW_FORMAT = DYNAMIC with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t4' (errno: 1478) create table t5 (id int primary key) engine = innodb key_block_size = 8 row_format = default; ERROR HY000: Can't create table 'test.t5' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t5' (errno: 1478) SELECT table_schema, table_name, row_format FROM information_schema.tables WHERE engine='innodb'; @@ -266,26 +266,26 @@ drop table t1; create table t1 (id int primary key) engine = innodb key_block_size = 9 row_format = redundant; ERROR HY000: Can't create table 'test.t1' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] -Error 1478 InnoDB: cannot specify ROW_FORMAT = REDUNDANT with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] +Warning 1478 InnoDB: cannot specify ROW_FORMAT = REDUNDANT with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t1' (errno: 1478) create table t2 (id int primary key) engine = innodb key_block_size = 9 row_format = compact; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] -Error 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] +Warning 1478 InnoDB: cannot specify ROW_FORMAT = COMPACT with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t2' (errno: 1478) create table t2 (id int primary key) engine = innodb key_block_size = 9 row_format = dynamic; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] -Error 1478 InnoDB: cannot specify ROW_FORMAT = DYNAMIC with KEY_BLOCK_SIZE. +Warning 1478 InnoDB: invalid KEY_BLOCK_SIZE = 9. Valid values are [1, 2, 4, 8, 16] +Warning 1478 InnoDB: cannot specify ROW_FORMAT = DYNAMIC with KEY_BLOCK_SIZE. Error 1005 Can't create table 'test.t2' (errno: 1478) SELECT table_schema, table_name, row_format FROM information_schema.tables WHERE engine='innodb'; @@ -293,45 +293,45 @@ table_schema table_name row_format set global innodb_file_per_table = off; create table t1 (id int primary key) engine = innodb key_block_size = 1; ERROR HY000: Can't create table 'test.t1' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. Error 1005 Can't create table 'test.t1' (errno: 1478) create table t2 (id int primary key) engine = innodb key_block_size = 2; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. Error 1005 Can't create table 'test.t2' (errno: 1478) create table t3 (id int primary key) engine = innodb key_block_size = 4; ERROR HY000: Can't create table 'test.t3' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. Error 1005 Can't create table 'test.t3' (errno: 1478) create table t4 (id int primary key) engine = innodb key_block_size = 8; ERROR HY000: Can't create table 'test.t4' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. Error 1005 Can't create table 'test.t4' (errno: 1478) create table t5 (id int primary key) engine = innodb key_block_size = 16; ERROR HY000: Can't create table 'test.t5' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_per_table. Error 1005 Can't create table 'test.t5' (errno: 1478) create table t6 (id int primary key) engine = innodb row_format = compressed; ERROR HY000: Can't create table 'test.t6' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: ROW_FORMAT=COMPRESSED requires innodb_file_per_table. +Warning 1478 InnoDB: ROW_FORMAT=COMPRESSED requires innodb_file_per_table. Error 1005 Can't create table 'test.t6' (errno: 1478) create table t7 (id int primary key) engine = innodb row_format = dynamic; ERROR HY000: Can't create table 'test.t7' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: ROW_FORMAT=DYNAMIC requires innodb_file_per_table. +Warning 1478 InnoDB: ROW_FORMAT=DYNAMIC requires innodb_file_per_table. Error 1005 Can't create table 'test.t7' (errno: 1478) create table t8 (id int primary key) engine = innodb row_format = compact; create table t9 (id int primary key) engine = innodb row_format = redundant; @@ -345,45 +345,45 @@ set global innodb_file_per_table = on; set global innodb_file_format = `0`; create table t1 (id int primary key) engine = innodb key_block_size = 1; ERROR HY000: Can't create table 'test.t1' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t1' (errno: 1478) create table t2 (id int primary key) engine = innodb key_block_size = 2; ERROR HY000: Can't create table 'test.t2' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t2' (errno: 1478) create table t3 (id int primary key) engine = innodb key_block_size = 4; ERROR HY000: Can't create table 'test.t3' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t3' (errno: 1478) create table t4 (id int primary key) engine = innodb key_block_size = 8; ERROR HY000: Can't create table 'test.t4' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t4' (errno: 1478) create table t5 (id int primary key) engine = innodb key_block_size = 16; ERROR HY000: Can't create table 'test.t5' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. +Warning 1478 InnoDB: KEY_BLOCK_SIZE requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t5' (errno: 1478) create table t6 (id int primary key) engine = innodb row_format = compressed; ERROR HY000: Can't create table 'test.t6' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: ROW_FORMAT=COMPRESSED requires innodb_file_format > Antelope. +Warning 1478 InnoDB: ROW_FORMAT=COMPRESSED requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t6' (errno: 1478) create table t7 (id int primary key) engine = innodb row_format = dynamic; ERROR HY000: Can't create table 'test.t7' (errno: 1478) -show errors; +show warnings; Level Code Message -Error 1478 InnoDB: ROW_FORMAT=DYNAMIC requires innodb_file_format > Antelope. +Warning 1478 InnoDB: ROW_FORMAT=DYNAMIC requires innodb_file_format > Antelope. Error 1005 Can't create table 'test.t7' (errno: 1478) create table t8 (id int primary key) engine = innodb row_format = compact; create table t9 (id int primary key) engine = innodb row_format = redundant; diff --git a/mysql-test/r/innodb.result b/mysql-test/r/innodb.result index 021bf34a132..9cfa825447b 100644 --- a/mysql-test/r/innodb.result +++ b/mysql-test/r/innodb.result @@ -3090,7 +3090,7 @@ ERROR HY000: Lock wait timeout exceeded; try restarting transaction commit; drop table t1, t2, t3, t5, t6, t8, t9; CREATE TABLE t1 (DB_ROW_ID int) engine=innodb; -ERROR HY000: Can't create table 'test.t1' (errno: -1) +ERROR 42000: Incorrect column name 'DB_ROW_ID' CREATE TABLE t1 ( a BIGINT(20) NOT NULL, PRIMARY KEY (a) diff --git a/mysql-test/r/innodb_bug36169.result b/mysql-test/r/innodb_bug36169.result index 0f43ab23f7a..aa80e4d7aa4 100644 --- a/mysql-test/r/innodb_bug36169.result +++ b/mysql-test/r/innodb_bug36169.result @@ -1,4 +1,2 @@ -set @old_innodb_file_format=@@innodb_file_format; -set @old_innodb_file_per_table=@@innodb_file_per_table; SET GLOBAL innodb_file_format='Barracuda'; SET GLOBAL innodb_file_per_table=ON; diff --git a/mysql-test/r/innodb_bug44369.result b/mysql-test/r/innodb_bug44369.result index 9cf79aeffab..ff25c774aa2 100644 --- a/mysql-test/r/innodb_bug44369.result +++ b/mysql-test/r/innodb_bug44369.result @@ -1,14 +1,6 @@ create table bug44369 (DB_ROW_ID int) engine=innodb; -ERROR HY000: Can't create table 'test.bug44369' (errno: -1) +ERROR 42000: Incorrect column name 'DB_ROW_ID' create table bug44369 (db_row_id int) engine=innodb; -ERROR HY000: Can't create table 'test.bug44369' (errno: -1) -show warnings; -Level Code Message -Warning 1005 Error creating table 'test/bug44369' with column name 'db_row_id'. 'db_row_id' is a reserved name. Please try to re-create the table with a different column name. -Error 1005 Can't create table 'test.bug44369' (errno: -1) +ERROR 42000: Incorrect column name 'db_row_id' create table bug44369 (db_TRX_Id int) engine=innodb; -ERROR HY000: Can't create table 'test.bug44369' (errno: -1) -show warnings; -Level Code Message -Warning 1005 Error creating table 'test/bug44369' with column name 'db_TRX_Id'. 'db_TRX_Id' is a reserved name. Please try to re-create the table with a different column name. -Error 1005 Can't create table 'test.bug44369' (errno: -1) +ERROR 42000: Incorrect column name 'db_TRX_Id' diff --git a/mysql-test/r/innodb_bug44571.result b/mysql-test/r/innodb_bug44571.result new file mode 100644 index 00000000000..955b55ac673 --- /dev/null +++ b/mysql-test/r/innodb_bug44571.result @@ -0,0 +1,7 @@ +CREATE TABLE bug44571 (foo INT) ENGINE=InnoDB; +ALTER TABLE bug44571 CHANGE foo bar INT; +ALTER TABLE bug44571 ADD INDEX bug44571b (foo); +ERROR 42000: Key column 'foo' doesn't exist in table +ALTER TABLE bug44571 ADD INDEX bug44571b (bar); +CREATE INDEX bug44571c ON bug44571 (bar); +DROP TABLE bug44571; diff --git a/mysql-test/r/innodb_bug46676.result b/mysql-test/r/innodb_bug46676.result new file mode 100644 index 00000000000..996799ce931 --- /dev/null +++ b/mysql-test/r/innodb_bug46676.result @@ -0,0 +1,9 @@ +SET foreign_key_checks=0; +CREATE TABLE t1 (id int, foreign key (id) references t2(id)) ENGINE=INNODB; +CREATE TABLE t2 (id int, foreign key (id) references t1(id)) ENGINE=INNODB; +SET foreign_key_checks=1; +SELECT COUNT(*) FROM information_schema.key_column_usage WHERE REFERENCED_TABLE_NAME in ('t1', 't2'); +COUNT(*) +2 +SET foreign_key_checks=0; +DROP TABLE t1, t2; diff --git a/mysql-test/r/innodb_bug47167.result b/mysql-test/r/innodb_bug47167.result new file mode 100644 index 00000000000..cbec363d78f --- /dev/null +++ b/mysql-test/r/innodb_bug47167.result @@ -0,0 +1,24 @@ +set @old_innodb_file_format_check=@@innodb_file_format_check; +select @old_innodb_file_format_check; +@old_innodb_file_format_check +Antelope +set global innodb_file_format_check = Barracuda; +select @@innodb_file_format_check; +@@innodb_file_format_check +Barracuda +set global innodb_file_format_check = DEFAULT; +select @@innodb_file_format_check; +@@innodb_file_format_check +Barracuda +set global innodb_file_format_check = @old_innodb_file_format_check; +select @@innodb_file_format_check; +@@innodb_file_format_check +Antelope +set global innodb_file_format_check = cheetah; +ERROR HY000: Incorrect arguments to SET +set global innodb_file_format_check = Bear; +ERROR HY000: Incorrect arguments to SET +set global innodb_file_format_check = on; +ERROR HY000: Incorrect arguments to SET +set global innodb_file_format_check = off; +ERROR HY000: Incorrect arguments to SET diff --git a/mysql-test/r/innodb_file_format.result b/mysql-test/r/innodb_file_format.result index 7b13a310fe8..d58c4ce8b28 100644 --- a/mysql-test/r/innodb_file_format.result +++ b/mysql-test/r/innodb_file_format.result @@ -1,4 +1,3 @@ -set @old_innodb_file_format=@@innodb_file_format; call mtr.add_suppression("InnoDB: invalid innodb_file_format_check value"); select @@innodb_file_format; @@innodb_file_format @@ -32,8 +31,6 @@ select @@innodb_file_format_check; @@innodb_file_format_check Barracuda set global innodb_file_format_check=default; -Warnings: -Warning 1210 Ignoring SET innodb_file_format=on select @@innodb_file_format_check; @@innodb_file_format_check Barracuda @@ -44,5 +41,4 @@ ERROR HY000: Incorrect arguments to SET select @@innodb_file_format_check; @@innodb_file_format_check Barracuda -set global innodb_file_format=@old_innodb_file_format; -set global innodb_file_format_check=Antelope; +set global innodb_file_format_check=antelope; diff --git a/mysql-test/r/innodb_xtradb_bug317074.result b/mysql-test/r/innodb_xtradb_bug317074.result index 5789c07282e..fff87ced019 100644 --- a/mysql-test/r/innodb_xtradb_bug317074.result +++ b/mysql-test/r/innodb_xtradb_bug317074.result @@ -1,6 +1,5 @@ SET @old_innodb_file_format=@@innodb_file_format; SET @old_innodb_file_per_table=@@innodb_file_per_table; -SET @old_innodb_file_format_check=@@innodb_file_format_check; SET GLOBAL innodb_file_format='Barracuda'; SET GLOBAL innodb_file_per_table=ON; DROP TABLE IF EXISTS `test1`; @@ -29,4 +28,4 @@ ALTER TABLE test1 ENGINE=MyISAM; DROP TABLE test1; SET GLOBAL innodb_file_format=@old_innodb_file_format; SET GLOBAL innodb_file_per_table=@old_innodb_file_per_table; -SET GLOBAL innodb_file_format_check=@old_innodb_file_format_check; +set global innodb_file_format_check=Antelope; diff --git a/mysql-test/t/innodb-analyze.test b/mysql-test/t/innodb-analyze.test index de5a486a355..9bdb9db697c 100644 --- a/mysql-test/t/innodb-analyze.test +++ b/mysql-test/t/innodb-analyze.test @@ -11,7 +11,7 @@ -- disable_result_log -- enable_warnings -SET @old_innodb_stats_sample_pages=@@innodb_stats_sample_pages; +let $sample_pages=`select @@innodb_stats_sample_pages`; SET GLOBAL innodb_stats_sample_pages=0; # check that the value has been adjusted to 1 @@ -62,4 +62,4 @@ SET GLOBAL innodb_stats_sample_pages=16; ANALYZE TABLE innodb_analyze; DROP TABLE innodb_analyze; -SET GLOBAL innodb_stats_sample_pages=@old_innodb_stats_sample_pages; +EVAL SET GLOBAL innodb_stats_sample_pages=$sample_pages; diff --git a/mysql-test/t/innodb-consistent-master.opt b/mysql-test/t/innodb-consistent-master.opt new file mode 100644 index 00000000000..cb48f1aaf60 --- /dev/null +++ b/mysql-test/t/innodb-consistent-master.opt @@ -0,0 +1 @@ +--loose-innodb_lock_wait_timeout=2 diff --git a/mysql-test/t/innodb-consistent.test b/mysql-test/t/innodb-consistent.test new file mode 100644 index 00000000000..5a7f4dc392d --- /dev/null +++ b/mysql-test/t/innodb-consistent.test @@ -0,0 +1,58 @@ +-- source include/not_embedded.inc +-- source include/have_innodb.inc + +--disable_warnings +drop table if exists t1; +--enable_warnings + +# REPLACE INTO ... SELECT and INSERT INTO ... SELECT should do +# a consistent read of the source table. + +connect (a,localhost,root,,); +connect (b,localhost,root,,); +connection a; +set session transaction isolation level read committed; +create table t1(a int not null) engine=innodb DEFAULT CHARSET=latin1; +create table t2 like t1; +insert into t2 values (1),(2),(3),(4),(5),(6),(7); +set autocommit=0; + +# REPLACE INTO ... SELECT case +begin; +# this should not result in any locks on t2. +replace into t1 select * from t2; + +connection b; +set session transaction isolation level read committed; +set autocommit=0; +# should not cuase a lock wait. +delete from t2 where a=5; +commit; +delete from t2; +commit; +connection a; +commit; + +# INSERT INTO ... SELECT case +begin; +# this should not result in any locks on t2. +insert into t1 select * from t2; + +connection b; +set session transaction isolation level read committed; +set autocommit=0; +# should not cuase a lock wait. +delete from t2 where a=5; +commit; +delete from t2; +commit; +connection a; +commit; + +select * from t1; +drop table t1; +drop table t2; + +connection default; +disconnect a; +disconnect b; diff --git a/mysql-test/t/innodb-index.test b/mysql-test/t/innodb-index.test index b8fa2e93916..38e5036b8be 100644 --- a/mysql-test/t/innodb-index.test +++ b/mysql-test/t/innodb-index.test @@ -1,6 +1,6 @@ -- source include/have_innodb.inc -SET @save_innodb_file_format_check=@@global.innodb_file_format_check; +let $innodb_file_format_check_orig=`select @@innodb_file_format_check`; create table t1(a int not null, b int, c char(10) not null, d varchar(20)) engine = innodb; insert into t1 values (5,5,'oo','oo'),(4,4,'tr','tr'),(3,4,'ad','ad'),(2,3,'ak','ak'); @@ -404,6 +404,7 @@ create index t1u on t1 (u(1)); drop table t1; eval set global innodb_file_per_table=$per_table; eval set global innodb_file_format=$format; +eval set global innodb_file_format_check=$format; # # Test to check whether CREATE INDEX handles implicit foreign key @@ -541,4 +542,9 @@ disconnect b; DROP TABLE t1; -SET GLOBAL innodb_file_format_check=@save_innodb_file_format_check; +# +# restore environment to the state it was before this test execution +# + +-- disable_query_log +eval SET GLOBAL innodb_file_format_check=$innodb_file_format_check_orig; diff --git a/mysql-test/t/innodb-master.opt b/mysql-test/t/innodb-master.opt index 4901efb416c..72c88068345 100644 --- a/mysql-test/t/innodb-master.opt +++ b/mysql-test/t/innodb-master.opt @@ -1 +1 @@ ---binlog_cache_size=32768 --innodb_lock_wait_timeout=1 +--binlog_cache_size=32768 --loose_innodb_lock_wait_timeout=1 diff --git a/mysql-test/t/innodb-semi-consistent-master.opt b/mysql-test/t/innodb-semi-consistent-master.opt index e76299453d3..cb48f1aaf60 100644 --- a/mysql-test/t/innodb-semi-consistent-master.opt +++ b/mysql-test/t/innodb-semi-consistent-master.opt @@ -1 +1 @@ ---innodb_lock_wait_timeout=2 +--loose-innodb_lock_wait_timeout=2 diff --git a/mysql-test/t/innodb-use-sys-malloc-master.opt b/mysql-test/t/innodb-use-sys-malloc-master.opt index 889834add01..fc8582b5887 100644 --- a/mysql-test/t/innodb-use-sys-malloc-master.opt +++ b/mysql-test/t/innodb-use-sys-malloc-master.opt @@ -1,2 +1 @@ ---innodb-use-sys-malloc=true ---innodb-use-sys-malloc=true +--loose-innodb-use-sys-malloc=true diff --git a/mysql-test/t/innodb-zip.test b/mysql-test/t/innodb-zip.test index f3f00aea26d..7514da087eb 100644 --- a/mysql-test/t/innodb-zip.test +++ b/mysql-test/t/innodb-zip.test @@ -178,11 +178,11 @@ set innodb_strict_mode = on; --error ER_CANT_CREATE_TABLE create table t1 (id int primary key) engine = innodb key_block_size = 0; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 9; -show errors; +show warnings; create table t3 (id int primary key) engine = innodb key_block_size = 1; @@ -208,22 +208,22 @@ key_block_size = 8 row_format = compressed; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 8 row_format = redundant; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t3 (id int primary key) engine = innodb key_block_size = 8 row_format = compact; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t4 (id int primary key) engine = innodb key_block_size = 8 row_format = dynamic; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t5 (id int primary key) engine = innodb key_block_size = 8 row_format = default; -show errors; +show warnings; SELECT table_schema, table_name, row_format FROM information_schema.tables WHERE engine='innodb'; @@ -233,17 +233,17 @@ drop table t1; --error ER_CANT_CREATE_TABLE create table t1 (id int primary key) engine = innodb key_block_size = 9 row_format = redundant; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 9 row_format = compact; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 9 row_format = dynamic; -show errors; +show warnings; SELECT table_schema, table_name, row_format FROM information_schema.tables WHERE engine='innodb'; @@ -253,25 +253,25 @@ set global innodb_file_per_table = off; --error ER_CANT_CREATE_TABLE create table t1 (id int primary key) engine = innodb key_block_size = 1; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 2; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t3 (id int primary key) engine = innodb key_block_size = 4; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t4 (id int primary key) engine = innodb key_block_size = 8; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t5 (id int primary key) engine = innodb key_block_size = 16; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t6 (id int primary key) engine = innodb row_format = compressed; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t7 (id int primary key) engine = innodb row_format = dynamic; -show errors; +show warnings; create table t8 (id int primary key) engine = innodb row_format = compact; create table t9 (id int primary key) engine = innodb row_format = redundant; @@ -285,25 +285,25 @@ set global innodb_file_format = `0`; --error ER_CANT_CREATE_TABLE create table t1 (id int primary key) engine = innodb key_block_size = 1; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t2 (id int primary key) engine = innodb key_block_size = 2; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t3 (id int primary key) engine = innodb key_block_size = 4; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t4 (id int primary key) engine = innodb key_block_size = 8; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t5 (id int primary key) engine = innodb key_block_size = 16; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t6 (id int primary key) engine = innodb row_format = compressed; -show errors; +show warnings; --error ER_CANT_CREATE_TABLE create table t7 (id int primary key) engine = innodb row_format = dynamic; -show errors; +show warnings; create table t8 (id int primary key) engine = innodb row_format = compact; create table t9 (id int primary key) engine = innodb row_format = redundant; diff --git a/mysql-test/t/innodb.test b/mysql-test/t/innodb.test index 599c6f10d23..19e5180dc8b 100644 --- a/mysql-test/t/innodb.test +++ b/mysql-test/t/innodb.test @@ -2270,7 +2270,7 @@ disconnect j; drop table t1, t2, t3, t5, t6, t8, t9; # bug 18934, "InnoDB crashes when table uses column names like DB_ROW_ID" ---error 1005 +--error ER_WRONG_COLUMN_NAME CREATE TABLE t1 (DB_ROW_ID int) engine=innodb; # diff --git a/mysql-test/t/innodb_bug34300.test b/mysql-test/t/innodb_bug34300.test index 0068fcd7b60..5b1939e648e 100644 --- a/mysql-test/t/innodb_bug34300.test +++ b/mysql-test/t/innodb_bug34300.test @@ -9,7 +9,7 @@ -- disable_result_log # set packet size and reconnect -SET @save_max_allowed_packet=@@global.max_allowed_packet; +let $max_packet=`select @@global.max_allowed_packet`; SET @@global.max_allowed_packet=16777216; --connect (newconn, localhost, root,,) @@ -33,4 +33,4 @@ SELECT f4, f8 FROM bug34300; DROP TABLE bug34300; disconnect newconn; connection default; -SET @@global.max_allowed_packet=@save_max_allowed_packet; +EVAL SET @@global.max_allowed_packet=$max_packet; diff --git a/mysql-test/t/innodb_bug36169.test b/mysql-test/t/innodb_bug36169.test index 22a2f6ddada..5bf55193b5c 100644 --- a/mysql-test/t/innodb_bug36169.test +++ b/mysql-test/t/innodb_bug36169.test @@ -4,9 +4,9 @@ # -- source include/have_innodb.inc -set @old_innodb_file_format=@@innodb_file_format; -set @old_innodb_file_per_table=@@innodb_file_per_table; +let $file_format=`select @@innodb_file_format`; +let $file_per_table=`select @@innodb_file_per_table`; SET GLOBAL innodb_file_format='Barracuda'; SET GLOBAL innodb_file_per_table=ON; @@ -1155,5 +1155,5 @@ DROP TABLE IF EXISTS table4; DROP TABLE IF EXISTS table5; DROP TABLE IF EXISTS table6; -set global innodb_file_format=@old_innodb_file_format; -set global innodb_file_per_table=@old_innodb_file_per_table; +EVAL SET GLOBAL innodb_file_format=$file_format; +EVAL SET GLOBAL innodb_file_per_table=$file_per_table; diff --git a/mysql-test/t/innodb_bug36172.test b/mysql-test/t/innodb_bug36172.test index 6fc6906be2f..67af2b2f7c2 100644 --- a/mysql-test/t/innodb_bug36172.test +++ b/mysql-test/t/innodb_bug36172.test @@ -13,9 +13,10 @@ SET storage_engine=InnoDB; -- disable_query_log -- disable_result_log -set @old_innodb_file_per_table=@@innodb_file_per_table; -set @old_innodb_file_format=@@innodb_file_format; +let $file_format=`select @@innodb_file_format`; +let $file_format_check=`select @@innodb_file_format_check`; +let $file_per_table=`select @@innodb_file_per_table`; SET GLOBAL innodb_file_format='Barracuda'; SET GLOBAL innodb_file_per_table=on; @@ -27,6 +28,6 @@ INSERT IGNORE INTO `table0` SET `col19` = '19940127002709', `col20` = 2383927.90 CHECK TABLE table0 EXTENDED; DROP TABLE table0; -set global innodb_file_per_table=@old_innodb_file_per_table; -set global innodb_file_format=@old_innodb_file_format; -set global innodb_file_format_check=Antelope; +EVAL SET GLOBAL innodb_file_format=$file_format; +EVAL SET GLOBAL innodb_file_format_check=$file_format_check; +EVAL SET GLOBAL innodb_file_per_table=$file_per_table; diff --git a/mysql-test/t/innodb_bug42101-nonzero-master.opt b/mysql-test/t/innodb_bug42101-nonzero-master.opt index d71dbe17d5b..455d66a06b8 100644 --- a/mysql-test/t/innodb_bug42101-nonzero-master.opt +++ b/mysql-test/t/innodb_bug42101-nonzero-master.opt @@ -1 +1 @@ ---innodb_commit_concurrency=1 +--loose_innodb_commit_concurrency=1 diff --git a/mysql-test/t/innodb_bug44369.test b/mysql-test/t/innodb_bug44369.test index 238dc3d8fb1..f5d85cd5815 100644 --- a/mysql-test/t/innodb_bug44369.test +++ b/mysql-test/t/innodb_bug44369.test @@ -6,16 +6,12 @@ --source include/have_innodb.inc # This create table operation should fail. ---error ER_CANT_CREATE_TABLE +--error ER_WRONG_COLUMN_NAME create table bug44369 (DB_ROW_ID int) engine=innodb; # This create should fail as well ---error ER_CANT_CREATE_TABLE +--error ER_WRONG_COLUMN_NAME create table bug44369 (db_row_id int) engine=innodb; -show warnings; - ---error ER_CANT_CREATE_TABLE +--error ER_WRONG_COLUMN_NAME create table bug44369 (db_TRX_Id int) engine=innodb; - -show warnings; diff --git a/mysql-test/t/innodb_bug44571.test b/mysql-test/t/innodb_bug44571.test new file mode 100644 index 00000000000..f5d79e3070a --- /dev/null +++ b/mysql-test/t/innodb_bug44571.test @@ -0,0 +1,13 @@ +# +# Bug#44571 InnoDB Plugin crashes on ADD INDEX +# http://bugs.mysql.com/44571 +# +-- source include/have_innodb.inc + +CREATE TABLE bug44571 (foo INT) ENGINE=InnoDB; +ALTER TABLE bug44571 CHANGE foo bar INT; +-- error ER_KEY_COLUMN_DOES_NOT_EXITS +ALTER TABLE bug44571 ADD INDEX bug44571b (foo); +ALTER TABLE bug44571 ADD INDEX bug44571b (bar); +CREATE INDEX bug44571c ON bug44571 (bar); +DROP TABLE bug44571; diff --git a/mysql-test/t/innodb_bug46676.test b/mysql-test/t/innodb_bug46676.test new file mode 100644 index 00000000000..440666c4226 --- /dev/null +++ b/mysql-test/t/innodb_bug46676.test @@ -0,0 +1,16 @@ +# This is the test for bug 46676: mysqld got exception 0xc0000005 +# It is reproducible with InnoDB plugin 1.0.4 + MySQL 5.1.37. +# But no longer reproducible after MySQL 5.1.38 (with plugin 1.0.5). + +--source include/have_innodb.inc + +SET foreign_key_checks=0; +CREATE TABLE t1 (id int, foreign key (id) references t2(id)) ENGINE=INNODB; +CREATE TABLE t2 (id int, foreign key (id) references t1(id)) ENGINE=INNODB; +SET foreign_key_checks=1; + +# Server crashes +SELECT COUNT(*) FROM information_schema.key_column_usage WHERE REFERENCED_TABLE_NAME in ('t1', 't2'); + +SET foreign_key_checks=0; +DROP TABLE t1, t2; diff --git a/mysql-test/t/innodb_bug47167.test b/mysql-test/t/innodb_bug47167.test new file mode 100644 index 00000000000..88e927c01bd --- /dev/null +++ b/mysql-test/t/innodb_bug47167.test @@ -0,0 +1,45 @@ +# This is the unit test for bug *47167. +# It tests setting the global variable +# "innodb_file_format_check" with a +# user-Defined Variable. + +--source include/have_innodb.inc + +# Save the value (Antelope) in 'innodb_file_format_check' to +# 'old_innodb_file_format_check' +set @old_innodb_file_format_check=@@innodb_file_format_check; + +# @old_innodb_file_format_check shall have the value of 'Antelope' +select @old_innodb_file_format_check; + +# Reset the value in 'innodb_file_format_check' to 'Barracuda' +set global innodb_file_format_check = Barracuda; + +select @@innodb_file_format_check; + +# Set 'innodb_file_format_check' to its default value, which +# is the latest file format supported in the current release. +set global innodb_file_format_check = DEFAULT; + +select @@innodb_file_format_check; + +# Put the saved value back to 'innodb_file_format_check' +set global innodb_file_format_check = @old_innodb_file_format_check; + +# Check whether 'innodb_file_format_check' get its original value. +select @@innodb_file_format_check; + +# Following are negative tests, all should fail. +--disable_warnings +--error ER_WRONG_ARGUMENTS +set global innodb_file_format_check = cheetah; + +--error ER_WRONG_ARGUMENTS +set global innodb_file_format_check = Bear; + +--error ER_WRONG_ARGUMENTS +set global innodb_file_format_check = on; + +--error ER_WRONG_ARGUMENTS +set global innodb_file_format_check = off; +--enable_warnings diff --git a/mysql-test/t/innodb_file_format.test b/mysql-test/t/innodb_file_format.test index 0999517bdd8..6de026ad97f 100644 --- a/mysql-test/t/innodb_file_format.test +++ b/mysql-test/t/innodb_file_format.test @@ -1,5 +1,4 @@ -- source include/have_innodb.inc -set @old_innodb_file_format=@@innodb_file_format; call mtr.add_suppression("InnoDB: invalid innodb_file_format_check value"); @@ -29,6 +28,4 @@ set global innodb_file_format=on; --error ER_WRONG_ARGUMENTS set global innodb_file_format=off; select @@innodb_file_format_check; - -set global innodb_file_format=@old_innodb_file_format; -set global innodb_file_format_check=Antelope; +set global innodb_file_format_check=antelope; diff --git a/mysql-test/t/innodb_information_schema.test b/mysql-test/t/innodb_information_schema.test index 45fb158b0da..fc1d38d8d14 100644 --- a/mysql-test/t/innodb_information_schema.test +++ b/mysql-test/t/innodb_information_schema.test @@ -109,19 +109,18 @@ SELECT * FROM ```t'\"_str` WHERE c1 = '3' FOR UPDATE; -- send SELECT * FROM ```t'\"_str` WHERE c1 = '4' FOR UPDATE; +-- enable_result_log -- connection con_verify_innodb_locks - -# Loop, giving time for the above 2 queries to execute before continuing. -# Without this, it sometimes happens that the SELECT FROM innodb_locks +# Wait for the above queries to execute before continuing. +# Without this, it sometimes happens that the SELECT from innodb_locks # executes before some of them, resulting in less than expected number -# of rows being selected from innodb_locks. -SET @counter := 0; -while (`SELECT (@counter := @counter + 1) <= 50 AND COUNT(*) != 14 FROM INFORMATION_SCHEMA.INNODB_LOCKS`) -{ - sleep 0.1; -} - --- enable_result_log +# of rows being selected from innodb_locks. If there is a bug and there +# are no 14 rows in innodb_locks then this test will fail with timeout. +let $count = 14; +let $table = INFORMATION_SCHEMA.INNODB_LOCKS; +-- source include/wait_until_rows_count.inc +# the above enables the query log, re-disable it +-- disable_query_log SELECT lock_mode, lock_type, lock_table, lock_index, lock_rec, lock_data FROM INFORMATION_SCHEMA.INNODB_LOCKS ORDER BY lock_data; diff --git a/mysql-test/t/innodb_xtradb_bug317074.test b/mysql-test/t/innodb_xtradb_bug317074.test index d3f7210d330..3b58c34598a 100644 --- a/mysql-test/t/innodb_xtradb_bug317074.test +++ b/mysql-test/t/innodb_xtradb_bug317074.test @@ -2,7 +2,7 @@ SET @old_innodb_file_format=@@innodb_file_format; SET @old_innodb_file_per_table=@@innodb_file_per_table; -SET @old_innodb_file_format_check=@@innodb_file_format_check; +let $innodb_file_format_check_orig=`select @@innodb_file_format_check`; SET GLOBAL innodb_file_format='Barracuda'; SET GLOBAL innodb_file_per_table=ON; @@ -45,4 +45,4 @@ ALTER TABLE test1 ENGINE=MyISAM; DROP TABLE test1; SET GLOBAL innodb_file_format=@old_innodb_file_format; SET GLOBAL innodb_file_per_table=@old_innodb_file_per_table; -SET GLOBAL innodb_file_format_check=@old_innodb_file_format_check; +eval set global innodb_file_format_check=$innodb_file_format_check_orig; diff --git a/storage/xtradb/CMakeLists.txt b/storage/xtradb/CMakeLists.txt index ec1e5f7ed38..38a9700df20 100644 --- a/storage/xtradb/CMakeLists.txt +++ b/storage/xtradb/CMakeLists.txt @@ -13,36 +13,41 @@ # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA -IF (CMAKE_SIZEOF_VOID_P MATCHES 8) - SET(WIN64 TRUE) -ENDIF (CMAKE_SIZEOF_VOID_P MATCHES 8) +# This is the CMakeLists for InnoDB Plugin -# Check type sizes -include(CheckTypeSize) - -# Currently, the checked results are not used. -CHECK_TYPE_SIZE(int SIZEOF_INT) -CHECK_TYPE_SIZE(long SIZEOF_LONG) -CHECK_TYPE_SIZE(void* SIZEOF_VOID_P) SET(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DSAFEMALLOC -DSAFE_MUTEX") SET(CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} -DSAFEMALLOC -DSAFE_MUTEX") -INCLUDE("${PROJECT_SOURCE_DIR}/storage/mysql_storage_engine.cmake") + +# Starting at 5.1.38, MySQL CMake files are simplified. But the plugin +# CMakeLists.txt still needs to work with previous versions of MySQL. +IF (MYSQL_VERSION_ID GREATER "50137") + INCLUDE("${PROJECT_SOURCE_DIR}/storage/mysql_storage_engine.cmake") +ENDIF (MYSQL_VERSION_ID GREATER "50137") + +IF (CMAKE_SIZEOF_VOID_P MATCHES 8) + SET(WIN64 TRUE) +ENDIF (CMAKE_SIZEOF_VOID_P MATCHES 8) + ADD_DEFINITIONS(-D_WIN32 -D_LIB -DMYSQL_SERVER) -INCLUDE_DIRECTORIES(${CMAKE_SOURCE_DIR}/include ${CMAKE_SOURCE_DIR}/zlib - ${CMAKE_SOURCE_DIR}/storage/xtradb/include - ${CMAKE_SOURCE_DIR}/storage/xtradb/handler - ${CMAKE_SOURCE_DIR}/sql - ${CMAKE_SOURCE_DIR}/regex - ${CMAKE_SOURCE_DIR}/extra/yassl/include) +# Include directories under innobase +INCLUDE_DIRECTORIES(${CMAKE_SOURCE_DIR}/storage/xtradb/include + ${CMAKE_SOURCE_DIR}/storage/xtradb/handler) + +# Include directories under mysql +INCLUDE_DIRECTORIES(${CMAKE_SOURCE_DIR}/include + ${CMAKE_SOURCE_DIR}/sql + ${CMAKE_SOURCE_DIR}/regex + ${CMAKE_SOURCE_DIR}/zlib + ${CMAKE_SOURCE_DIR}/extra/yassl/include) # Removing compiler optimizations for innodb/mem/* files on 64-bit Windows # due to 64-bit compiler error, See MySQL Bug #19424, #36366, #34297 -IF(MSVC AND $(WIN64)) +IF (MSVC AND $(WIN64)) SET_SOURCE_FILES_PROPERTIES(mem/mem0mem.c mem/mem0pool.c PROPERTIES COMPILE_FLAGS -Od) -ENDIF(MSVC AND $(WIN64)) +ENDIF (MSVC AND $(WIN64)) SET(INNOBASE_SOURCES btr/btr0btr.c btr/btr0cur.c btr/btr0pcur.c btr/btr0sea.c buf/buf0buddy.c buf/buf0buf.c buf/buf0flu.c buf/buf0lru.c buf/buf0rea.c @@ -77,5 +82,20 @@ SET(INNOBASE_SOURCES btr/btr0btr.c btr/btr0cur.c btr/btr0pcur.c btr/btr0sea.c usr/usr0sess.c ut/ut0byte.c ut/ut0dbg.c ut/ut0mem.c ut/ut0rnd.c ut/ut0ut.c ut/ut0vec.c ut/ut0list.c ut/ut0wqueue.c) +ADD_DEFINITIONS(-DHAVE_WINDOWS_ATOMICS -DIB_HAVE_PAUSE_INSTRUCTION) -MYSQL_STORAGE_ENGINE(INNOBASE) +IF (MYSQL_VERSION_ID GREATER "50137") + MYSQL_STORAGE_ENGINE(INNOBASE) + # Use ha_innodb for plugin name, if plugin is built + GET_TARGET_PROPERTY(LIB_LOCATION ha_innobase LOCATION) + IF(LIB_LOCATION) + SET_TARGET_PROPERTIES(ha_innobase PROPERTIES OUTPUT_NAME ha_innodb) + ENDIF(LIB_LOCATION) +ELSE (MYSQL_VERSION_ID GREATER "50137") + IF (NOT SOURCE_SUBLIBS) + ADD_DEFINITIONS(-D_WIN32 -DMYSQL_SERVER) + ADD_LIBRARY(innobase STATIC ${INNOBASE_SOURCES}) + # Require mysqld_error.h, which is built as part of the GenError + ADD_DEPENDENCIES(innobase GenError) + ENDIF (NOT SOURCE_SUBLIBS) +ENDIF (MYSQL_VERSION_ID GREATER "50137") diff --git a/storage/xtradb/ChangeLog b/storage/xtradb/ChangeLog index 2b04c06f0e8..1a6e07fd147 100644 --- a/storage/xtradb/ChangeLog +++ b/storage/xtradb/ChangeLog @@ -1,3 +1,319 @@ +2009-11-20 The InnoDB Team + + * handler/ha_innodb.cc: + Add a workaround to prevent a crash due to Bug#45961 DDL on + partitioned innodb tables leaves data dictionary in an inconsistent + state + +2009-11-19 The InnoDB Team + + * btr/btr0btr.c: + Fix Bug#48469 when innodb tablespace is configured too small, crash + and corruption! + +2009-11-19 The InnoDB Team + + * data/data0type.c: + Fix Bug#48526 Data type for float and double is incorrectly reported + in InnoDB table monitor + +2009-11-19 The InnoDB Team + + * CMakeLists.txt: + Fix Bug#48317 cannot build innodb as static library + +2009-11-18 The InnoDB Team + + * handler/handler0alter.cc: + Fix Bug#48782 On lock wait timeout, CREATE INDEX (creating primary key) + attempts DROP TABLE + +2009-11-17 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb.result, + mysql-test/innodb.test, mysql-test/innodb_bug44369.result, + mysql-test/innodb_bug44369.test, mysql-test/patches/innodb-index.diff, + row/row0mysql.c: + Report duplicate table names to the client connection, not to the + error log. + +2009-11-12 The InnoDB Team + + * handler/ha_innodb.cc, include/db0err.h, row/row0merge.c, + row/row0mysql.c: + Allow CREATE INDEX to be interrupted. + Also, when CHECK TABLE is interrupted, report ER_QUERY_INTERRUPTED. + +2009-11-11 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb_bug47167.result, + mysql-test/innodb_bug47167.test, mysql-test/innodb_file_format.result: + Fix Bug#47167 "set global innodb_file_format_check" cannot set value + by User-Defined Variable + +2009-11-11 The InnoDB Team + + * include/os0file.h, os/os0file.c: + Fix Bug#3139 Mysql crashes: 'windows error 995' after several selects + on a large DB + +2009-11-04 The InnoDB Team + + * handler/ha_innodb.cc: + Fix Bug#32430 'show innodb status' causes errors + Invalid (old?) table or database name in logs + +2009-11-02 The InnoDB Team + + * btr/btr0sea.c, buf/buf0buf.c, dict/dict0dict.c, fil/fil0fil.c, + ibuf/ibuf0ibuf.c, include/btr0sea.h, include/dict0dict.h, + include/fil0fil.h, include/ibuf0ibuf.h, include/lock0lock.h, + include/log0log.h, include/log0recv.h, include/mem0mem.h, + include/mem0pool.h, include/os0file.h, include/pars0pars.h, + include/srv0srv.h, include/thr0loc.h, include/trx0i_s.h, + include/trx0purge.h, include/trx0rseg.h, include/trx0sys.h, + include/trx0undo.h, include/usr0sess.h, lock/lock0lock.c, + log/log0log.c, log/log0recv.c, mem/mem0dbg.c, mem/mem0pool.c, + os/os0file.c, os/os0sync.c, os/os0thread.c, pars/lexyy.c, + pars/pars0lex.l, que/que0que.c, srv/srv0srv.c, srv/srv0start.c, + sync/sync0arr.c, sync/sync0sync.c, thr/thr0loc.c, trx/trx0i_s.c, + trx/trx0purge.c, trx/trx0rseg.c, trx/trx0sys.c, trx/trx0undo.c, + usr/usr0sess.c, ut/ut0mem.c: + Fix Bug #45992 innodb memory not freed after shutdown + Fix Bug #46656 InnoDB plugin: memory leaks (Valgrind) + +2009-10-29 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb-autoinc.result, + mysql-test/innodb-autoinc.test: + Fix Bug#47125 auto_increment start value is ignored if an index is + created and engine=innodb + +2009-10-29 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb_bug47777.result, + mysql-test/innodb_bug47777.test: + Fix Bug#47777 innodb dies with spatial pk: Failing assertion: buf <= + original_buf + buf_len + +2009-10-29 The InnoDB Team + + * handler/ha_innodb.cc: + Fix Bug#38996 Race condition in ANALYZE TABLE + +2009-10-29 The InnoDB Team + + * handler/ha_innodb.cc: + Fix bug#42383: Can't create table 'test.bug39438' + +2009-10-29 The InnoDB Team + + * os/os0proc.c: + Fix Bug#48237 Error handling in os_mem_alloc_large appears to + be incorrect + +2009-10-29 The InnoDB Team + + * buf/buf0buf.c, buf/buf0lru.c, include/buf0buf.h, include/buf0buf.ic: + Fix corruption of the buf_pool->LRU_old list and improve debug + assertions. + +2009-10-28 The InnoDB Team + + * srv/srv0start.c: + Fix Bug#41490 After enlargement of InnoDB page size, the error message + become inaccurate + +2009-10-26 The InnoDB Team + + * row/row0ins.c: + When allocating a data tuple, zero out the system fields in order + to avoid Valgrind warnings about uninitialized fields in + dtuple_validate(). + +2009-10-22 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb-zip.result, + mysql-test/innodb-zip.test, mysql-test/innodb_bug44369.result, + mysql-test/innodb_bug44369.test: + Fix Bug#47233 Innodb calls push_warning(MYSQL_ERROR::WARN_LEVEL_ERROR) + +2009-10-19 The InnoDB Team + + * mysql-test/innodb_information_schema.test: + Fix Bug#47808 innodb_information_schema.test fails when run under + valgrind + +2009-10-15 The InnoDB Team + + * include/page0page.ic: + Fix Bug#47058 Failure to compile innodb_plugin on solaris 10u7 + spro + cc/CC 5.10 + +2009-10-13 The InnoDB Team + + * buf/buf0flu.c: + Call fsync() on datafiles after a batch of pages is written to disk + even when skip_innodb_doublewrite is set. + +2009-10-05 The InnoDB Team + + * buf/buf0buf.c: + Do not invalidate buffer pool while an LRU batch is active. Added code + to buf_pool_invalidate() to wait for the running batches to finish. + +2009-10-01 The InnoDB Team + + * handler/ha_innodb.cc: + Fix Bug#47763 typo in error message: Failed to open table %s after %lu + attemtps. + +2009-10-01 The InnoDB Team + + * fsp/fsp0fsp.c, row/row0merge.c: + Clean up after a crash during DROP INDEX. When InnoDB crashes + while dropping an index, ensure that the index will be completely + dropped during crash recovery. The MySQL .frm file may still + contain the dropped index, but there is little that we can do + about it. + +2009-09-28 The InnoDB Team + + * handler/ha_innodb.cc: + When a secondary index exists in the MySQL .frm file but not in + the InnoDB data dictionary, return an error instead of letting an + assertion fail in index_read. + +2009-09-28 The InnoDB Team + + * btr/btr0btr.c, buf/buf0buf.c, include/page0page.h, + include/page0zip.h, page/page0cur.c, page/page0page.c, + page/page0zip.c: + Do not write to PAGE_INDEX_ID when restoring an uncompressed page + after a compression failure. The field should only be written + when creating a B-tree page. This fix addresses a race condition + in a debug assertion. + +2009-09-28 The InnoDB Team + + * fil/fil0fil.c: + Try to prevent the reuse of tablespace identifiers after InnoDB + has crashed during table creation. Also, refuse to start if files + with duplicate tablespace identifiers are encountered. + +2009-09-25 The InnoDB Team + + * include/os0file.h, os/os0file.c: + Fix Bug#47055 unconditional exit(1) on ERROR_WORKING_SET_QUOTA + 1453 (0x5AD) for InnoDB backend + +2009-09-19 The InnoDB Team + + * handler/ha_innodb.cc, mysql-test/innodb-consistent-master.opt, + mysql-test/innodb-consistent.result, + mysql-test/innodb-consistent.test: + Fix Bug#37232 Innodb might get too many read locks for DML with + repeatable-read + +2009-09-19 The InnoDB Team + + * fsp/fsp0fsp.c: + Fix Bug#31183 Tablespace full problems not reported in error log, + error message unclear + +2009-09-17 The InnoDB Team + + * mysql-test/innodb-zip.result, mysql-test/innodb-zip.test: + Make the test pass with zlib 1.2.3.3. Apparently, the definition + of compressBound() has changed between zlib versions, and the + maximum record size of a table with 1K compressed page size has + been reduced by one byte. This is an arbitrary test. In practical + applications, for good write performance, the compressed page size + should be chosen to be bigger than the absolute minimum. + +2009-09-16 The InnoDB Team + + * handler/ha_innodb.cc: + Fix Bug#46256 drop table with unknown collation crashes innodb + +2009-09-16 The InnoDB Team + + * dict/dict0dict.c, handler/ha_innodb.cc, + mysql-test/innodb_bug44369.result, mysql-test/innodb_bug44369.test, + row/row0mysql.c: + Fix Bug#44369 InnoDB: Does not uniformly disallow disallowed column + names + +2009-09-16 The InnoDB Team + + * handler/ha_innodb.cc, include/db0err.h, + mysql-test/innodb_bug46000.result, mysql-test/innodb_bug46000.test: + Fix Bug#46000 using index called GEN_CLUST_INDEX crashes server + +2009-09-02 The InnoDB Team + + * include/lock0lock.h, include/row0mysql.h, lock/lock0lock.c, + row/row0mysql.c: + Fix a regression introduced by the fix for MySQL bug#26316. We check + whether a transaction holds any AUTOINC locks before we acquire + the kernel mutex and release those locks. + +2009-08-27 The InnoDB Team + + * dict/dict0dict.c, include/dict0dict.h, + mysql-test/innodb_bug44571.result, mysql-test/innodb_bug44571.test: + Fix Bug#44571 InnoDB Plugin crashes on ADD INDEX + +2009-08-27 The InnoDB Team + + * row/row0merge.c: + Fix a bug in the merge sort that can corrupt indexes in fast index + creation. Add some consistency checks. Check that the number of + records remains constant in every merge sort pass. + +2009-08-27 The InnoDB Team + + * buf/buf0buf.c, buf/buf0lru.c, buf/buf0rea.c, handler/ha_innodb.cc, + include/buf0buf.h, include/buf0buf.ic, include/buf0lru.h, + include/ut0ut.h, ut/ut0ut.c: + Make it possible to tune the buffer pool LRU eviction policy to be + more resistant against index scans. Introduce the settable global + variables innodb_old_blocks_pct and innodb_old_blocks_time for + controlling the buffer pool eviction policy. The parameter + innodb_old_blocks_pct (5..95) controls the desired amount of "old" + blocks in the LRU list. The default is 37, corresponding to the + old fixed ratio of 3/8. Each time a block is accessed, it will be + moved to the "new" blocks if its first access was at least + innodb_old_blocks_time milliseconds ago (default 0, meaning every + block). The idea is that in index scans, blocks will be accessed + a few times within innodb_old_blocks_time, and they will remain in + the "old" section of the LRU list. Thus, when innodb_old_blocks_time + is nonzero, blocks retrieved for one-time index scans will be more + likely candidates for eviction than blocks that are accessed in + random patterns. + +2009-08-26 The InnoDB Team + + * handler/ha_innodb.cc, os/os0file.c: + Fix Bug#42885 buf_read_ahead_random, buf_read_ahead_linear counters, + thread wakeups + +2009-08-20 The InnoDB Team + + * lock/lock0lock.c: + Fix Bug#46650 Innodb assertion autoinc_lock == lock in + lock_table_remove_low on INSERT SELECT + +2009-08-13 The InnoDB Team + + * handler/handler0alter.cc: + Fix Bug#46657 InnoDB plugin: invalid read in index_merge_innodb test + (Valgrind) + +2009-08-11 The InnoDB Team + + InnoDB Plugin 1.0.4 released + 2009-07-20 The InnoDB Team * buf/buf0rea.c, handler/ha_innodb.cc, include/srv0srv.h, diff --git a/storage/xtradb/Makefile.am b/storage/xtradb/Makefile.am index 3aaf7a58ef9..53413dfaeb8 100644 --- a/storage/xtradb/Makefile.am +++ b/storage/xtradb/Makefile.am @@ -22,7 +22,7 @@ MYSQLLIBdir= $(pkglibdir) pkgplugindir= $(pkglibdir)/plugin INCLUDES= -I$(top_srcdir)/include -I$(top_builddir)/include \ -I$(top_srcdir)/regex \ - -I$(top_srcdir)/storage/xtradb/include \ + -I$(srcdir)/include \ -I$(top_srcdir)/sql \ -I$(srcdir) @ZLIB_INCLUDES@ @@ -31,7 +31,6 @@ DEFS= @DEFS@ noinst_HEADERS= \ handler/ha_innodb.h \ - handler/handler0vars.h \ handler/i_s.h \ include/btr0btr.h \ include/btr0btr.ic \ diff --git a/storage/xtradb/btr/btr0btr.c b/storage/xtradb/btr/btr0btr.c index 6ba9b36207b..086b3a0a599 100644 --- a/storage/xtradb/btr/btr0btr.c +++ b/storage/xtradb/btr/btr0btr.c @@ -790,14 +790,21 @@ btr_create( } else { /* It is a non-ibuf tree: create a file segment for leaf pages */ - fseg_create(space, page_no, - PAGE_HEADER + PAGE_BTR_SEG_LEAF, mtr); + if (!fseg_create(space, page_no, + PAGE_HEADER + PAGE_BTR_SEG_LEAF, mtr)) { + /* Not enough space for new segment, free root + segment before return. */ + btr_free_root(space, zip_size, page_no, mtr); + + return(FIL_NULL); + } + /* The fseg create acquires a second latch on the page, therefore we must declare it: */ buf_block_dbg_add_level(block, SYNC_TREE_NODE_NEW); } - /* Create a new index page on the the allocated segment page */ + /* Create a new index page on the allocated segment page */ page_zip = buf_block_get_page_zip(block); if (UNIV_LIKELY_NULL(page_zip)) { @@ -1011,7 +1018,26 @@ btr_page_reorganize_low( (!page_zip_compress(page_zip, page, index, NULL))) { /* Restore the old page and exit. */ - buf_frame_copy(page, temp_page); + +#if defined UNIV_DEBUG || defined UNIV_ZIP_DEBUG + /* Check that the bytes that we skip are identical. */ + ut_a(!memcmp(page, temp_page, PAGE_HEADER)); + ut_a(!memcmp(PAGE_HEADER + PAGE_N_RECS + page, + PAGE_HEADER + PAGE_N_RECS + temp_page, + PAGE_DATA - (PAGE_HEADER + PAGE_N_RECS))); + ut_a(!memcmp(UNIV_PAGE_SIZE - FIL_PAGE_DATA_END + page, + UNIV_PAGE_SIZE - FIL_PAGE_DATA_END + temp_page, + FIL_PAGE_DATA_END)); +#endif /* UNIV_DEBUG || UNIV_ZIP_DEBUG */ + + memcpy(PAGE_HEADER + page, PAGE_HEADER + temp_page, + PAGE_N_RECS - PAGE_N_DIR_SLOTS); + memcpy(PAGE_DATA + page, PAGE_DATA + temp_page, + UNIV_PAGE_SIZE - PAGE_DATA - FIL_PAGE_DATA_END); + +#if defined UNIV_DEBUG || defined UNIV_ZIP_DEBUG + ut_a(!memcmp(page, temp_page, UNIV_PAGE_SIZE)); +#endif /* UNIV_DEBUG || UNIV_ZIP_DEBUG */ goto func_exit; } @@ -1902,7 +1928,7 @@ func_start: n_uniq, &heap); /* If the new record is less than the existing record - the the split in the middle will copy the existing + the split in the middle will copy the existing record to the new node. */ if (cmp_dtuple_rec(tuple, first_rec, offsets) < 0) { split_rec = page_get_middle_rec(page); diff --git a/storage/xtradb/btr/btr0sea.c b/storage/xtradb/btr/btr0sea.c index 14f4b94f51b..e59d51e056e 100644 --- a/storage/xtradb/btr/btr0sea.c +++ b/storage/xtradb/btr/btr0sea.c @@ -175,6 +175,21 @@ btr_search_sys_create( btr_search_sys->hash_index = ha_create(hash_size, 0, 0); } +/*****************************************************************//** +Frees the adaptive search system at a database shutdown. */ +UNIV_INTERN +void +btr_search_sys_free(void) +/*=====================*/ +{ + mem_free(btr_search_latch_temp); + btr_search_latch_temp = NULL; + mem_heap_free(btr_search_sys->hash_index->heap); + hash_table_free(btr_search_sys->hash_index); + mem_free(btr_search_sys); + btr_search_sys = NULL; +} + /********************************************************************//** Disable the adaptive hash search system and empty the index. */ UNIV_INTERN @@ -957,7 +972,7 @@ btr_search_guess_on_hash( /* Increment the page get statistics though we did not really fix the page: for user info only */ - buf_pool->n_page_gets++; + buf_pool->stat.n_page_gets++; return(TRUE); diff --git a/storage/xtradb/buf/buf0buddy.c b/storage/xtradb/buf/buf0buddy.c index 6ee7a71a2e5..586a8dee1a2 100644 --- a/storage/xtradb/buf/buf0buddy.c +++ b/storage/xtradb/buf/buf0buddy.c @@ -531,11 +531,10 @@ buf_buddy_relocate( UNIV_MEM_ASSERT_W(src, size); mutex = buf_page_get_mutex_enter(bpage); - ut_a(mutex); mutex_enter(&zip_free_mutex); - if (buf_page_can_relocate(bpage)) { + if (mutex && buf_page_can_relocate(bpage)) { /* Relocate the compressed page. */ ut_a(bpage->zip.data == src); memcpy(dst, src, size); @@ -563,7 +562,9 @@ success: rw_lock_x_unlock(&page_hash_latch); } - mutex_exit(mutex); + if (mutex) { + mutex_exit(mutex); + } } else if (i == buf_buddy_get_slot(sizeof(buf_page_t))) { /* This must be a buf_page_t object. */ UNIV_MEM_ASSERT_RW(src, size); diff --git a/storage/xtradb/buf/buf0buf.c b/storage/xtradb/buf/buf0buf.c index 8da0a87751d..617b0e01b09 100644 --- a/storage/xtradb/buf/buf0buf.c +++ b/storage/xtradb/buf/buf0buf.c @@ -51,6 +51,40 @@ Created 11/5/1995 Heikki Tuuri #include "dict0dict.h" #include "log0recv.h" #include "page0zip.h" +#include "trx0trx.h" + +/* prototypes for new functions added to ha_innodb.cc */ +trx_t* innobase_get_trx(); + +inline void _increment_page_get_statistics(buf_block_t* block, trx_t* trx) +{ + ulint block_hash; + ulint block_hash_byte; + byte block_hash_offset; + + ut_ad(block); + + if (!innobase_get_slow_log() || !trx || !trx->take_stats) + return; + + if (!trx->distinct_page_access_hash) { + trx->distinct_page_access_hash = mem_alloc(DPAH_SIZE); + memset(trx->distinct_page_access_hash, 0, DPAH_SIZE); + } + + block_hash = ut_hash_ulint((block->page.space << 20) + block->page.space + + block->page.offset, DPAH_SIZE << 3); + block_hash_byte = block_hash >> 3; + block_hash_offset = (byte) block_hash & 0x07; + if (block_hash_byte < 0 || block_hash_byte >= DPAH_SIZE) + fprintf(stderr, "!!! block_hash_byte = %lu block_hash_offset = %lu !!!\n", block_hash_byte, block_hash_offset); + if (block_hash_offset < 0 || block_hash_offset > 7) + fprintf(stderr, "!!! block_hash_byte = %lu block_hash_offset = %lu !!!\n", block_hash_byte, block_hash_offset); + if ((trx->distinct_page_access_hash[block_hash_byte] & ((byte) 0x01 << block_hash_offset)) == 0) + trx->distinct_page_access++; + trx->distinct_page_access_hash[block_hash_byte] |= (byte) 0x01 << block_hash_offset; + return; +} /* IMPLEMENTATION OF THE BUFFER POOL @@ -845,16 +879,35 @@ buf_chunk_not_freed( block = chunk->blocks; for (i = chunk->size; i--; block++) { - mutex_enter(&block->mutex); - - if (buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE - && !buf_flush_ready_for_replace(&block->page)) { + ibool ready; + switch (buf_block_get_state(block)) { + case BUF_BLOCK_ZIP_FREE: + case BUF_BLOCK_ZIP_PAGE: + case BUF_BLOCK_ZIP_DIRTY: + /* The uncompressed buffer pool should never + contain compressed block descriptors. */ + ut_error; + break; + case BUF_BLOCK_NOT_USED: + case BUF_BLOCK_READY_FOR_USE: + case BUF_BLOCK_MEMORY: + case BUF_BLOCK_REMOVE_HASH: + /* Skip blocks that are not being used for + file pages. */ + break; + case BUF_BLOCK_FILE_PAGE: + mutex_enter(&block->mutex); + ready = buf_flush_ready_for_replace(&block->page); mutex_exit(&block->mutex); - return(block); - } - mutex_exit(&block->mutex); + if (!ready) { + + return(block); + } + + break; + } } return(NULL); @@ -985,8 +1038,6 @@ buf_pool_init(void) buf_pool->no_flush[i] = os_event_create(NULL); } - buf_pool->ulint_clock = 1; - /* 3. Initialize LRU fields --------------------------- */ /* All fields are initialized by mem_zalloc(). */ @@ -1024,7 +1075,11 @@ buf_pool_free(void) os_mem_free_large(chunk->mem, chunk->mem_size); } - buf_pool->n_chunks = 0; + mem_free(buf_pool->chunks); + hash_table_free(buf_pool->page_hash); + hash_table_free(buf_pool->zip_hash); + mem_free(buf_pool); + buf_pool = NULL; } /********************************************************************//** @@ -1171,10 +1226,15 @@ buf_relocate( #ifdef UNIV_LRU_DEBUG /* buf_pool->LRU_old must be the first item in the LRU list whose "old" flag is set. */ + ut_a(buf_pool->LRU_old->old); ut_a(!UT_LIST_GET_PREV(LRU, buf_pool->LRU_old) || !UT_LIST_GET_PREV(LRU, buf_pool->LRU_old)->old); ut_a(!UT_LIST_GET_NEXT(LRU, buf_pool->LRU_old) || UT_LIST_GET_NEXT(LRU, buf_pool->LRU_old)->old); + } else { + /* Check that the "old" flag is consistent in + the block and its neighbours. */ + buf_page_set_old(dpage, buf_page_is_old(dpage)); #endif /* UNIV_LRU_DEBUG */ } @@ -1510,35 +1570,8 @@ buf_pool_resize(void) } /********************************************************************//** -Moves the block to the start of the LRU list if there is a danger -that the block would drift out of the buffer pool. */ -UNIV_INLINE -void -buf_block_make_young( -/*=================*/ - buf_page_t* bpage) /*!< in: block to make younger */ -{ - ut_ad(!buf_pool_mutex_own()); - - /* Note that we read freed_page_clock's without holding any mutex: - this is allowed since the result is used only in heuristics */ - - if (buf_page_peek_if_too_old(bpage)) { - - //buf_pool_mutex_enter(); - mutex_enter(&LRU_list_mutex); - /* There has been freeing activity in the LRU list: - best to move to the head of the LRU list */ - - buf_LRU_make_block_young(bpage); - //buf_pool_mutex_exit(); - mutex_exit(&LRU_list_mutex); - } -} - -/********************************************************************//** Moves a page to the start of the buffer pool LRU list. This high-level -function can be used to prevent an important page from from slipping out of +function can be used to prevent an important page from slipping out of the buffer pool. */ UNIV_INTERN void @@ -1558,6 +1591,42 @@ buf_page_make_young( } /********************************************************************//** +Sets the time of the first access of a page and moves a page to the +start of the buffer pool LRU list if it is too old. This high-level +function can be used to prevent an important page from slipping +out of the buffer pool. */ +static +void +buf_page_set_accessed_make_young( +/*=============================*/ + buf_page_t* bpage, /*!< in/out: buffer block of a + file page */ + unsigned access_time) /*!< in: bpage->access_time + read under mutex protection, + or 0 if unknown */ +{ + ut_ad(!buf_pool_mutex_own()); + ut_a(buf_page_in_file(bpage)); + + if (buf_page_peek_if_too_old(bpage)) { + //buf_pool_mutex_enter(); + mutex_enter(&LRU_list_mutex); + buf_LRU_make_block_young(bpage); + //buf_pool_mutex_exit(); + mutex_exit(&LRU_list_mutex); + } else if (!access_time) { + ulint time_ms = ut_time_ms(); + mutex_t* block_mutex = buf_page_get_mutex_enter(bpage); + //buf_pool_mutex_enter(); + if (block_mutex) { + buf_page_set_accessed(bpage, time_ms); + mutex_exit(block_mutex); + } + //buf_pool_mutex_exit(); + } +} + +/********************************************************************//** Resets the check_index_page_at_flush field of a page if found in the buffer pool. */ UNIV_INTERN @@ -1696,11 +1765,20 @@ buf_page_get_zip( buf_page_t* bpage; mutex_t* block_mutex; ibool must_read; + unsigned access_time; + trx_t* trx = NULL; + ulint sec; + ulint ms; + ib_uint64_t start_time; + ib_uint64_t finish_time; #ifndef UNIV_LOG_DEBUG ut_ad(!ibuf_inside()); #endif - buf_pool->n_page_gets++; + if (innobase_get_slow_log()) { + trx = innobase_get_trx(); + } + buf_pool->stat.n_page_gets++; for (;;) { //buf_pool_mutex_enter(); @@ -1716,7 +1794,7 @@ lookup: //buf_pool_mutex_exit(); rw_lock_s_unlock(&page_hash_latch); - buf_read_page(space, zip_size, offset); + buf_read_page(space, zip_size, offset, trx); #if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG ut_a(++buf_dbg_counter % 37 || buf_validate()); @@ -1770,14 +1848,13 @@ err_exit: got_block: must_read = buf_page_get_io_fix(bpage) == BUF_IO_READ; + access_time = buf_page_is_accessed(bpage); //buf_pool_mutex_exit(); - buf_page_set_accessed(bpage, TRUE); - mutex_exit(block_mutex); - buf_block_make_young(bpage); + buf_page_set_accessed_make_young(bpage, access_time); #ifdef UNIV_DEBUG_FILE_ACCESSES ut_a(!bpage->file_page_was_freed); @@ -1793,6 +1870,13 @@ got_block: /* Let us wait until the read operation completes */ + if (innobase_get_slow_log() && trx && trx->take_stats) + { + ut_usectime(&sec, &ms); + start_time = (ib_uint64_t)sec * 1000000 + ms; + } else { + start_time = 0; + } for (;;) { enum buf_io_fix io_fix; @@ -1807,6 +1891,12 @@ got_block: break; } } + if (innobase_get_slow_log() && trx && trx->take_stats && start_time) + { + ut_usectime(&sec, &ms); + finish_time = (ib_uint64_t)sec * 1000000 + ms; + trx->io_reads_wait_timer += (ulint)(finish_time - start_time); + } } #ifdef UNIV_IBUF_COUNT_DEBUG @@ -1870,7 +1960,7 @@ buf_zip_decompress( switch (fil_page_get_type(frame)) { case FIL_PAGE_INDEX: if (page_zip_decompress(&block->page.zip, - block->frame)) { + block->frame, TRUE)) { return(TRUE); } @@ -2058,10 +2148,15 @@ buf_page_get_gen( mtr_t* mtr) /*!< in: mini-transaction */ { buf_block_t* block; - ibool accessed; + unsigned access_time; ulint fix_type; ibool must_read; mutex_t* block_mutex; + trx_t* trx = NULL; + ulint sec; + ulint ms; + ib_uint64_t start_time; + ib_uint64_t finish_time; ut_ad(mtr); ut_ad((rw_latch == RW_S_LATCH) @@ -2075,14 +2170,16 @@ buf_page_get_gen( #ifndef UNIV_LOG_DEBUG ut_ad(!ibuf_inside() || ibuf_page(space, zip_size, offset, NULL)); #endif - buf_pool->n_page_gets++; + if (innobase_get_slow_log()) { + trx = innobase_get_trx(); + } + buf_pool->stat.n_page_gets++; loop: block = guess; //buf_pool_mutex_enter(); if (block) { block_mutex = buf_page_get_mutex_enter((buf_page_t*)block); - ut_a(block_mutex); /* If the guess is a compressed page descriptor that has been allocated by buf_buddy_alloc(), it may have @@ -2092,7 +2189,9 @@ loop: the guess may be pointing to a buffer pool chunk that has been released when resizing the buffer pool. */ - if (!buf_block_is_uncompressed(block) + if (!block_mutex) { + block = guess = NULL; + } else if (!buf_block_is_uncompressed(block) || offset != block->page.offset || space != block->page.space || buf_block_get_state(block) != BUF_BLOCK_FILE_PAGE) { @@ -2127,7 +2226,7 @@ loop2: return(NULL); } - buf_read_page(space, zip_size, offset); + buf_read_page(space, zip_size, offset, trx); #if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG ut_a(++buf_dbg_counter % 37 || buf_validate()); @@ -2351,17 +2450,17 @@ wait_until_unfixed: UNIV_MEM_ASSERT_RW(&block->page, sizeof block->page); buf_block_buf_fix_inc(block, file, line); - //buf_pool_mutex_exit(); - /* Check if this is the first access to the page */ + //mutex_exit(&block->mutex); - accessed = buf_page_is_accessed(&block->page); + /* Check if this is the first access to the page */ - buf_page_set_accessed(&block->page, TRUE); + access_time = buf_page_is_accessed(&block->page); + //buf_pool_mutex_exit(); mutex_exit(block_mutex); - buf_block_make_young(&block->page); + buf_page_set_accessed_make_young(&block->page, access_time); #ifdef UNIV_DEBUG_FILE_ACCESSES ut_a(!block->page.file_page_was_freed); @@ -2379,6 +2478,13 @@ wait_until_unfixed: /* Let us wait until the read operation completes */ + if (innobase_get_slow_log() && trx && trx->take_stats) + { + ut_usectime(&sec, &ms); + start_time = (ib_uint64_t)sec * 1000000 + ms; + } else { + start_time = 0; + } for (;;) { enum buf_io_fix io_fix; @@ -2393,6 +2499,12 @@ wait_until_unfixed: break; } } + if (innobase_get_slow_log() && trx && trx->take_stats && start_time) + { + ut_usectime(&sec, &ms); + finish_time = (ib_uint64_t)sec * 1000000 + ms; + trx->io_reads_wait_timer += (ulint)(finish_time - start_time); + } } fix_type = MTR_MEMO_BUF_FIX; @@ -2414,17 +2526,21 @@ wait_until_unfixed: mtr_memo_push(mtr, block, fix_type); - if (!accessed) { + if (!access_time) { /* In the case of a first access, try to apply linear read-ahead */ - buf_read_ahead_linear(space, zip_size, offset); + buf_read_ahead_linear(space, zip_size, offset, trx); } #ifdef UNIV_IBUF_COUNT_DEBUG ut_a(ibuf_count_get(buf_block_get_space(block), buf_block_get_page_no(block)) == 0); #endif + if (innobase_get_slow_log()) { + _increment_page_get_statistics(block, trx); + } + return(block); } @@ -2444,9 +2560,10 @@ buf_page_optimistic_get_func( ulint line, /*!< in: line where called */ mtr_t* mtr) /*!< in: mini-transaction */ { - ibool accessed; + unsigned access_time; ibool success; ulint fix_type; + trx_t* trx = NULL; ut_ad(mtr && block); ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH)); @@ -2461,14 +2578,16 @@ buf_page_optimistic_get_func( } buf_block_buf_fix_inc(block, file, line); - accessed = buf_page_is_accessed(&block->page); - buf_page_set_accessed(&block->page, TRUE); mutex_exit(&block->mutex); - buf_block_make_young(&block->page); + /* Check if this is the first access to the page. + We do a dirty read on purpose, to avoid mutex contention. + This field is only used for heuristic purposes; it does not + affect correctness. */ - /* Check if this is the first access to the page */ + access_time = buf_page_is_accessed(&block->page); + buf_page_set_accessed_make_young(&block->page, access_time); ut_ad(!ibuf_inside() || ibuf_page(buf_block_get_space(block), @@ -2520,21 +2639,28 @@ buf_page_optimistic_get_func( #ifdef UNIV_DEBUG_FILE_ACCESSES ut_a(block->page.file_page_was_freed == FALSE); #endif - if (UNIV_UNLIKELY(!accessed)) { + if (innobase_get_slow_log()) { + trx = innobase_get_trx(); + } + + if (UNIV_UNLIKELY(!access_time)) { /* In the case of a first access, try to apply linear read-ahead */ buf_read_ahead_linear(buf_block_get_space(block), buf_block_get_zip_size(block), - buf_block_get_page_no(block)); + buf_block_get_page_no(block), trx); } #ifdef UNIV_IBUF_COUNT_DEBUG ut_a(ibuf_count_get(buf_block_get_space(block), buf_block_get_page_no(block)) == 0); #endif - buf_pool->n_page_gets++; + buf_pool->stat.n_page_gets++; + if (innobase_get_slow_log()) { + _increment_page_get_statistics(block, trx); + } return(TRUE); } @@ -2556,6 +2682,7 @@ buf_page_get_known_nowait( { ibool success; ulint fix_type; + trx_t* trx = NULL; ut_ad(mtr); ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH)); @@ -2581,8 +2708,24 @@ buf_page_get_known_nowait( mutex_exit(&block->mutex); - if (mode == BUF_MAKE_YOUNG) { - buf_block_make_young(&block->page); + if (mode == BUF_MAKE_YOUNG && buf_page_peek_if_too_old(&block->page)) { + //buf_pool_mutex_enter(); + mutex_enter(&LRU_list_mutex); + buf_LRU_make_block_young(&block->page); + //buf_pool_mutex_exit(); + mutex_exit(&LRU_list_mutex); + } else if (!buf_page_is_accessed(&block->page)) { + /* Above, we do a dirty read on purpose, to avoid + mutex contention. The field buf_page_t::access_time + is only used for heuristic purposes. Writes to the + field must be protected by mutex, however. */ + ulint time_ms = ut_time_ms(); + + //buf_pool_mutex_enter(); + mutex_enter(&block->mutex); + buf_page_set_accessed(&block->page, time_ms); + //buf_pool_mutex_exit(); + mutex_exit(&block->mutex); } ut_ad(!ibuf_inside() || (mode == BUF_KEEP_OLD)); @@ -2621,7 +2764,12 @@ buf_page_get_known_nowait( || (ibuf_count_get(buf_block_get_space(block), buf_block_get_page_no(block)) == 0)); #endif - buf_pool->n_page_gets++; + buf_pool->stat.n_page_gets++; + + if (innobase_get_slow_log()) { + trx = innobase_get_trx(); + _increment_page_get_statistics(block, trx); + } return(TRUE); } @@ -2700,7 +2848,7 @@ buf_page_try_get_func( #endif /* UNIV_DEBUG_FILE_ACCESSES */ buf_block_dbg_add_level(block, SYNC_NO_ORDER_CHECK); - buf_pool->n_page_gets++; + buf_pool->stat.n_page_gets++; #ifdef UNIV_IBUF_COUNT_DEBUG ut_a(ibuf_count_get(buf_block_get_space(block), @@ -2719,10 +2867,10 @@ buf_page_init_low( buf_page_t* bpage) /*!< in: block to init */ { bpage->flush_type = BUF_FLUSH_LRU; - bpage->accessed = FALSE; bpage->io_fix = BUF_IO_NONE; bpage->buf_fix_count = 0; bpage->freed_page_clock = 0; + bpage->access_time = 0; bpage->newest_modification = 0; bpage->oldest_modification = 0; HASH_INVALIDATE(bpage, hash); @@ -3044,6 +3192,7 @@ buf_page_create( buf_frame_t* frame; buf_block_t* block; buf_block_t* free_block = NULL; + ulint time_ms = ut_time_ms(); ut_ad(mtr); ut_ad(space || !zip_size); @@ -3095,7 +3244,7 @@ buf_page_create( buf_LRU_add_block(&block->page, FALSE); buf_block_buf_fix_inc(block, __FILE__, __LINE__); - buf_pool->n_pages_created++; + buf_pool->stat.n_pages_created++; if (zip_size) { void* data; @@ -3132,13 +3281,13 @@ buf_page_create( rw_lock_x_unlock(&block->lock); } + buf_page_set_accessed(&block->page, time_ms); + //buf_pool_mutex_exit(); mutex_exit(&LRU_list_mutex); mtr_memo_push(mtr, block, MTR_MEMO_BUF_FIX); - buf_page_set_accessed(&block->page, TRUE); - mutex_exit(&block->mutex); /* Delete possible entries for the page from the insert buffer: @@ -3355,7 +3504,7 @@ corrupt: ut_ad(buf_pool->n_pend_reads > 0); buf_pool->n_pend_reads--; - buf_pool->n_pages_read++; + buf_pool->stat.n_pages_read++; if (uncompressed) { rw_lock_x_unlock_gen(&((buf_block_t*) bpage)->lock, @@ -3380,7 +3529,7 @@ corrupt: BUF_IO_WRITE); } - buf_pool->n_pages_written++; + buf_pool->stat.n_pages_written++; break; @@ -3411,7 +3560,32 @@ void buf_pool_invalidate(void) /*=====================*/ { - ibool freed; + ibool freed; + enum buf_flush i; + + buf_pool_mutex_enter(); + + for (i = BUF_FLUSH_LRU; i < BUF_FLUSH_N_TYPES; i++) { + + /* As this function is called during startup and + during redo application phase during recovery, InnoDB + is single threaded (apart from IO helper threads) at + this stage. No new write batch can be in intialization + stage at this point. */ + ut_ad(buf_pool->init_flush[i] == FALSE); + + /* However, it is possible that a write batch that has + been posted earlier is still not complete. For buffer + pool invalidation to proceed we must ensure there is NO + write activity happening. */ + if (buf_pool->n_flush[i] > 0) { + buf_pool_mutex_exit(); + buf_flush_wait_batch_end(i); + buf_pool_mutex_enter(); + } + } + + buf_pool_mutex_exit(); ut_ad(buf_all_freed()); @@ -3427,6 +3601,14 @@ buf_pool_invalidate(void) ut_ad(UT_LIST_GET_LEN(buf_pool->LRU) == 0); ut_ad(UT_LIST_GET_LEN(buf_pool->unzip_LRU) == 0); + buf_pool->freed_page_clock = 0; + buf_pool->LRU_old = NULL; + buf_pool->LRU_old_len = 0; + buf_pool->LRU_flush_ended = 0; + + memset(&buf_pool->stat, 0x00, sizeof(buf_pool->stat)); + buf_refresh_io_stats(); + //buf_pool_mutex_exit(); mutex_exit(&LRU_list_mutex); } @@ -3706,6 +3888,7 @@ buf_print(void) "n pending decompressions %lu\n" "n pending reads %lu\n" "n pending flush LRU %lu list %lu single page %lu\n" + "pages made young %lu, not young %lu\n" "pages read %lu, created %lu, written %lu\n", (ulong) size, (ulong) UT_LIST_GET_LEN(buf_pool->LRU), @@ -3716,8 +3899,11 @@ buf_print(void) (ulong) buf_pool->n_flush[BUF_FLUSH_LRU], (ulong) buf_pool->n_flush[BUF_FLUSH_LIST], (ulong) buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE], - (ulong) buf_pool->n_pages_read, buf_pool->n_pages_created, - (ulong) buf_pool->n_pages_written); + (ulong) buf_pool->stat.n_pages_made_young, + (ulong) buf_pool->stat.n_pages_not_made_young, + (ulong) buf_pool->stat.n_pages_read, + (ulong) buf_pool->stat.n_pages_created, + (ulong) buf_pool->stat.n_pages_written); /* Count the number of blocks belonging to each index in the buffer */ @@ -3927,10 +4113,9 @@ buf_print_io( { time_t current_time; double time_elapsed; - ulint size; + ulint n_gets_diff; ut_ad(buf_pool); - size = buf_pool->curr_size; //buf_pool_mutex_enter(); mutex_enter(&LRU_list_mutex); @@ -3943,13 +4128,15 @@ buf_print_io( "Buffer pool size, bytes %lu\n" "Free buffers %lu\n" "Database pages %lu\n" + "Old database pages %lu\n" "Modified db pages %lu\n" "Pending reads %lu\n" "Pending writes: LRU %lu, flush list %lu, single page %lu\n", - (ulong) size, - (ulong) size * UNIV_PAGE_SIZE, + (ulong) buf_pool->curr_size, + (ulong) buf_pool->curr_size * UNIV_PAGE_SIZE, (ulong) UT_LIST_GET_LEN(buf_pool->free), (ulong) UT_LIST_GET_LEN(buf_pool->LRU), + (ulong) buf_pool->LRU_old_len, (ulong) UT_LIST_GET_LEN(buf_pool->flush_list), (ulong) buf_pool->n_pend_reads, (ulong) buf_pool->n_flush[BUF_FLUSH_LRU] @@ -3961,37 +4148,66 @@ buf_print_io( current_time = time(NULL); time_elapsed = 0.001 + difftime(current_time, buf_pool->last_printout_time); - buf_pool->last_printout_time = current_time; fprintf(file, + "Pages made young %lu, not young %lu\n" + "%.2f youngs/s, %.2f non-youngs/s\n" "Pages read %lu, created %lu, written %lu\n" "%.2f reads/s, %.2f creates/s, %.2f writes/s\n", - (ulong) buf_pool->n_pages_read, - (ulong) buf_pool->n_pages_created, - (ulong) buf_pool->n_pages_written, - (buf_pool->n_pages_read - buf_pool->n_pages_read_old) + (ulong) buf_pool->stat.n_pages_made_young, + (ulong) buf_pool->stat.n_pages_not_made_young, + (buf_pool->stat.n_pages_made_young + - buf_pool->old_stat.n_pages_made_young) + / time_elapsed, + (buf_pool->stat.n_pages_not_made_young + - buf_pool->old_stat.n_pages_not_made_young) / time_elapsed, - (buf_pool->n_pages_created - buf_pool->n_pages_created_old) + (ulong) buf_pool->stat.n_pages_read, + (ulong) buf_pool->stat.n_pages_created, + (ulong) buf_pool->stat.n_pages_written, + (buf_pool->stat.n_pages_read + - buf_pool->old_stat.n_pages_read) / time_elapsed, - (buf_pool->n_pages_written - buf_pool->n_pages_written_old) + (buf_pool->stat.n_pages_created + - buf_pool->old_stat.n_pages_created) + / time_elapsed, + (buf_pool->stat.n_pages_written + - buf_pool->old_stat.n_pages_written) / time_elapsed); - if (buf_pool->n_page_gets > buf_pool->n_page_gets_old) { - fprintf(file, "Buffer pool hit rate %lu / 1000\n", + n_gets_diff = buf_pool->stat.n_page_gets - buf_pool->old_stat.n_page_gets; + + if (n_gets_diff) { + fprintf(file, + "Buffer pool hit rate %lu / 1000," + " young-making rate %lu / 1000 not %lu / 1000\n", (ulong) - (1000 - ((1000 * (buf_pool->n_pages_read - - buf_pool->n_pages_read_old)) - / (buf_pool->n_page_gets - - buf_pool->n_page_gets_old)))); + (1000 - ((1000 * (buf_pool->stat.n_pages_read + - buf_pool->old_stat.n_pages_read)) + / (buf_pool->stat.n_page_gets + - buf_pool->old_stat.n_page_gets))), + (ulong) + (1000 * (buf_pool->stat.n_pages_made_young + - buf_pool->old_stat.n_pages_made_young) + / n_gets_diff), + (ulong) + (1000 * (buf_pool->stat.n_pages_not_made_young + - buf_pool->old_stat.n_pages_not_made_young) + / n_gets_diff)); } else { fputs("No buffer pool page gets since the last printout\n", file); } - buf_pool->n_page_gets_old = buf_pool->n_page_gets; - buf_pool->n_pages_read_old = buf_pool->n_pages_read; - buf_pool->n_pages_created_old = buf_pool->n_pages_created; - buf_pool->n_pages_written_old = buf_pool->n_pages_written; + /* Statistics about read ahead algorithm */ + fprintf(file, "Pages read ahead %.2f/s," + " evicted without access %.2f/s\n", + (buf_pool->stat.n_ra_pages_read + - buf_pool->old_stat.n_ra_pages_read) + / time_elapsed, + (buf_pool->stat.n_ra_pages_evicted + - buf_pool->old_stat.n_ra_pages_evicted) + / time_elapsed); /* Print some values to help us with visualizing what is happening with LRU eviction. */ @@ -4003,6 +4219,7 @@ buf_print_io( buf_LRU_stat_sum.io, buf_LRU_stat_cur.io, buf_LRU_stat_sum.unzip, buf_LRU_stat_cur.unzip); + buf_refresh_io_stats(); //buf_pool_mutex_exit(); mutex_exit(&LRU_list_mutex); mutex_exit(&free_list_mutex); @@ -4018,10 +4235,7 @@ buf_refresh_io_stats(void) /*======================*/ { buf_pool->last_printout_time = time(NULL); - buf_pool->n_page_gets_old = buf_pool->n_page_gets; - buf_pool->n_pages_read_old = buf_pool->n_pages_read; - buf_pool->n_pages_created_old = buf_pool->n_pages_created; - buf_pool->n_pages_written_old = buf_pool->n_pages_written; + buf_pool->old_stat = buf_pool->stat; } /*********************************************************************//** diff --git a/storage/xtradb/buf/buf0flu.c b/storage/xtradb/buf/buf0flu.c index d465483691a..9fb4fe7d9ea 100644 --- a/storage/xtradb/buf/buf0flu.c +++ b/storage/xtradb/buf/buf0flu.c @@ -331,6 +331,28 @@ buf_flush_write_complete( } /********************************************************************//** +Flush a batch of writes to the datafiles that have already been +written by the OS. */ +static +void +buf_flush_sync_datafiles(void) +/*==========================*/ +{ + /* Wake possible simulated aio thread to actually post the + writes to the operating system */ + os_aio_simulated_wake_handler_threads(); + + /* Wait that all async writes to tablespaces have been posted to + the OS */ + os_aio_wait_until_no_pending_writes(); + + /* Now we flush the data to disk (for example, with fsync) */ + fil_flush_file_spaces(FIL_TABLESPACE); + + return; +} + +/********************************************************************//** Flushes possible buffered writes from the doublewrite memory buffer to disk, and also wakes up the aio thread if simulated aio is used. It is very important to call this function after a batch of writes has been posted, @@ -347,8 +369,8 @@ buf_flush_buffered_writes(void) ulint i; if (!srv_use_doublewrite_buf || trx_doublewrite == NULL) { - os_aio_simulated_wake_handler_threads(); - + /* Sync the writes to the disk. */ + buf_flush_sync_datafiles(); return; } @@ -556,22 +578,10 @@ flush: buf_LRU_stat_inc_io(); } - /* Wake possible simulated aio thread to actually post the - writes to the operating system */ - - os_aio_simulated_wake_handler_threads(); - - /* Wait that all async writes to tablespaces have been posted to - the OS */ - - os_aio_wait_until_no_pending_writes(); - - /* Now we flush the data to disk (for example, with fsync) */ - - fil_flush_file_spaces(FIL_TABLESPACE); + /* Sync the writes to the disk. */ + buf_flush_sync_datafiles(); /* We can now reuse the doublewrite memory buffer: */ - trx_doublewrite->first_free = 0; mutex_exit(&(trx_doublewrite->mutex)); @@ -994,9 +1004,7 @@ buf_flush_try_neighbors( || buf_page_is_old(bpage)) { mutex_t* block_mutex = buf_page_get_mutex_enter(bpage); - ut_a(block_mutex); - - if (buf_flush_ready_for_flush(bpage, flush_type) + if (block_mutex && buf_flush_ready_for_flush(bpage, flush_type) && (i == offset || !bpage->buf_fix_count)) { /* We only try to flush those neighbors != offset where the buf fix count is @@ -1012,7 +1020,7 @@ buf_flush_try_neighbors( //buf_pool_mutex_enter(); rw_lock_s_lock(&page_hash_latch); - } else { + } else if (block_mutex) { mutex_exit(block_mutex); } } @@ -1050,6 +1058,7 @@ buf_flush_batch( min_n), otherwise ignored */ { buf_page_t* bpage; + buf_page_t* prev_bpage = NULL; ulint page_count = 0; ulint old_page_count; ulint space; @@ -1103,6 +1112,9 @@ flush_next: mutex_enter(&flush_list_mutex); remaining = UT_LIST_GET_LEN(buf_pool->flush_list); bpage = UT_LIST_GET_LAST(buf_pool->flush_list); + if (bpage) { + prev_bpage = UT_LIST_GET_PREV(flush_list, bpage); + } mutex_exit(&flush_list_mutex); if (!bpage || bpage->oldest_modification >= lsn_limit) { @@ -1123,11 +1135,14 @@ flush_next: mutex_t*block_mutex = buf_page_get_mutex_enter(bpage); ibool ready; - ut_a(buf_page_in_file(bpage)); + //ut_a(buf_page_in_file(bpage)); - ut_a(block_mutex); - ready = buf_flush_ready_for_flush(bpage, flush_type); - mutex_exit(block_mutex); + if (block_mutex) { + ready = buf_flush_ready_for_flush(bpage, flush_type); + mutex_exit(block_mutex); + } else { + ready = FALSE; + } if (ready) { space = buf_page_get_space(bpage); @@ -1162,6 +1177,13 @@ flush_next: mutex_enter(&flush_list_mutex); bpage = UT_LIST_GET_PREV(flush_list, bpage); //ut_ad(!bpage || bpage->in_flush_list); /* optimistic */ + if (bpage != prev_bpage) { + /* the search may warp.. retrying */ + bpage = NULL; + } + if (bpage) { + prev_bpage = UT_LIST_GET_PREV(flush_list, bpage); + } mutex_exit(&flush_list_mutex); remaining--; } @@ -1271,13 +1293,13 @@ buf_flush_LRU_recommendation(void) } block_mutex = buf_page_get_mutex_enter(bpage); - ut_a(block_mutex); - - if (buf_flush_ready_for_replace(bpage)) { + if (block_mutex && buf_flush_ready_for_replace(bpage)) { n_replaceable++; } - mutex_exit(block_mutex); + if (block_mutex) { + mutex_exit(block_mutex); + } distance++; diff --git a/storage/xtradb/buf/buf0lru.c b/storage/xtradb/buf/buf0lru.c index 2270ea5dce2..58e2c23275b 100644 --- a/storage/xtradb/buf/buf0lru.c +++ b/storage/xtradb/buf/buf0lru.c @@ -49,18 +49,22 @@ Created 11/5/1995 Heikki Tuuri #include "log0recv.h" #include "srv0srv.h" -/** The number of blocks from the LRU_old pointer onward, including the block -pointed to, must be 3/8 of the whole LRU list length, except that the -tolerance defined below is allowed. Note that the tolerance must be small -enough such that for even the BUF_LRU_OLD_MIN_LEN long LRU list, the -LRU_old pointer is not allowed to point to either end of the LRU list. */ +/** The number of blocks from the LRU_old pointer onward, including +the block pointed to, must be buf_LRU_old_ratio/BUF_LRU_OLD_RATIO_DIV +of the whole LRU list length, except that the tolerance defined below +is allowed. Note that the tolerance must be small enough such that for +even the BUF_LRU_OLD_MIN_LEN long LRU list, the LRU_old pointer is not +allowed to point to either end of the LRU list. */ #define BUF_LRU_OLD_TOLERANCE 20 -/** The whole LRU list length is divided by this number to determine an -initial segment in buf_LRU_get_recent_limit */ - -#define BUF_LRU_INITIAL_RATIO 8 +/** The minimum amount of non-old blocks when the LRU_old list exists +(that is, when there are more than BUF_LRU_OLD_MIN_LEN blocks). +@see buf_LRU_old_adjust_len */ +#define BUF_LRU_NON_OLD_MIN_LEN 5 +#if BUF_LRU_NON_OLD_MIN_LEN >= BUF_LRU_OLD_MIN_LEN +# error "BUF_LRU_NON_OLD_MIN_LEN >= BUF_LRU_OLD_MIN_LEN" +#endif /** When dropping the search hash index entries before deleting an ibd file, we build a local array of pages belonging to that tablespace @@ -107,6 +111,15 @@ UNIV_INTERN buf_LRU_stat_t buf_LRU_stat_sum; /* @} */ +/** @name Heuristics for detecting index scan @{ */ +/** Reserve this much/BUF_LRU_OLD_RATIO_DIV of the buffer pool for +"old" blocks. Protected by buf_pool_mutex. */ +UNIV_INTERN uint buf_LRU_old_ratio; +/** Move blocks to "new" LRU list only if the first access was at +least this many milliseconds ago. Not protected by any mutex or latch. */ +UNIV_INTERN uint buf_LRU_old_threshold_ms; +/* @} */ + /******************************************************************//** Takes a block out of the LRU list and page hash table. If the block is compressed-only (BUF_BLOCK_ZIP_PAGE), @@ -255,9 +268,12 @@ scan_again: mutex_t* block_mutex = buf_page_get_mutex_enter(bpage); buf_page_t* prev_bpage; - ut_a(block_mutex); prev_bpage = UT_LIST_GET_PREV(LRU, bpage); + if (!block_mutex) { + goto next_page; + } + ut_a(buf_page_in_file(bpage)); if (buf_page_get_state(bpage) != BUF_BLOCK_FILE_PAGE @@ -360,9 +376,13 @@ scan_again: ut_a(buf_page_in_file(bpage)); - ut_a(block_mutex); prev_bpage = UT_LIST_GET_PREV(LRU, bpage); + if (!block_mutex) { + bpage = prev_bpage; + continue; + } + if (buf_page_get_space(bpage) == id) { if (bpage->buf_fix_count > 0 || buf_page_get_io_fix(bpage) != BUF_IO_NONE) { @@ -450,45 +470,6 @@ next_page: } } -/******************************************************************//** -Gets the minimum LRU_position field for the blocks in an initial segment -(determined by BUF_LRU_INITIAL_RATIO) of the LRU list. The limit is not -guaranteed to be precise, because the ulint_clock may wrap around. -@return the limit; zero if could not determine it */ -UNIV_INTERN -ulint -buf_LRU_get_recent_limit(void) -/*==========================*/ -{ - const buf_page_t* bpage; - ulint len; - ulint limit; - - //buf_pool_mutex_enter(); - mutex_enter(&LRU_list_mutex); - - len = UT_LIST_GET_LEN(buf_pool->LRU); - - if (len < BUF_LRU_OLD_MIN_LEN) { - /* The LRU list is too short to do read-ahead */ - - //buf_pool_mutex_exit(); - mutex_exit(&LRU_list_mutex); - - return(0); - } - - bpage = UT_LIST_GET_FIRST(buf_pool->LRU); - - limit = buf_page_get_LRU_position(bpage); - len /= BUF_LRU_INITIAL_RATIO; - - //buf_pool_mutex_exit(); - mutex_exit(&LRU_list_mutex); - - return(limit > len ? (limit - len) : 0); -} - /********************************************************************//** Insert a compressed block into buf_pool->zip_clean in the LRU order. */ UNIV_INTERN @@ -631,10 +612,13 @@ restart: bpage = UT_LIST_GET_PREV(LRU, bpage), distance--) { enum buf_lru_free_block_status freed; + unsigned accessed; mutex_t* block_mutex = buf_page_get_mutex_enter(bpage); - ut_a(block_mutex); + if (!block_mutex) { + goto restart; + } if (!bpage->in_LRU_list || !buf_page_in_file(bpage)) { @@ -645,11 +629,18 @@ restart: ut_ad(buf_page_in_file(bpage)); ut_ad(bpage->in_LRU_list); + accessed = buf_page_is_accessed(bpage); freed = buf_LRU_free_block(bpage, TRUE, NULL, have_LRU_mutex); mutex_exit(block_mutex); switch (freed) { case BUF_LRU_FREED: + /* Keep track of pages that are evicted without + ever being accessed. This gives us a measure of + the effectiveness of readahead */ + if (!accessed) { + ++buf_pool->stat.n_ra_pages_evicted; + } return(TRUE); case BUF_LRU_NOT_FREED: @@ -1027,8 +1018,10 @@ buf_LRU_old_adjust_len(void) ut_a(buf_pool->LRU_old); //ut_ad(buf_pool_mutex_own()); ut_ad(mutex_own(&LRU_list_mutex)); -#if 3 * (BUF_LRU_OLD_MIN_LEN / 8) <= BUF_LRU_OLD_TOLERANCE + 5 -# error "3 * (BUF_LRU_OLD_MIN_LEN / 8) <= BUF_LRU_OLD_TOLERANCE + 5" + ut_ad(buf_LRU_old_ratio >= BUF_LRU_OLD_RATIO_MIN); + ut_ad(buf_LRU_old_ratio <= BUF_LRU_OLD_RATIO_MAX); +#if BUF_LRU_OLD_RATIO_MIN * BUF_LRU_OLD_MIN_LEN <= BUF_LRU_OLD_RATIO_DIV * (BUF_LRU_OLD_TOLERANCE + 5) +# error "BUF_LRU_OLD_RATIO_MIN * BUF_LRU_OLD_MIN_LEN <= BUF_LRU_OLD_RATIO_DIV * (BUF_LRU_OLD_TOLERANCE + 5)" #endif #ifdef UNIV_LRU_DEBUG /* buf_pool->LRU_old must be the first item in the LRU list @@ -1040,34 +1033,39 @@ buf_LRU_old_adjust_len(void) || UT_LIST_GET_NEXT(LRU, buf_pool->LRU_old)->old); #endif /* UNIV_LRU_DEBUG */ + old_len = buf_pool->LRU_old_len; + new_len = ut_min(UT_LIST_GET_LEN(buf_pool->LRU) + * buf_LRU_old_ratio / BUF_LRU_OLD_RATIO_DIV, + UT_LIST_GET_LEN(buf_pool->LRU) + - (BUF_LRU_OLD_TOLERANCE + + BUF_LRU_NON_OLD_MIN_LEN)); + for (;;) { - old_len = buf_pool->LRU_old_len; - new_len = 3 * (UT_LIST_GET_LEN(buf_pool->LRU) / 8); + buf_page_t* LRU_old = buf_pool->LRU_old; - ut_ad(buf_pool->LRU_old->in_LRU_list); - ut_a(buf_pool->LRU_old); + ut_a(LRU_old); + ut_ad(LRU_old->in_LRU_list); #ifdef UNIV_LRU_DEBUG - ut_a(buf_pool->LRU_old->old); + ut_a(LRU_old->old); #endif /* UNIV_LRU_DEBUG */ /* Update the LRU_old pointer if necessary */ - if (old_len < new_len - BUF_LRU_OLD_TOLERANCE) { + if (old_len + BUF_LRU_OLD_TOLERANCE < new_len) { - buf_pool->LRU_old = UT_LIST_GET_PREV( - LRU, buf_pool->LRU_old); + buf_pool->LRU_old = LRU_old = UT_LIST_GET_PREV( + LRU, LRU_old); #ifdef UNIV_LRU_DEBUG - ut_a(!buf_pool->LRU_old->old); + ut_a(!LRU_old->old); #endif /* UNIV_LRU_DEBUG */ - buf_page_set_old(buf_pool->LRU_old, TRUE); - buf_pool->LRU_old_len++; + old_len = ++buf_pool->LRU_old_len; + buf_page_set_old(LRU_old, TRUE); } else if (old_len > new_len + BUF_LRU_OLD_TOLERANCE) { - buf_page_set_old(buf_pool->LRU_old, FALSE); - buf_pool->LRU_old = UT_LIST_GET_NEXT( - LRU, buf_pool->LRU_old); - buf_pool->LRU_old_len--; + buf_pool->LRU_old = UT_LIST_GET_NEXT(LRU, LRU_old); + old_len = --buf_pool->LRU_old_len; + buf_page_set_old(LRU_old, FALSE); } else { return; } @@ -1092,12 +1090,13 @@ buf_LRU_old_init(void) the adjust function to move the LRU_old pointer to the right position */ - bpage = UT_LIST_GET_FIRST(buf_pool->LRU); - - while (bpage != NULL) { + for (bpage = UT_LIST_GET_LAST(buf_pool->LRU); bpage != NULL; + bpage = UT_LIST_GET_PREV(LRU, bpage)) { ut_ad(bpage->in_LRU_list); - buf_page_set_old(bpage, TRUE); - bpage = UT_LIST_GET_NEXT(LRU, bpage); + ut_ad(buf_page_in_file(bpage)); + /* This loop temporarily violates the + assertions of buf_page_set_old(). */ + bpage->old = TRUE; } buf_pool->LRU_old = UT_LIST_GET_FIRST(buf_pool->LRU); @@ -1152,16 +1151,19 @@ buf_LRU_remove_block( if (UNIV_UNLIKELY(bpage == buf_pool->LRU_old)) { - /* Below: the previous block is guaranteed to exist, because - the LRU_old pointer is only allowed to differ by the - tolerance value from strict 3/8 of the LRU list length. */ + /* Below: the previous block is guaranteed to exist, + because the LRU_old pointer is only allowed to differ + by BUF_LRU_OLD_TOLERANCE from strict + buf_LRU_old_ratio/BUF_LRU_OLD_RATIO_DIV of the LRU + list length. */ + buf_page_t* prev_bpage = UT_LIST_GET_PREV(LRU, bpage); - buf_pool->LRU_old = UT_LIST_GET_PREV(LRU, bpage); - ut_a(buf_pool->LRU_old); + ut_a(prev_bpage); #ifdef UNIV_LRU_DEBUG - ut_a(!buf_pool->LRU_old->old); + ut_a(!prev_bpage->old); #endif /* UNIV_LRU_DEBUG */ - buf_page_set_old(buf_pool->LRU_old, TRUE); + buf_pool->LRU_old = prev_bpage; + buf_page_set_old(prev_bpage, TRUE); buf_pool->LRU_old_len++; } @@ -1172,10 +1174,19 @@ buf_LRU_remove_block( buf_unzip_LRU_remove_block_if_needed(bpage); - /* If the LRU list is so short that LRU_old not defined, return */ + /* If the LRU list is so short that LRU_old is not defined, + clear the "old" flags and return */ if (UT_LIST_GET_LEN(buf_pool->LRU) < BUF_LRU_OLD_MIN_LEN) { + for (bpage = UT_LIST_GET_FIRST(buf_pool->LRU); bpage != NULL; + bpage = UT_LIST_GET_NEXT(LRU, bpage)) { + /* This loop temporarily violates the + assertions of buf_page_set_old(). */ + bpage->old = FALSE; + } + buf_pool->LRU_old = NULL; + buf_pool->LRU_old_len = 0; return; } @@ -1227,8 +1238,6 @@ buf_LRU_add_block_to_end_low( /*=========================*/ buf_page_t* bpage) /*!< in: control block */ { - buf_page_t* last_bpage; - ut_ad(buf_pool); ut_ad(bpage); //ut_ad(buf_pool_mutex_own()); @@ -1236,31 +1245,18 @@ buf_LRU_add_block_to_end_low( ut_a(buf_page_in_file(bpage)); - last_bpage = UT_LIST_GET_LAST(buf_pool->LRU); - - if (last_bpage) { - bpage->LRU_position = last_bpage->LRU_position; - } else { - bpage->LRU_position = buf_pool_clock_tic(); - } - ut_ad(!bpage->in_LRU_list); UT_LIST_ADD_LAST(LRU, buf_pool->LRU, bpage); bpage->in_LRU_list = TRUE; - buf_page_set_old(bpage, TRUE); - - if (UT_LIST_GET_LEN(buf_pool->LRU) >= BUF_LRU_OLD_MIN_LEN) { - - buf_pool->LRU_old_len++; - } - if (UT_LIST_GET_LEN(buf_pool->LRU) > BUF_LRU_OLD_MIN_LEN) { ut_ad(buf_pool->LRU_old); /* Adjust the length of the old block list if necessary */ + buf_page_set_old(bpage, TRUE); + buf_pool->LRU_old_len++; buf_LRU_old_adjust_len(); } else if (UT_LIST_GET_LEN(buf_pool->LRU) == BUF_LRU_OLD_MIN_LEN) { @@ -1269,6 +1265,8 @@ buf_LRU_add_block_to_end_low( defined: init it */ buf_LRU_old_init(); + } else { + buf_page_set_old(bpage, buf_pool->LRU_old != NULL); } /* If this is a zipped block with decompressed frame as well @@ -1302,7 +1300,6 @@ buf_LRU_add_block_low( UT_LIST_ADD_FIRST(LRU, buf_pool->LRU, bpage); - bpage->LRU_position = buf_pool_clock_tic(); bpage->freed_page_clock = buf_pool->freed_page_clock; } else { #ifdef UNIV_LRU_DEBUG @@ -1317,23 +1314,17 @@ buf_LRU_add_block_low( UT_LIST_INSERT_AFTER(LRU, buf_pool->LRU, buf_pool->LRU_old, bpage); buf_pool->LRU_old_len++; - - /* We copy the LRU position field of the previous block - to the new block */ - - bpage->LRU_position = (buf_pool->LRU_old)->LRU_position; } bpage->in_LRU_list = TRUE; - buf_page_set_old(bpage, old); - if (UT_LIST_GET_LEN(buf_pool->LRU) > BUF_LRU_OLD_MIN_LEN) { ut_ad(buf_pool->LRU_old); /* Adjust the length of the old block list if necessary */ + buf_page_set_old(bpage, old); buf_LRU_old_adjust_len(); } else if (UT_LIST_GET_LEN(buf_pool->LRU) == BUF_LRU_OLD_MIN_LEN) { @@ -1342,6 +1333,8 @@ buf_LRU_add_block_low( defined: init it */ buf_LRU_old_init(); + } else { + buf_page_set_old(bpage, buf_pool->LRU_old != NULL); } /* If this is a zipped block with decompressed frame as well @@ -1375,6 +1368,13 @@ buf_LRU_make_block_young( /*=====================*/ buf_page_t* bpage) /*!< in: control block */ { + //ut_ad(buf_pool_mutex_own()); + ut_ad(mutex_own(&LRU_list_mutex)); + + if (bpage->old) { + buf_pool->stat.n_pages_made_young++; + } + buf_LRU_remove_block(bpage); buf_LRU_add_block_low(bpage, FALSE); } @@ -1571,15 +1571,6 @@ not_freed: buf_pool->LRU_old = b; } -#ifdef UNIV_LRU_DEBUG - ut_a(prev_b->old - || !UT_LIST_GET_NEXT(LRU, b) - || UT_LIST_GET_NEXT(LRU, b)->old); - } else { - ut_a(!prev_b->old - || !UT_LIST_GET_NEXT(LRU, b) - || !UT_LIST_GET_NEXT(LRU, b)->old); -#endif /* UNIV_LRU_DEBUG */ } lru_len = UT_LIST_GET_LEN(buf_pool->LRU); @@ -1595,6 +1586,11 @@ not_freed: defined: init it */ buf_LRU_old_init(); } +#ifdef UNIV_LRU_DEBUG + /* Check that the "old" flag is consistent + in the block and its neighbours. */ + buf_page_set_old(b, buf_page_is_old(b)); +#endif /* UNIV_LRU_DEBUG */ } else { b->in_LRU_list = FALSE; buf_LRU_add_block_low(b, buf_page_is_old(b)); @@ -1985,6 +1981,52 @@ buf_LRU_block_free_hashed_page( buf_LRU_block_free_non_file_page(block, have_page_hash_mutex); } +/**********************************************************************//** +Updates buf_LRU_old_ratio. +@return updated old_pct */ +UNIV_INTERN +uint +buf_LRU_old_ratio_update( +/*=====================*/ + uint old_pct,/*!< in: Reserve this percentage of + the buffer pool for "old" blocks. */ + ibool adjust) /*!< in: TRUE=adjust the LRU list; + FALSE=just assign buf_LRU_old_ratio + during the initialization of InnoDB */ +{ + uint ratio; + + ratio = old_pct * BUF_LRU_OLD_RATIO_DIV / 100; + if (ratio < BUF_LRU_OLD_RATIO_MIN) { + ratio = BUF_LRU_OLD_RATIO_MIN; + } else if (ratio > BUF_LRU_OLD_RATIO_MAX) { + ratio = BUF_LRU_OLD_RATIO_MAX; + } + + if (adjust) { + //buf_pool_mutex_enter(); + mutex_enter(&LRU_list_mutex); + + if (ratio != buf_LRU_old_ratio) { + buf_LRU_old_ratio = ratio; + + if (UT_LIST_GET_LEN(buf_pool->LRU) + >= BUF_LRU_OLD_MIN_LEN) { + buf_LRU_old_adjust_len(); + } + } + + //buf_pool_mutex_exit(); + mutex_exit(&LRU_list_mutex); + } else { + buf_LRU_old_ratio = ratio; + } + + /* the reverse of + ratio = old_pct * BUF_LRU_OLD_RATIO_DIV / 100 */ + return((uint) (ratio * 100 / (double) BUF_LRU_OLD_RATIO_DIV + 0.5)); +} + /********************************************************************//** Update the historical stats that we are collecting for LRU eviction policy at the end of each interval. */ @@ -2023,6 +2065,218 @@ func_exit: memset(&buf_LRU_stat_cur, 0, sizeof buf_LRU_stat_cur); } +/********************************************************************//** +Dump the LRU page list to the specific file. */ +#define LRU_DUMP_FILE "ib_lru_dump" + +UNIV_INTERN +ibool +buf_LRU_file_dump(void) +/*===================*/ +{ + os_file_t dump_file = -1; + ibool success; + byte* buffer_base = NULL; + byte* buffer = NULL; + buf_page_t* bpage; + ulint buffers; + ulint offset; + ibool ret = FALSE; + ulint i; + + for (i = 0; i < srv_n_data_files; i++) { + if (strstr(srv_data_file_names[i], LRU_DUMP_FILE) != NULL) { + fprintf(stderr, + " InnoDB: The name '%s' seems to be used for" + " innodb_data_file_path. Dumping LRU list is not" + " done for safeness.\n", LRU_DUMP_FILE); + goto end; + } + } + + buffer_base = ut_malloc(2 * UNIV_PAGE_SIZE); + buffer = ut_align(buffer_base, UNIV_PAGE_SIZE); + if (!buffer) { + fprintf(stderr, + " InnoDB: cannot allocate buffer.\n"); + goto end; + } + + dump_file = os_file_create(LRU_DUMP_FILE, OS_FILE_OVERWRITE, + OS_FILE_NORMAL, OS_DATA_FILE, &success); + if (!success) { + os_file_get_last_error(TRUE); + fprintf(stderr, + " InnoDB: cannot open %s\n", LRU_DUMP_FILE); + goto end; + } + + mutex_enter(&LRU_list_mutex); + bpage = UT_LIST_GET_LAST(buf_pool->LRU); + + buffers = offset = 0; + while (bpage != NULL) { + if (offset == 0) { + memset(buffer, 0, UNIV_PAGE_SIZE); + } + + mach_write_to_4(buffer + offset * 4, bpage->space); + offset++; + mach_write_to_4(buffer + offset * 4, bpage->offset); + offset++; + + if (offset == UNIV_PAGE_SIZE/4) { + success = os_file_write(LRU_DUMP_FILE, dump_file, buffer, + (buffers << UNIV_PAGE_SIZE_SHIFT) & 0xFFFFFFFFUL, + (buffers >> (32 - UNIV_PAGE_SIZE_SHIFT)), + UNIV_PAGE_SIZE); + if (!success) { + mutex_exit(&LRU_list_mutex); + fprintf(stderr, + " InnoDB: cannot write page %lu of %s\n", + buffers, LRU_DUMP_FILE); + goto end; + } + buffers++; + offset = 0; + } + + bpage = UT_LIST_GET_PREV(LRU, bpage); + } + mutex_exit(&LRU_list_mutex); + + if (offset == 0) { + memset(buffer, 0, UNIV_PAGE_SIZE); + } + + mach_write_to_4(buffer + offset * 4, 0xFFFFFFFFUL); + offset++; + mach_write_to_4(buffer + offset * 4, 0xFFFFFFFFUL); + offset++; + + success = os_file_write(LRU_DUMP_FILE, dump_file, buffer, + (buffers << UNIV_PAGE_SIZE_SHIFT) & 0xFFFFFFFFUL, + (buffers >> (32 - UNIV_PAGE_SIZE_SHIFT)), + UNIV_PAGE_SIZE); + if (!success) { + goto end; + } + + ret = TRUE; +end: + if (dump_file != -1) + os_file_close(dump_file); + if (buffer_base) + ut_free(buffer_base); + + return(ret); +} +/********************************************************************//** +Read the pages based on the specific file.*/ +UNIV_INTERN +ibool +buf_LRU_file_restore(void) +/*======================*/ +{ + os_file_t dump_file = -1; + ibool success; + byte* buffer_base = NULL; + byte* buffer = NULL; + ulint buffers; + ulint offset; + ulint reads = 0; + ulint req = 0; + ibool terminated = FALSE; + ibool ret = FALSE; + + buffer_base = ut_malloc(2 * UNIV_PAGE_SIZE); + buffer = ut_align(buffer_base, UNIV_PAGE_SIZE); + if (!buffer) { + fprintf(stderr, + " InnoDB: cannot allocate buffer.\n"); + goto end; + } + + dump_file = os_file_create_simple_no_error_handling( + LRU_DUMP_FILE, OS_FILE_OPEN, OS_FILE_READ_ONLY, &success); + if (!success) { + os_file_get_last_error(TRUE); + fprintf(stderr, + " InnoDB: cannot open %s\n", LRU_DUMP_FILE); + goto end; + } + + buffers = 0; + while (!terminated) { + success = os_file_read(dump_file, buffer, + (buffers << UNIV_PAGE_SIZE_SHIFT) & 0xFFFFFFFFUL, + (buffers >> (32 - UNIV_PAGE_SIZE_SHIFT)), + UNIV_PAGE_SIZE); + if (!success) { + fprintf(stderr, + " InnoDB: cannot read page %lu of %s," + " or meet unexpected terminal.", + buffers, LRU_DUMP_FILE); + goto end; + } + + for (offset = 0; offset < UNIV_PAGE_SIZE/4; offset += 2) { + ulint space_id, zip_size, page_no; + ulint err; + ib_int64_t tablespace_version; + + space_id = mach_read_from_4(buffer + offset * 4); + page_no = mach_read_from_4(buffer + (offset + 1) * 4); + if (space_id == 0xFFFFFFFFUL + || page_no == 0xFFFFFFFFUL) { + terminated = TRUE; + break; + } + + if (offset % 16 == 15) { + os_aio_simulated_wake_handler_threads(); + buf_flush_free_margin(FALSE); + } + + zip_size = fil_space_get_zip_size(space_id); + if (UNIV_UNLIKELY(zip_size == ULINT_UNDEFINED)) { + continue; + } + + if (fil_area_is_exist(space_id, zip_size, page_no, 0, + zip_size ? zip_size : UNIV_PAGE_SIZE)) { + + tablespace_version = fil_space_get_version(space_id); + + req++; + reads += buf_read_page_low(&err, FALSE, BUF_READ_ANY_PAGE + | OS_AIO_SIMULATED_WAKE_LATER, + space_id, zip_size, TRUE, + tablespace_version, page_no, NULL); + buf_LRU_stat_inc_io(); + } + } + + buffers++; + } + + os_aio_simulated_wake_handler_threads(); + buf_flush_free_margin(FALSE); + + ut_print_timestamp(stderr); + fprintf(stderr, + " InnoDB: reading pages based on the dumped LRU list was done." + " (requested: %lu, read: %lu)\n", req, reads); + ret = TRUE; +end: + if (dump_file != -1) + os_file_close(dump_file); + if (buffer_base) + ut_free(buffer_base); + + return(ret); +} + #if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG /**********************************************************************//** Validates the LRU list. @@ -2036,7 +2290,6 @@ buf_LRU_validate(void) buf_block_t* block; ulint old_len; ulint new_len; - ulint LRU_pos; ut_ad(buf_pool); //buf_pool_mutex_enter(); @@ -2046,7 +2299,11 @@ buf_LRU_validate(void) ut_a(buf_pool->LRU_old); old_len = buf_pool->LRU_old_len; - new_len = 3 * (UT_LIST_GET_LEN(buf_pool->LRU) / 8); + new_len = ut_min(UT_LIST_GET_LEN(buf_pool->LRU) + * buf_LRU_old_ratio / BUF_LRU_OLD_RATIO_DIV, + UT_LIST_GET_LEN(buf_pool->LRU) + - (BUF_LRU_OLD_TOLERANCE + + BUF_LRU_NON_OLD_MIN_LEN)); ut_a(old_len >= new_len - BUF_LRU_OLD_TOLERANCE); ut_a(old_len <= new_len + BUF_LRU_OLD_TOLERANCE); } @@ -2077,28 +2334,24 @@ buf_LRU_validate(void) } if (buf_page_is_old(bpage)) { - old_len++; - } + const buf_page_t* prev + = UT_LIST_GET_PREV(LRU, bpage); + const buf_page_t* next + = UT_LIST_GET_NEXT(LRU, bpage); - if (buf_pool->LRU_old && (old_len == 1)) { - ut_a(buf_pool->LRU_old == bpage); - } + if (!old_len++) { + ut_a(buf_pool->LRU_old == bpage); + } else { + ut_a(!prev || buf_page_is_old(prev)); + } - LRU_pos = buf_page_get_LRU_position(bpage); + ut_a(!next || buf_page_is_old(next)); + } bpage = UT_LIST_GET_NEXT(LRU, bpage); - - if (bpage) { - /* If the following assert fails, it may - not be an error: just the buf_pool clock - has wrapped around */ - ut_a(LRU_pos >= buf_page_get_LRU_position(bpage)); - } } - if (buf_pool->LRU_old) { - ut_a(buf_pool->LRU_old_len == old_len); - } + ut_a(buf_pool->LRU_old_len == old_len); mutex_exit(&LRU_list_mutex); mutex_enter(&free_list_mutex); @@ -2149,9 +2402,6 @@ buf_LRU_print(void) //buf_pool_mutex_enter(); mutex_enter(&LRU_list_mutex); - fprintf(stderr, "Pool ulint clock %lu\n", - (ulong) buf_pool->ulint_clock); - bpage = UT_LIST_GET_FIRST(buf_pool->LRU); while (bpage != NULL) { @@ -2182,18 +2432,16 @@ buf_LRU_print(void) const byte* frame; case BUF_BLOCK_FILE_PAGE: frame = buf_block_get_frame((buf_block_t*) bpage); - fprintf(stderr, "\nLRU pos %lu type %lu" + fprintf(stderr, "\ntype %lu" " index id %lu\n", - (ulong) buf_page_get_LRU_position(bpage), (ulong) fil_page_get_type(frame), (ulong) ut_dulint_get_low( btr_page_get_index_id(frame))); break; case BUF_BLOCK_ZIP_PAGE: frame = bpage->zip.data; - fprintf(stderr, "\nLRU pos %lu type %lu size %lu" + fprintf(stderr, "\ntype %lu size %lu" " index id %lu\n", - (ulong) buf_page_get_LRU_position(bpage), (ulong) fil_page_get_type(frame), (ulong) buf_page_get_zip_size(bpage), (ulong) ut_dulint_get_low( @@ -2201,8 +2449,7 @@ buf_LRU_print(void) break; default: - fprintf(stderr, "\nLRU pos %lu !state %lu!\n", - (ulong) buf_page_get_LRU_position(bpage), + fprintf(stderr, "\n!state %lu!\n", (ulong) buf_page_get_state(bpage)); break; } diff --git a/storage/xtradb/buf/buf0rea.c b/storage/xtradb/buf/buf0rea.c index f2dbe939c92..08aaac788d5 100644 --- a/storage/xtradb/buf/buf0rea.c +++ b/storage/xtradb/buf/buf0rea.c @@ -38,14 +38,6 @@ Created 11/5/1995 Heikki Tuuri #include "srv0start.h" #include "srv0srv.h" -/** The size in blocks of the area where the random read-ahead algorithm counts -the accessed pages when deciding whether to read-ahead */ -#define BUF_READ_AHEAD_RANDOM_AREA BUF_READ_AHEAD_AREA - -/** There must be at least this many pages in buf_pool in the area to start -a random read-ahead */ -#define BUF_READ_AHEAD_RANDOM_THRESHOLD (1 + BUF_READ_AHEAD_RANDOM_AREA / 2) - /** The linear read-ahead area size */ #define BUF_READ_AHEAD_LINEAR_AREA BUF_READ_AHEAD_AREA @@ -62,8 +54,9 @@ flag is cleared and the x-lock released by an i/o-handler thread. @return 1 if a read request was queued, 0 if the page already resided in buf_pool, or if the page is in the doublewrite buffer blocks in which case it is never read into the pool, or if the tablespace does -not exist or is being dropped */ -static +not exist or is being dropped +@return 1 if read request is issued. 0 if it is not */ +UNIV_INTERN ulint buf_read_page_low( /*==============*/ @@ -82,7 +75,8 @@ buf_read_page_low( treat the tablespace as dropped; this is a timestamp we use to stop dangling page reads from a tablespace which we have DISCARDed + IMPORTed back */ - ulint offset) /*!< in: page number */ + ulint offset, /*!< in: page number */ + trx_t* trx) { buf_page_t* bpage; ulint wake_later; @@ -183,15 +177,15 @@ not_to_recover: ut_ad(buf_page_in_file(bpage)); if (zip_size) { - *err = fil_io(OS_FILE_READ | wake_later, + *err = _fil_io(OS_FILE_READ | wake_later, sync, space, zip_size, offset, 0, zip_size, - bpage->zip.data, bpage); + bpage->zip.data, bpage, trx); } else { ut_a(buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE); - *err = fil_io(OS_FILE_READ | wake_later, + *err = _fil_io(OS_FILE_READ | wake_later, sync, space, 0, offset, 0, UNIV_PAGE_SIZE, - ((buf_block_t*) bpage)->frame, bpage); + ((buf_block_t*) bpage)->frame, bpage, trx); } ut_a(*err == DB_SUCCESS); @@ -205,206 +199,33 @@ not_to_recover: } /********************************************************************//** -Applies a random read-ahead in buf_pool if there are at least a threshold -value of accessed pages from the random read-ahead area. Does not read any -page, not even the one at the position (space, offset), if the read-ahead -mechanism is not activated. NOTE 1: the calling thread may own latches on -pages: to avoid deadlocks this function must be written such that it cannot -end up waiting for these latches! NOTE 2: the calling thread must want -access to the page given: this rule is set to prevent unintended read-aheads -performed by ibuf routines, a situation which could result in a deadlock if -the OS does not support asynchronous i/o. -@return number of page read requests issued; NOTE that if we read ibuf -pages, it may happen that the page at the given page number does not -get read even if we return a positive value! */ -static -ulint -buf_read_ahead_random( -/*==================*/ - ulint space, /*!< in: space id */ - ulint zip_size,/*!< in: compressed page size in bytes, or 0 */ - ulint offset) /*!< in: page number of a page which the current thread - wants to access */ -{ - ib_int64_t tablespace_version; - ulint recent_blocks = 0; - ulint count; - ulint LRU_recent_limit; - ulint ibuf_mode; - ulint low, high; - ulint err; - ulint i; - ulint buf_read_ahead_random_area; - -// /* We have currently disabled random readahead */ -// return(0); - - if (!(srv_read_ahead & 1)) { - return(0); - } - - if (srv_startup_is_before_trx_rollback_phase) { - /* No read-ahead to avoid thread deadlocks */ - return(0); - } - - if (ibuf_bitmap_page(zip_size, offset) - || trx_sys_hdr_page(space, offset)) { - - /* If it is an ibuf bitmap page or trx sys hdr, we do - no read-ahead, as that could break the ibuf page access - order */ - - return(0); - } - - /* Remember the tablespace version before we ask te tablespace size - below: if DISCARD + IMPORT changes the actual .ibd file meanwhile, we - do not try to read outside the bounds of the tablespace! */ - - tablespace_version = fil_space_get_version(space); - - buf_read_ahead_random_area = BUF_READ_AHEAD_RANDOM_AREA; - - low = (offset / buf_read_ahead_random_area) - * buf_read_ahead_random_area; - high = (offset / buf_read_ahead_random_area + 1) - * buf_read_ahead_random_area; - if (high > fil_space_get_size(space)) { - - high = fil_space_get_size(space); - } - - /* Get the minimum LRU_position field value for an initial segment - of the LRU list, to determine which blocks have recently been added - to the start of the list. */ - - LRU_recent_limit = buf_LRU_get_recent_limit(); - - //buf_pool_mutex_enter(); - mutex_enter(&buf_pool_mutex); - - if (buf_pool->n_pend_reads - > buf_pool->curr_size / BUF_READ_AHEAD_PEND_LIMIT) { - //buf_pool_mutex_exit(); - mutex_exit(&buf_pool_mutex); - - return(0); - } - mutex_exit(&buf_pool_mutex); - - /* Count how many blocks in the area have been recently accessed, - that is, reside near the start of the LRU list. */ - - rw_lock_s_lock(&page_hash_latch); - for (i = low; i < high; i++) { - const buf_page_t* bpage = buf_page_hash_get(space, i); - - if (bpage - && buf_page_is_accessed(bpage) - && (buf_page_get_LRU_position(bpage) > LRU_recent_limit)) { - - recent_blocks++; - - if (recent_blocks >= BUF_READ_AHEAD_RANDOM_THRESHOLD) { - - //buf_pool_mutex_exit(); - rw_lock_s_unlock(&page_hash_latch); - goto read_ahead; - } - } - } - - //buf_pool_mutex_exit(); - rw_lock_s_unlock(&page_hash_latch); - /* Do nothing */ - return(0); - -read_ahead: - /* Read all the suitable blocks within the area */ - - if (ibuf_inside()) { - ibuf_mode = BUF_READ_IBUF_PAGES_ONLY; - } else { - ibuf_mode = BUF_READ_ANY_PAGE; - } - - count = 0; - - for (i = low; i < high; i++) { - /* It is only sensible to do read-ahead in the non-sync aio - mode: hence FALSE as the first parameter */ - - if (!ibuf_bitmap_page(zip_size, i)) { - count += buf_read_page_low( - &err, FALSE, - ibuf_mode | OS_AIO_SIMULATED_WAKE_LATER, - space, zip_size, FALSE, - tablespace_version, i); - if (err == DB_TABLESPACE_DELETED) { - ut_print_timestamp(stderr); - fprintf(stderr, - " InnoDB: Warning: in random" - " readahead trying to access\n" - "InnoDB: tablespace %lu page %lu,\n" - "InnoDB: but the tablespace does not" - " exist or is just being dropped.\n", - (ulong) space, (ulong) i); - } - } - } - - /* In simulated aio we wake the aio handler threads only after - queuing all aio requests, in native aio the following call does - nothing: */ - - os_aio_simulated_wake_handler_threads(); - -#ifdef UNIV_DEBUG - if (buf_debug_prints && (count > 0)) { - fprintf(stderr, - "Random read-ahead space %lu offset %lu pages %lu\n", - (ulong) space, (ulong) offset, - (ulong) count); - } -#endif /* UNIV_DEBUG */ - - ++srv_read_ahead_rnd; - return(count); -} - -/********************************************************************//** High-level function which reads a page asynchronously from a file to the buffer buf_pool if it is not already there. Sets the io_fix flag and sets an exclusive lock on the buffer frame. The flag is cleared and the x-lock -released by the i/o-handler thread. Does a random read-ahead if it seems -sensible. -@return number of page read requests issued: this can be greater than -1 if read-ahead occurred */ +released by the i/o-handler thread. +@return TRUE if page has been read in, FALSE in case of failure */ UNIV_INTERN -ulint +ibool buf_read_page( /*==========*/ ulint space, /*!< in: space id */ ulint zip_size,/*!< in: compressed page size in bytes, or 0 */ - ulint offset) /*!< in: page number */ + ulint offset, /*!< in: page number */ + trx_t* trx) { ib_int64_t tablespace_version; ulint count; - ulint count2; ulint err; tablespace_version = fil_space_get_version(space); - count = buf_read_ahead_random(space, zip_size, offset); - /* We do the i/o in the synchronous aio mode to save thread switches: hence TRUE */ - count2 = buf_read_page_low(&err, TRUE, BUF_READ_ANY_PAGE, space, - zip_size, FALSE, - tablespace_version, offset); - srv_buf_pool_reads+= count2; + count = buf_read_page_low(&err, TRUE, BUF_READ_ANY_PAGE, space, + zip_size, FALSE, + tablespace_version, offset, trx); + srv_buf_pool_reads += count; if (err == DB_TABLESPACE_DELETED) { ut_print_timestamp(stderr); fprintf(stderr, @@ -421,14 +242,14 @@ buf_read_page( /* Increment number of I/O operations used for LRU policy. */ buf_LRU_stat_inc_io(); - return(count + count2); + return(count > 0); } /********************************************************************//** Applies linear read-ahead if in the buf_pool the page is a border page of a linear read-ahead area and all the pages in the area have been accessed. Does not read any page if the read-ahead mechanism is not activated. Note -that the the algorithm looks at the 'natural' adjacent successor and +that the algorithm looks at the 'natural' adjacent successor and predecessor of the page, which on the leaf level of a B-tree are the next and previous page in the chain of leaves. To know these, the page specified in (space, offset) must already be present in the buf_pool. Thus, the @@ -454,8 +275,9 @@ buf_read_ahead_linear( /*==================*/ ulint space, /*!< in: space id */ ulint zip_size,/*!< in: compressed page size in bytes, or 0 */ - ulint offset) /*!< in: page number of a page; NOTE: the current thread + ulint offset, /*!< in: page number of a page; NOTE: the current thread must want access to this page (see NOTE 3 above) */ + trx_t* trx) { ib_int64_t tablespace_version; buf_page_t* bpage; @@ -557,9 +379,17 @@ buf_read_ahead_linear( fail_count++; } else if (pred_bpage) { - int res = (ut_ulint_cmp( - buf_page_get_LRU_position(bpage), - buf_page_get_LRU_position(pred_bpage))); + /* Note that buf_page_is_accessed() returns + the time of the first access. If some blocks + of the extent existed in the buffer pool at + the time of a linear access pattern, the first + access times may be nonmonotonic, even though + the latest access times were linear. The + threshold (srv_read_ahead_factor) should help + a little against this. */ + int res = ut_ulint_cmp( + buf_page_is_accessed(bpage), + buf_page_is_accessed(pred_bpage)); /* Accesses not in the right order */ if (res != 0 && res != asc_or_desc) { fail_count++; @@ -670,7 +500,7 @@ buf_read_ahead_linear( count += buf_read_page_low( &err, FALSE, ibuf_mode | OS_AIO_SIMULATED_WAKE_LATER, - space, zip_size, FALSE, tablespace_version, i); + space, zip_size, FALSE, tablespace_version, i, trx); if (err == DB_TABLESPACE_DELETED) { ut_print_timestamp(stderr); fprintf(stderr, @@ -705,7 +535,7 @@ buf_read_ahead_linear( LRU policy decision. */ buf_LRU_stat_inc_io(); - ++srv_read_ahead_seq; + buf_pool->stat.n_ra_pages_read += count; return(count); } @@ -760,7 +590,7 @@ buf_read_ibuf_merge_pages( buf_read_page_low(&err, sync && (i + 1 == n_stored), BUF_READ_ANY_PAGE, space_ids[i], zip_size, TRUE, space_versions[i], - page_nos[i]); + page_nos[i], NULL); if (UNIV_UNLIKELY(err == DB_TABLESPACE_DELETED)) { tablespace_deleted: @@ -857,12 +687,12 @@ buf_read_recv_pages( if ((i + 1 == n_stored) && sync) { buf_read_page_low(&err, TRUE, BUF_READ_ANY_PAGE, space, zip_size, TRUE, tablespace_version, - page_nos[i]); + page_nos[i], NULL); } else { buf_read_page_low(&err, FALSE, BUF_READ_ANY_PAGE | OS_AIO_SIMULATED_WAKE_LATER, space, zip_size, TRUE, - tablespace_version, page_nos[i]); + tablespace_version, page_nos[i], NULL); } } diff --git a/storage/xtradb/data/data0type.c b/storage/xtradb/data/data0type.c index 8429775e7d8..e834fd2ec55 100644 --- a/storage/xtradb/data/data0type.c +++ b/storage/xtradb/data/data0type.c @@ -237,6 +237,22 @@ dtype_print( fputs("DATA_SYS", stderr); break; + case DATA_FLOAT: + fputs("DATA_FLOAT", stderr); + break; + + case DATA_DOUBLE: + fputs("DATA_DOUBLE", stderr); + break; + + case DATA_DECIMAL: + fputs("DATA_DECIMAL", stderr); + break; + + case DATA_VARMYSQL: + fputs("DATA_VARMYSQL", stderr); + break; + default: fprintf(stderr, "type %lu", (ulong) mtype); break; diff --git a/storage/xtradb/dict/dict0crea.c b/storage/xtradb/dict/dict0crea.c index 10e3ff409f3..f548d89bc68 100644 --- a/storage/xtradb/dict/dict0crea.c +++ b/storage/xtradb/dict/dict0crea.c @@ -1387,7 +1387,7 @@ dict_create_add_foreign_field_to_dictionary( Add a single foreign key definition to the data dictionary tables in the database. We also generate names to constraints that were not named by the user. A generated constraint has a name of the format -databasename/tablename_ibfk_<number>, where the numbers start from 1, and +databasename/tablename_ibfk_NUMBER, where the numbers start from 1, and are given locally for this table, that is, the number is not global, as in the old format constraints < 4.0.18 it used to be. @return error code or DB_SUCCESS */ diff --git a/storage/xtradb/dict/dict0dict.c b/storage/xtradb/dict/dict0dict.c index 050182bc831..3f5dd2930e3 100644 --- a/storage/xtradb/dict/dict0dict.c +++ b/storage/xtradb/dict/dict0dict.c @@ -82,9 +82,10 @@ static char dict_ibfk[] = "_ibfk_"; /*******************************************************************//** Tries to find column names for the index and sets the col field of the -index. */ +index. +@return TRUE if the column names were found */ static -void +ibool dict_index_find_cols( /*=================*/ dict_table_t* table, /*!< in: table */ @@ -1261,7 +1262,7 @@ dict_index_too_big_for_undo( = TRX_UNDO_PAGE_HDR - TRX_UNDO_PAGE_HDR_SIZE + 2 /* next record pointer */ + 1 /* type_cmpl */ - + 11 /* trx->undo_no */ - 11 /* table->id */ + + 11 /* trx->undo_no */ + 11 /* table->id */ + 1 /* rec_get_info_bits() */ + 11 /* DB_TRX_ID */ + 11 /* DB_ROLL_PTR */ @@ -1440,7 +1441,11 @@ dict_index_too_big_for_tree( goto add_field_size; } + if (srv_relax_table_creation) { + field_max_size = dict_col_get_min_size(col); + } else { field_max_size = dict_col_get_max_size(col); + } field_ext_max_size = field_max_size < 256 ? 1 : 2; if (field->prefix_len) { @@ -1493,7 +1498,7 @@ add_field_size: /**********************************************************************//** Adds an index to the dictionary cache. -@return DB_SUCCESS or DB_TOO_BIG_RECORD */ +@return DB_SUCCESS, DB_TOO_BIG_RECORD, or DB_CORRUPTION */ UNIV_INTERN ulint dict_index_add_to_cache( @@ -1519,7 +1524,10 @@ dict_index_add_to_cache( ut_a(!dict_index_is_clust(index) || UT_LIST_GET_LEN(table->indexes) == 0); - dict_index_find_cols(table, index); + if (!dict_index_find_cols(table, index)) { + + return(DB_CORRUPTION); + } /* Build the cache internal representation of the index, containing also the added system fields */ @@ -1732,9 +1740,10 @@ dict_index_remove_from_cache( /*******************************************************************//** Tries to find column names for the index and sets the col field of the -index. */ +index. +@return TRUE if the column names were found */ static -void +ibool dict_index_find_cols( /*=================*/ dict_table_t* table, /*!< in: table */ @@ -1759,17 +1768,21 @@ dict_index_find_cols( } } +#ifdef UNIV_DEBUG /* It is an error not to find a matching column. */ fputs("InnoDB: Error: no matching column for ", stderr); ut_print_name(stderr, NULL, FALSE, field->name); fputs(" in ", stderr); dict_index_name_print(stderr, NULL, index); fputs("!\n", stderr); - ut_error; +#endif /* UNIV_DEBUG */ + return(FALSE); found: ; } + + return(TRUE); } #endif /* !UNIV_HOTBACKUP */ @@ -4711,6 +4724,26 @@ dict_ind_init(void) dict_ind_redundant->cached = dict_ind_compact->cached = TRUE; } +/**********************************************************************//** +Frees dict_ind_redundant and dict_ind_compact. */ +static +void +dict_ind_free(void) +/*===============*/ +{ + dict_table_t* table; + + table = dict_ind_compact->table; + dict_mem_index_free(dict_ind_compact); + dict_ind_compact = NULL; + dict_mem_table_free(table); + + table = dict_ind_redundant->table; + dict_mem_index_free(dict_ind_redundant); + dict_ind_redundant = NULL; + dict_mem_table_free(table); +} + #ifndef UNIV_HOTBACKUP /**********************************************************************//** Get index by name @@ -4836,4 +4869,55 @@ dict_table_check_for_dup_indexes( } } #endif /* UNIV_DEBUG */ + +/************************************************************************** +Closes the data dictionary module. */ +UNIV_INTERN +void +dict_close(void) +/*============*/ +{ + ulint i; + + /* Free the hash elements. We don't remove them from the table + because we are going to destroy the table anyway. */ + for (i = 0; i < hash_get_n_cells(dict_sys->table_hash); i++) { + dict_table_t* table; + + table = HASH_GET_FIRST(dict_sys->table_hash, i); + + while (table) { + dict_table_t* prev_table = table; + + table = HASH_GET_NEXT(name_hash, prev_table); +#ifdef UNIV_DEBUG + ut_a(prev_table->magic_n == DICT_TABLE_MAGIC_N); +#endif + /* Acquire only because it's a pre-condition. */ + mutex_enter(&dict_sys->mutex); + + dict_table_remove_from_cache(prev_table); + + mutex_exit(&dict_sys->mutex); + } + } + + hash_table_free(dict_sys->table_hash); + + /* The elements are the same instance as in dict_sys->table_hash, + therefore we don't delete the individual elements. */ + hash_table_free(dict_sys->table_id_hash); + + dict_ind_free(); + + mutex_free(&dict_sys->mutex); + + rw_lock_free(&dict_operation_lock); + memset(&dict_operation_lock, 0x0, sizeof(dict_operation_lock)); + + mutex_free(&dict_foreign_err_mutex); + + mem_free(dict_sys); + dict_sys = NULL; +} #endif /* !UNIV_HOTBACKUP */ diff --git a/storage/xtradb/fil/fil0fil.c b/storage/xtradb/fil/fil0fil.c index 8d81fd41d58..9fa401a4b3e 100644 --- a/storage/xtradb/fil/fil0fil.c +++ b/storage/xtradb/fil/fil0fil.c @@ -327,6 +327,17 @@ fil_get_space_id_for_table( /*=======================*/ const char* name); /*!< in: table name in the standard 'databasename/tablename' format */ +/*******************************************************************//** +Frees a space object from the tablespace memory cache. Closes the files in +the chain but does not delete them. There must not be any pending i/o's or +flushes on the files. */ +static +ibool +fil_space_free( +/*===========*/ + /* out: TRUE if success */ + ulint id, /* in: space id */ + ibool own_mutex);/* in: TRUE if own system->mutex */ /********************************************************************//** Reads data from a space to a buffer. Remember that the possible incomplete blocks at the end of file are ignored: they are not taken into account when @@ -600,6 +611,11 @@ fil_node_create( UT_LIST_ADD_LAST(chain, space->chain, node); + if (id < SRV_LOG_SPACE_FIRST_ID && fil_system->max_assigned_id < id) { + + fil_system->max_assigned_id = id; + } + mutex_exit(&fil_system->mutex); } @@ -619,12 +635,10 @@ fil_node_open_file( ulint size_high; ibool ret; ibool success; -#ifndef UNIV_HOTBACKUP byte* buf2; byte* page; ulint space_id; ulint flags; -#endif /* !UNIV_HOTBACKUP */ ut_ad(mutex_own(&(system->mutex))); ut_a(node->n_pending == 0); @@ -660,9 +674,12 @@ fil_node_open_file( size_bytes = (((ib_int64_t)size_high) << 32) + (ib_int64_t)size_low; #ifdef UNIV_HOTBACKUP - node->size = (ulint) (size_bytes / UNIV_PAGE_SIZE); - /* TODO: adjust to zip_size, like below? */ -#else + if (space->id == 0) { + node->size = (ulint) (size_bytes / UNIV_PAGE_SIZE); + os_file_close(node->handle); + goto add_size; + } +#endif /* UNIV_HOTBACKUP */ ut_a(space->purpose != FIL_LOG); ut_a(space->id != 0); @@ -741,7 +758,10 @@ fil_node_open_file( (size_bytes / dict_table_flags_to_zip_size(flags)); } -#endif + +#ifdef UNIV_HOTBACKUP +add_size: +#endif /* UNIV_HOTBACKUP */ space->size += node->size; } @@ -961,7 +981,7 @@ close_more: " while the maximum\n" "InnoDB: allowed value would be %lu.\n" "InnoDB: You may need to raise the value of" - " innodb_max_files_open in\n" + " innodb_open_files in\n" "InnoDB: my.cnf.\n", (ulong) fil_system->n_open, (ulong) fil_system->max_n_open); @@ -1141,7 +1161,7 @@ try_again: mutex_exit(&fil_system->mutex); - fil_space_free(namesake_id); + fil_space_free(namesake_id, FALSE); goto try_again; } @@ -1266,17 +1286,21 @@ Frees a space object from the tablespace memory cache. Closes the files in the chain but does not delete them. There must not be any pending i/o's or flushes on the files. @return TRUE if success */ -UNIV_INTERN +static ibool fil_space_free( /*===========*/ - ulint id) /*!< in: space id */ + /* out: TRUE if success */ + ulint id, /* in: space id */ + ibool own_mutex) /* in: TRUE if own system->mutex */ { fil_space_t* space; fil_space_t* namespace; fil_node_t* fil_node; - mutex_enter(&fil_system->mutex); + if (!own_mutex) { + mutex_enter(&fil_system->mutex); + } space = fil_space_get_by_id(id); @@ -1323,7 +1347,9 @@ fil_space_free( ut_a(0 == UT_LIST_GET_LEN(space->chain)); - mutex_exit(&fil_system->mutex); + if (!own_mutex) { + mutex_exit(&fil_system->mutex); + } rw_lock_free(&(space->latch)); @@ -1541,7 +1567,7 @@ fil_open_log_and_system_tablespace_files(void) fprintf(stderr, "InnoDB: Warning: you must" " raise the value of" - " innodb_max_open_files in\n" + " innodb_open_files in\n" "InnoDB: my.cnf! Remember that" " InnoDB keeps all log files" " and all system\n" @@ -1583,6 +1609,8 @@ fil_close_all_files(void) space = UT_LIST_GET_FIRST(fil_system->space_list); while (space != NULL) { + fil_space_t* prev_space = space; + node = UT_LIST_GET_FIRST(space->chain); while (node != NULL) { @@ -1592,6 +1620,7 @@ fil_close_all_files(void) node = UT_LIST_GET_NEXT(chain, node); } space = UT_LIST_GET_NEXT(space_list, space); + fil_space_free(prev_space->id, TRUE); } mutex_exit(&fil_system->mutex); @@ -2223,7 +2252,7 @@ try_again: #endif /* printf("Deleting tablespace %s id %lu\n", space->name, id); */ - success = fil_space_free(id); + success = fil_space_free(id, FALSE); if (success) { success = os_file_delete(path); @@ -2929,7 +2958,6 @@ fil_open_single_table_tablespace( byte* page; ulint space_id; ulint space_flags; - ibool ret = TRUE; filepath = fil_make_ibd_name(name, FALSE); @@ -3330,7 +3358,7 @@ skip_write: (ulong) space_id, (ulong) space_flags, (ulong) id, (ulong) flags); - ret = FALSE; + success = FALSE; goto func_exit; } @@ -3350,7 +3378,7 @@ func_exit: os_file_close(file); mem_free(filepath); - return(ret); + return(success); } #endif /* !UNIV_HOTBACKUP */ @@ -3566,7 +3594,7 @@ fil_load_single_table_tablespace( fprintf(stderr, "InnoDB: Renaming tablespace %s of id %lu,\n" "InnoDB: to %s_ibbackup_old_vers_<timestamp>\n" - "InnoDB: because its size %lld is too small" + "InnoDB: because its size %" PRId64 " is too small" " (< 4 pages 16 kB each),\n" "InnoDB: or the space id in the file header" " is not sensible.\n" @@ -3628,7 +3656,17 @@ fil_load_single_table_tablespace( if (!success) { - goto func_exit; + if (srv_force_recovery > 0) { + fprintf(stderr, + "InnoDB: innodb_force_recovery" + " was set to %lu. Continuing crash recovery\n" + "InnoDB: even though the tablespace creation" + " of this table failed.\n", + srv_force_recovery); + goto func_exit; + } + + exit(1); } /* We do not use the size information we have about the file, because @@ -4163,7 +4201,7 @@ fil_extend_space_to_desired_size( node->name, node->handle, buf, offset_low, offset_high, page_size * n_pages, - NULL, NULL); + NULL, NULL, NULL); #endif if (success) { node->size += n_pages; @@ -4490,7 +4528,7 @@ Reads or writes data. This operation is asynchronous (aio). i/o on a tablespace which does not exist */ UNIV_INTERN ulint -fil_io( +_fil_io( /*===*/ ulint type, /*!< in: OS_FILE_READ or OS_FILE_WRITE, ORed to OS_FILE_LOG, if a log i/o @@ -4515,8 +4553,9 @@ fil_io( void* buf, /*!< in/out: buffer where to store read data or from where to write; in aio this must be appropriately aligned */ - void* message) /*!< in: message for aio handler if non-sync + void* message, /*!< in: message for aio handler if non-sync aio used, else ignored */ + trx_t* trx) { ulint mode; fil_space_t* space; @@ -4686,7 +4725,7 @@ fil_io( #else /* Queue the aio request */ ret = os_aio(type, mode | wake_later, node->name, node->handle, buf, - offset_low, offset_high, len, node, message); + offset_low, offset_high, len, node, message, trx); #endif ut_a(ret); @@ -4706,6 +4745,78 @@ fil_io( return(DB_SUCCESS); } +/********************************************************************//** +Confirm whether the parameters are valid or not */ +UNIV_INTERN +ibool +fil_area_is_exist( +/*==============*/ + ulint space_id, /*!< in: space id */ + ulint zip_size, /*!< in: compressed page size in bytes; + 0 for uncompressed pages */ + ulint block_offset, /*!< in: offset in number of blocks */ + ulint byte_offset, /*!< in: remainder of offset in bytes; in + aio this must be divisible by the OS block + size */ + ulint len) /*!< in: how many bytes to read or write; this + must not cross a file boundary; in aio this + must be a block size multiple */ +{ + fil_space_t* space; + fil_node_t* node; + + /* Reserve the fil_system mutex and make sure that we can open at + least one file while holding it, if the file is not already open */ + + fil_mutex_enter_and_prepare_for_io(space_id); + + space = fil_space_get_by_id(space_id); + + if (!space) { + mutex_exit(&fil_system->mutex); + return(FALSE); + } + + node = UT_LIST_GET_FIRST(space->chain); + + for (;;) { + if (UNIV_UNLIKELY(node == NULL)) { + mutex_exit(&fil_system->mutex); + return(FALSE); + } + + if (space->id != 0 && node->size == 0) { + /* We do not know the size of a single-table tablespace + before we open the file */ + + break; + } + + if (node->size > block_offset) { + /* Found! */ + break; + } else { + block_offset -= node->size; + node = UT_LIST_GET_NEXT(chain, node); + } + } + + /* Open file if closed */ + fil_node_prepare_for_io(node, fil_system, space); + fil_node_complete_io(node, fil_system, OS_FILE_READ); + + /* Check that at least the start offset is within the bounds of a + single-table tablespace */ + if (UNIV_UNLIKELY(node->size <= block_offset) + && space->id != 0 && space->purpose == FIL_TABLESPACE) { + mutex_exit(&fil_system->mutex); + return(FALSE); + } + + mutex_exit(&fil_system->mutex); + return(TRUE); +} + #ifndef UNIV_HOTBACKUP /**********************************************************************//** Waits for an aio operation to complete. This function is used to write the @@ -5065,6 +5176,29 @@ fil_page_get_type( return(mach_read_from_2(page + FIL_PAGE_TYPE)); } +/******************************************************************** +Initializes the tablespace memory cache. */ +UNIV_INTERN +void +fil_close(void) +/*===========*/ +{ + /* The mutex should already have been freed. */ + ut_ad(fil_system->mutex.magic_n == 0); + + hash_table_free(fil_system->spaces); + + hash_table_free(fil_system->name_hash); + + ut_a(UT_LIST_GET_LEN(fil_system->LRU) == 0); + ut_a(UT_LIST_GET_LEN(fil_system->unflushed_spaces) == 0); + ut_a(UT_LIST_GET_LEN(fil_system->space_list) == 0); + + mem_free(fil_system); + + fil_system = NULL; +} + /************************************************************************* Return local hash table informations. */ diff --git a/storage/xtradb/fsp/fsp0fsp.c b/storage/xtradb/fsp/fsp0fsp.c index ce14723ba18..3cc4318fc06 100644 --- a/storage/xtradb/fsp/fsp0fsp.c +++ b/storage/xtradb/fsp/fsp0fsp.c @@ -232,6 +232,9 @@ the extent are free and which contain old tuple version to clean. */ #define XDES_ARR_OFFSET (FSP_HEADER_OFFSET + FSP_HEADER_SIZE) #ifndef UNIV_HOTBACKUP +/* Flag to indicate if we have printed the tablespace full error. */ +static ibool fsp_tbs_full_error_printed = FALSE; + /**********************************************************************//** Returns an extent to the free list of a space. */ static @@ -1099,7 +1102,7 @@ fsp_header_inc_size( /**********************************************************************//** Gets the current free limit of the system tablespace. The free limit -means the place of the first page which has never been put to the the +means the place of the first page which has never been put to the free list for allocation. The space above that address is initialized to zero. Sets also the global variable log_fsp_current_free_limit. @return free limit in megabytes */ @@ -1218,6 +1221,19 @@ fsp_try_extend_data_file( if (space == 0 && !srv_auto_extend_last_data_file) { + /* We print the error message only once to avoid + spamming the error log. Note that we don't need + to reset the flag to FALSE as dealing with this + error requires server restart. */ + if (fsp_tbs_full_error_printed == FALSE) { + fprintf(stderr, + "InnoDB: Error: Data file(s) ran" + " out of space.\n" + "Please add another data file or" + " use \'autoextend\' for the last" + " data file.\n"); + fsp_tbs_full_error_printed = TRUE; + } return(FALSE); } @@ -1832,6 +1848,8 @@ fsp_seg_inode_page_find_used( if (!ut_dulint_is_zero(mach_read_from_8(inode + FSEG_ID))) { /* This is used */ + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); return(i); } } @@ -1863,6 +1881,9 @@ fsp_seg_inode_page_find_free( return(i); } + + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); } return(ULINT_UNDEFINED); @@ -1981,6 +2002,8 @@ fsp_alloc_seg_inode( page + FSEG_INODE_PAGE_NODE, mtr); } + ut_ad(ut_dulint_is_zero(mach_read_from_8(inode + FSEG_ID)) + || mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); return(inode); } @@ -2018,7 +2041,7 @@ fsp_free_seg_inode( } mlog_write_dulint(inode + FSEG_ID, ut_dulint_zero, mtr); - mlog_write_ulint(inode + FSEG_MAGIC_N, 0, MLOG_4BYTES, mtr); + mlog_write_ulint(inode + FSEG_MAGIC_N, 0xfa051ce3, MLOG_4BYTES, mtr); if (ULINT_UNDEFINED == fsp_seg_inode_page_find_used(page, zip_size, mtr)) { @@ -2034,11 +2057,11 @@ fsp_free_seg_inode( /**********************************************************************//** Returns the file segment inode, page x-latched. -@return segment inode, page x-latched */ +@return segment inode, page x-latched; NULL if the inode is free */ static fseg_inode_t* -fseg_inode_get( -/*===========*/ +fseg_inode_try_get( +/*===============*/ fseg_header_t* header, /*!< in: segment header */ ulint space, /*!< in: space id */ ulint zip_size,/*!< in: compressed page size in bytes @@ -2054,12 +2077,38 @@ fseg_inode_get( inode = fut_get_ptr(space, zip_size, inode_addr, RW_X_LATCH, mtr); - ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); + if (UNIV_UNLIKELY + (ut_dulint_is_zero(mach_read_from_8(inode + FSEG_ID)))) { + + inode = NULL; + } else { + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); + } return(inode); } /**********************************************************************//** +Returns the file segment inode, page x-latched. +@return segment inode, page x-latched */ +static +fseg_inode_t* +fseg_inode_get( +/*===========*/ + fseg_header_t* header, /*!< in: segment header */ + ulint space, /*!< in: space id */ + ulint zip_size,/*!< in: compressed page size in bytes + or 0 for uncompressed pages */ + mtr_t* mtr) /*!< in: mtr handle */ +{ + fseg_inode_t* inode + = fseg_inode_try_get(header, space, zip_size, mtr); + ut_a(inode); + return(inode); +} + +/**********************************************************************//** Gets the page number from the nth fragment page slot. @return page number, FIL_NULL if not in use */ UNIV_INLINE @@ -2073,6 +2122,7 @@ fseg_get_nth_frag_page_no( ut_ad(inode && mtr); ut_ad(n < FSEG_FRAG_ARR_N_SLOTS); ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX)); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); return(mach_read_from_4(inode + FSEG_FRAG_ARR + n * FSEG_FRAG_SLOT_SIZE)); } @@ -2091,6 +2141,7 @@ fseg_set_nth_frag_page_no( ut_ad(inode && mtr); ut_ad(n < FSEG_FRAG_ARR_N_SLOTS); ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX)); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); mlog_write_ulint(inode + FSEG_FRAG_ARR + n * FSEG_FRAG_SLOT_SIZE, page_no, MLOG_4BYTES, mtr); @@ -2451,6 +2502,8 @@ fseg_fill_free_list( xdes_set_state(descr, XDES_FSEG, mtr); seg_id = mtr_read_dulint(inode + FSEG_ID, mtr); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); mlog_write_dulint(descr + XDES_ID, seg_id, mtr); flst_add_last(inode + FSEG_FREE, descr + XDES_FLST_NODE, mtr); @@ -2479,6 +2532,7 @@ fseg_alloc_free_extent( fil_addr_t first; ut_ad(!((page_offset(inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE)); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); if (flst_get_len(inode + FSEG_FREE, mtr) > 0) { /* Segment free list is not empty, allocate from it */ @@ -3136,6 +3190,8 @@ fseg_mark_page_used( ut_ad(seg_inode && mtr); ut_ad(!((page_offset(seg_inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE)); + ut_ad(mach_read_from_4(seg_inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); descr = xdes_get_descriptor(space, zip_size, page, mtr); @@ -3373,6 +3429,8 @@ fseg_free_extent( ut_a(xdes_get_state(descr, mtr) == XDES_FSEG); ut_a(0 == ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID, mtr), mtr_read_dulint(seg_inode + FSEG_ID, mtr))); + ut_ad(mach_read_from_4(seg_inode + FSEG_MAGIC_N) + == FSEG_MAGIC_N_VALUE); first_page_in_extent = page - (page % FSP_EXTENT_SIZE); @@ -3463,7 +3521,13 @@ fseg_free_step( ut_a(descr); ut_a(xdes_get_bit(descr, XDES_FREE_BIT, header_page % FSP_EXTENT_SIZE, mtr) == FALSE); - inode = fseg_inode_get(header, space, zip_size, mtr); + inode = fseg_inode_try_get(header, space, zip_size, mtr); + + if (UNIV_UNLIKELY(inode == NULL)) { + fprintf(stderr, "double free of inode from %u:%u\n", + (unsigned) space, (unsigned) header_page); + return(TRUE); + } descr = fseg_get_first_extent(inode, space, zip_size, mtr); @@ -3587,6 +3651,7 @@ fseg_get_first_extent( ut_ad(inode && mtr); ut_ad(space == page_get_space_id(page_align(inode))); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); first = fil_addr_null; @@ -3801,6 +3866,7 @@ fseg_print_low( (ulong) reserved, (ulong) used, (ulong) n_full, (ulong) n_frag, (ulong) n_free, (ulong) n_not_full, (ulong) n_used); + ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE); } #ifdef UNIV_BTR_PRINT diff --git a/storage/xtradb/handler/ha_innodb.cc b/storage/xtradb/handler/ha_innodb.cc index f723a58c6d9..b4c2f7abc16 100644 --- a/storage/xtradb/handler/ha_innodb.cc +++ b/storage/xtradb/handler/ha_innodb.cc @@ -79,6 +79,7 @@ with this program; if not, write to the Free Software Foundation, Inc., /* Include necessary InnoDB headers */ extern "C" { #include "univ.i" +#include "buf0lru.h" #include "btr0sea.h" #include "os0file.h" #include "os0thread.h" @@ -111,7 +112,6 @@ extern "C" { #include "ha_innodb.h" #include "i_s.h" -#include "handler0vars.h" #ifdef MYSQL_SERVER // Defined in trx0sys.c @@ -122,9 +122,12 @@ extern ib_int64_t trx_sys_mysql_relay_log_pos; #endif /* MYSQL_SERVER */ #ifndef MYSQL_SERVER -/* This is needed because of Bug #3596. Let us hope that pthread_mutex_t +# ifndef MYSQL_PLUGIN_IMPORT +# define MYSQL_PLUGIN_IMPORT /* nothing */ +# endif /* MYSQL_PLUGIN_IMPORT */ +/* This is needed because of Bug #3596. Let us hope that pthread_mutex_t is defined the same in both builds: the MySQL server and the InnoDB plugin. */ -extern pthread_mutex_t LOCK_thread_count; +extern MYSQL_PLUGIN_IMPORT pthread_mutex_t LOCK_thread_count; #if MYSQL_VERSION_ID < 50124 /* this is defined in mysql_priv.h inside #ifdef MYSQL_SERVER @@ -148,13 +151,9 @@ static bool innodb_inited = 0; /* In the Windows plugin, the return value of current_thd is undefined. Map it to NULL. */ -#if defined MYSQL_DYNAMIC_PLUGIN && defined __WIN__ -# undef current_thd -# define current_thd NULL -# define EQ_CURRENT_THD(thd) TRUE -#else /* MYSQL_DYNAMIC_PLUGIN && __WIN__ */ -# define EQ_CURRENT_THD(thd) ((thd) == current_thd) -#endif /* MYSQL_DYNAMIC_PLUGIN && __WIN__ */ + +#define EQ_CURRENT_THD(thd) ((thd) == current_thd) + static struct handlerton* innodb_hton_ptr; @@ -174,6 +173,10 @@ static ulong innobase_write_io_threads; static my_bool innobase_thread_concurrency_timer_based; static long long innobase_buffer_pool_size, innobase_log_file_size; +/** Percentage of the buffer pool to reserve for 'old' blocks. +Connected to buf_LRU_old_ratio. */ +static uint innobase_old_blocks_pct; + /* The default values for the following char* start-up parameters are determined in innobase_init below: */ @@ -188,9 +191,7 @@ file formats in the configuration file, but can only be set to any of the supported file formats during runtime. */ static char* innobase_file_format_check = NULL; -/* The following has a misleading name: starting from 4.0.5, this also -affects Windows: */ -static char* innobase_unix_file_flush_method = NULL; +static char* innobase_file_flush_method = NULL; /* Below we have boolean-valued start-up parameters, and their default values */ @@ -204,7 +205,7 @@ static my_bool innobase_use_doublewrite = TRUE; static my_bool innobase_use_checksums = TRUE; static my_bool innobase_extra_undoslots = FALSE; static my_bool innobase_fast_recovery = FALSE; -static my_bool innobase_use_purge_thread = FALSE; +static my_bool innobase_recovery_stats = TRUE; static my_bool innobase_locks_unsafe_for_binlog = FALSE; static my_bool innobase_overwrite_relay_log_info = FALSE; static my_bool innobase_rollback_on_timeout = FALSE; @@ -240,10 +241,10 @@ static void free_share(INNOBASE_SHARE *share); static int innobase_close_connection(handlerton *hton, THD* thd); static int innobase_commit(handlerton *hton, THD* thd, bool all); static int innobase_rollback(handlerton *hton, THD* thd, bool all); -static int innobase_rollback_to_savepoint(handlerton *hton, THD* thd, +static int innobase_rollback_to_savepoint(handlerton *hton, THD* thd, void *savepoint); static int innobase_savepoint(handlerton *hton, THD* thd, void *savepoint); -static int innobase_release_savepoint(handlerton *hton, THD* thd, +static int innobase_release_savepoint(handlerton *hton, THD* thd, void *savepoint); static handler *innobase_create_handler(handlerton *hton, TABLE_SHARE *table, @@ -287,10 +288,10 @@ innobase_file_format_check_on_off( /************************************************************//** Validate the file format check config parameters, as a side effect it sets the srv_check_file_format_at_startup variable. -@return true if valid config value */ +@return the format_id if valid config value, otherwise, return -1 */ static -bool -innobase_file_format_check_validate( +int +innobase_file_format_validate_and_set( /*================================*/ const char* format_check); /*!< in: parameter value */ /****************************************************************//** @@ -521,10 +522,10 @@ static SHOW_VAR innodb_status_variables[]= { (char*) &export_vars.innodb_buffer_pool_pages_misc, SHOW_LONG}, {"buffer_pool_pages_total", (char*) &export_vars.innodb_buffer_pool_pages_total, SHOW_LONG}, - {"buffer_pool_read_ahead_rnd", - (char*) &export_vars.innodb_buffer_pool_read_ahead_rnd, SHOW_LONG}, - {"buffer_pool_read_ahead_seq", - (char*) &export_vars.innodb_buffer_pool_read_ahead_seq, SHOW_LONG}, + {"buffer_pool_read_ahead", + (char*) &export_vars.innodb_buffer_pool_read_ahead, SHOW_LONG}, + {"buffer_pool_read_ahead_evicted", + (char*) &export_vars.innodb_buffer_pool_read_ahead_evicted, SHOW_LONG}, {"buffer_pool_read_requests", (char*) &export_vars.innodb_buffer_pool_read_requests, SHOW_LONG}, {"buffer_pool_reads", @@ -805,11 +806,20 @@ convert_error_code_to_mysql( case DB_SUCCESS: return(0); + case DB_INTERRUPTED: + my_error(ER_QUERY_INTERRUPTED, MYF(0)); + /* fall through */ case DB_ERROR: default: return(-1); /* unspecified error */ case DB_DUPLICATE_KEY: + /* Be cautious with returning this error, since + mysql could re-enter the storage layer to get + duplicated key info, the operation requires a + valid table handle and/or transaction information, + which might not always be available in the error + handling stage. */ return(HA_ERR_FOUND_DUPP_KEY); case DB_FOREIGN_DUPLICATE_KEY: @@ -896,17 +906,14 @@ convert_error_code_to_mysql( return(ER_PRIMARY_CANT_HAVE_NULL); case DB_TOO_MANY_CONCURRENT_TRXS: - /* Once MySQL add the appropriate code to errmsg.txt then - we can get rid of this #ifdef. NOTE: The code checked by - the #ifdef is the suggested name for the error condition - and the actual error code name could very well be different. - This will require some monitoring, ie. the status - of this request on our part.*/ -#ifdef ER_TOO_MANY_CONCURRENT_TRXS - return(ER_TOO_MANY_CONCURRENT_TRXS); -#else + /* New error code HA_ERR_TOO_MANY_CONCURRENT_TRXS is only + available in 5.1.38 and later, but the plugin should still + work with previous versions of MySQL. */ +#ifdef HA_ERR_TOO_MANY_CONCURRENT_TRXS + return(HA_ERR_TOO_MANY_CONCURRENT_TRXS); +#else /* HA_ERR_TOO_MANY_CONCURRENT_TRXS */ return(HA_ERR_RECORD_FILE_FULL); -#endif +#endif /* HA_ERR_TOO_MANY_CONCURRENT_TRXS */ case DB_UNSUPPORTED: return(HA_ERR_UNSUPPORTED); } @@ -980,7 +987,23 @@ innobase_get_cset_width( *mbminlen = cs->mbminlen; *mbmaxlen = cs->mbmaxlen; } else { - ut_a(cset == 0); + THD* thd = current_thd; + + if (thd && thd_sql_command(thd) == SQLCOM_DROP_TABLE) { + + /* Fix bug#46256: allow tables to be dropped if the + collation is not found, but issue a warning. */ + if ((global_system_variables.log_warnings) + && (cset != 0)){ + + sql_print_warning( + "Unknown collation #%lu.", cset); + } + } else { + + ut_a(cset == 0); + } + *mbminlen = *mbmaxlen = 0; } } @@ -1054,6 +1077,7 @@ innobase_get_charset( } #if defined (__WIN__) && defined (MYSQL_DYNAMIC_PLUGIN) +extern MYSQL_PLUGIN_IMPORT MY_TMPDIR mysql_tmpdir_list; /*******************************************************************//** Map an OS error to an errno value. The OS error number is stored in _doserrno and the mapped value is stored in errno) */ @@ -1341,6 +1365,16 @@ innobase_trx_init( trx->check_unique_secondary = !thd_test_options( thd, OPTION_RELAXED_UNIQUE_CHECKS); +#ifdef EXTENDED_SLOWLOG + if (thd_log_slow_verbosity(thd) & SLOG_V_INNODB) { + trx->take_stats = TRUE; + } else { + trx->take_stats = FALSE; + } +#else + trx->take_stats = FALSE; +#endif + DBUG_VOID_RETURN; } @@ -1397,6 +1431,32 @@ check_trx_exists( } +/************************************************************************* +Gets current trx. */ +extern "C" +trx_t* +innobase_get_trx() +{ + THD *thd=current_thd; + if (likely(thd != 0)) { + trx_t*& trx = thd_to_trx(thd); + return(trx); + } else { + return(NULL); + } +} + +extern "C" +ibool +innobase_get_slow_log() +{ +#ifdef EXTENDED_SLOWLOG + return((ibool) thd_opt_slow_log()); +#else + return(FALSE); +#endif +} + /*********************************************************************//** Construct ha_innobase handler. */ UNIV_INTERN @@ -1713,15 +1773,19 @@ innobase_convert_identifier( FALSE=id is an UTF-8 string */ { char nz[NAME_LEN + 1]; +#if MYSQL_VERSION_ID >= 50141 + char nz2[NAME_LEN + 1 + EXPLAIN_FILENAME_MAX_EXTRA_LENGTH]; +#else /* MYSQL_VERSION_ID >= 50141 */ char nz2[NAME_LEN + 1 + sizeof srv_mysql50_table_name_prefix]; +#endif /* MYSQL_VERSION_ID >= 50141 */ const char* s = id; int q; if (file_id) { - /* Decode the table name. The filename_to_tablename() - function expects a NUL-terminated string. The input and - output strings buffers must not be shared. */ + /* Decode the table name. The MySQL function expects + a NUL-terminated string. The input and output strings + buffers must not be shared. */ if (UNIV_UNLIKELY(idlen > (sizeof nz) - 1)) { idlen = (sizeof nz) - 1; @@ -1731,7 +1795,13 @@ innobase_convert_identifier( nz[idlen] = 0; s = nz2; +#if MYSQL_VERSION_ID >= 50141 + idlen = explain_filename((THD*) thd, nz, nz2, sizeof nz2, + EXPLAIN_PARTITIONS_AS_COMMENT); + goto no_quote; +#else /* MYSQL_VERSION_ID >= 50141 */ idlen = filename_to_tablename(nz, nz2, sizeof nz2); +#endif /* MYSQL_VERSION_ID >= 50141 */ } /* See if the identifier needs to be quoted. */ @@ -1742,6 +1812,9 @@ innobase_convert_identifier( } if (q == EOF) { +#if MYSQL_VERSION_ID >= 50141 +no_quote: +#endif /* MYSQL_VERSION_ID >= 50141 */ if (UNIV_UNLIKELY(idlen > buflen)) { idlen = buflen; } @@ -2224,8 +2297,8 @@ mem_free_and_error: /* Did the user specify a format name that we support ? As a side effect it will update the variable srv_check_file_format_at_startup */ - if (!innobase_file_format_check_validate( - innobase_file_format_check)) { + if (innobase_file_format_validate_and_set( + innobase_file_format_check) < 0) { sql_print_error("InnoDB: invalid " "innodb_file_format_check value: " @@ -2266,7 +2339,7 @@ innobase_change_buffering_inited_ok: /* --------------------------------------------------*/ - srv_file_flush_method_str = innobase_unix_file_flush_method; + srv_file_flush_method_str = innobase_file_flush_method; srv_n_log_groups = (ulint) innobase_mirrored_log_groups; srv_n_log_files = (ulint) innobase_log_files_in_group; @@ -2294,8 +2367,7 @@ innobase_change_buffering_inited_ok: srv_force_recovery = (ulint) innobase_force_recovery; srv_fast_recovery = (ibool) innobase_fast_recovery; - - srv_use_purge_thread = (ibool) innobase_use_purge_thread; + srv_recovery_stats = (ibool) innobase_recovery_stats; srv_use_doublewrite_buf = (ibool) innobase_use_doublewrite; srv_use_checksums = (ibool) innobase_use_checksums; @@ -2331,6 +2403,9 @@ innobase_change_buffering_inited_ok: ut_a(0 == strcmp(my_charset_latin1.name, "latin1_swedish_ci")); srv_latin1_ordering = my_charset_latin1.sort_order; + innobase_old_blocks_pct = buf_LRU_old_ratio_update( + innobase_old_blocks_pct, FALSE); + innobase_commit_concurrency_init_default(); /* Since we in this module access directly the fields of a trx @@ -2512,7 +2587,10 @@ innobase_alter_table_flags( { return(HA_ONLINE_ADD_INDEX_NO_WRITES | HA_ONLINE_DROP_INDEX_NO_WRITES - | HA_ONLINE_ADD_UNIQUE_INDEX_NO_WRITES + /* Current InnoDB doesn't sort unique indexes along mysqld's order + It is dangerous to use index. So it is disabled until + the bug http://bugs.mysql.com/47622 */ + /* | HA_ONLINE_ADD_UNIQUE_INDEX_NO_WRITES */ | HA_ONLINE_DROP_UNIQUE_INDEX_NO_WRITES | HA_ONLINE_ADD_PK_INDEX_NO_WRITES); } @@ -2678,6 +2756,19 @@ retry: } } + /* The following calls to read the MySQL binary log + file name and the position return consistent results: + 1) Other InnoDB transactions cannot intervene between + these calls as we are holding prepare_commit_mutex. + 2) Binary logging of other engines is not relevant + to InnoDB as all InnoDB requires is that committing + InnoDB transactions appear in the same order in the + MySQL binary log as they appear in InnoDB logs. + 3) A MySQL log file rotation cannot happen because + MySQL protects against this by having a counter of + transactions in prepared state and it only allows + a rotation when the counter drops to zero. See + LOCK_prep_xids and COND_prep_xids in log.cc. */ trx->mysql_log_file_name = mysql_bin_log_file_name(); trx->mysql_log_offset = (ib_int64_t) mysql_bin_log_file_pos(); @@ -2763,6 +2854,8 @@ innobase_rollback( innobase_release_stat_resources(trx); + trx->n_autoinc_rows = 0; /* Reset the number AUTO-INC rows required */ + /* If we had reserved the auto-inc lock for some table (if we come here to roll back the latest SQL statement) we release it now before a possibly lengthy rollback */ @@ -3324,7 +3417,7 @@ retry: if (is_part) { sql_print_error("Failed to open table %s after " - "%lu attemtps.\n", norm_name, + "%lu attempts.\n", norm_name, retries); } @@ -3928,7 +4021,6 @@ ha_innobase::store_key_val_for_row( as BLOB data in innodb. */ || mysql_type == MYSQL_TYPE_GEOMETRY) { - CHARSET_INFO* cs; ulint key_len; ulint true_len; @@ -5229,6 +5321,11 @@ ha_innobase::index_read( index = prebuilt->index; + if (UNIV_UNLIKELY(index == NULL)) { + prebuilt->index_usable = FALSE; + DBUG_RETURN(HA_ERR_CRASHED); + } + /* Note that if the index for which the search template is built is not necessarily prebuilt->index, but can also be the clustered index */ @@ -5388,6 +5485,7 @@ ha_innobase::change_active_index( if (UNIV_UNLIKELY(!prebuilt->index)) { sql_print_warning("InnoDB: change_active_index(%u) failed", keynr); + prebuilt->index_usable = FALSE; DBUG_RETURN(1); } @@ -5395,8 +5493,10 @@ ha_innobase::change_active_index( prebuilt->index); if (UNIV_UNLIKELY(!prebuilt->index_usable)) { - sql_print_warning("InnoDB: insufficient history for index %u", - keynr); + push_warning_printf(user_thd, MYSQL_ERROR::WARN_LEVEL_WARN, + HA_ERR_TABLE_DEF_CHANGED, + "InnoDB: insufficient history for index %u", + keynr); /* The caller seems to ignore this. Thus, we must check this again in row_search_for_mysql(). */ DBUG_RETURN(convert_error_code_to_mysql(DB_MISSING_HISTORY, @@ -5852,7 +5952,7 @@ create_table_def( number fits in one byte in prtype */ push_warning_printf( (THD*) trx->mysql_thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_CANT_CREATE_TABLE, "In InnoDB, charset-collation codes" " must be below 256." @@ -5884,17 +5984,8 @@ create_table_def( /* First check whether the column to be added has a system reserved name. */ if (dict_col_name_is_reserved(field->field_name)){ - push_warning_printf( - (THD*) trx->mysql_thd, - MYSQL_ERROR::WARN_LEVEL_WARN, - ER_CANT_CREATE_TABLE, - "Error creating table '%s' with " - "column name '%s'. '%s' is a " - "reserved name. Please try to " - "re-create the table with a " - "different column name.", - table->name, (char*) field->field_name, - (char*) field->field_name); + my_error(ER_WRONG_COLUMN_NAME, MYF(0), + field->field_name); dict_mem_table_free(table); trx_commit_for_mysql(trx); @@ -5916,6 +6007,14 @@ create_table_def( error = row_create_table_for_mysql(table, trx); + if (error == DB_DUPLICATE_KEY) { + char buf[100]; + innobase_convert_identifier(buf, sizeof buf, + table_name, strlen(table_name), + trx->mysql_thd, TRUE); + my_error(ER_TABLE_EXISTS_ERROR, MYF(0), buf); + } + error_ret: error = convert_error_code_to_mysql(error, flags, NULL); @@ -6066,7 +6165,6 @@ create_clustered_index_when_no_primary( /* We pass 0 as the space id, and determine at a lower level the space id where to store the table */ - index = dict_mem_index_create(table_name, innobase_index_reserve_name, 0, DICT_CLUSTERED, 0); @@ -6121,7 +6219,7 @@ create_options_are_valid( /* Valid value. */ break; default: - push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_ERROR, + push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: invalid" " KEY_BLOCK_SIZE = %lu." @@ -6135,7 +6233,7 @@ create_options_are_valid( /* If KEY_BLOCK_SIZE was specified, check for its dependencies. */ if (kbs_specified && !srv_file_per_table) { - push_warning(thd, MYSQL_ERROR::WARN_LEVEL_ERROR, + push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: KEY_BLOCK_SIZE" " requires innodb_file_per_table."); @@ -6143,7 +6241,7 @@ create_options_are_valid( } if (kbs_specified && srv_file_format < DICT_TF_FORMAT_ZIP) { - push_warning(thd, MYSQL_ERROR::WARN_LEVEL_ERROR, + push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: KEY_BLOCK_SIZE" " requires innodb_file_format >" @@ -6167,7 +6265,7 @@ create_options_are_valid( if (!srv_file_per_table) { push_warning_printf( thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: ROW_FORMAT=%s" " requires innodb_file_per_table.", @@ -6179,7 +6277,7 @@ create_options_are_valid( if (srv_file_format < DICT_TF_FORMAT_ZIP) { push_warning_printf( thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: ROW_FORMAT=%s" " requires innodb_file_format >" @@ -6196,7 +6294,7 @@ create_options_are_valid( && form->s->row_type == ROW_TYPE_DYNAMIC) { push_warning_printf( thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: cannot specify" " ROW_FORMAT = DYNAMIC with" @@ -6220,7 +6318,7 @@ create_options_are_valid( if (kbs_specified) { push_warning_printf( thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: cannot specify" " ROW_FORMAT = %s with" @@ -6233,7 +6331,7 @@ create_options_are_valid( default: push_warning(thd, - MYSQL_ERROR::WARN_LEVEL_ERROR, + MYSQL_ERROR::WARN_LEVEL_WARN, ER_ILLEGAL_HA_CREATE_OPTION, "InnoDB: invalid ROW_FORMAT specifier."); ret = FALSE; @@ -6297,13 +6395,15 @@ ha_innobase::create( 1. <database_name>/<table_name>: for normal table creation 2. full path: for temp table creation, or sym link - When srv_file_per_table is on, check for full path pattern, i.e. + When srv_file_per_table is on and mysqld_embedded is off, + check for full path pattern, i.e. X:\dir\..., X is a driver letter, or \\dir1\dir2\..., UNC path returns error if it is in full path format, but not creating a temp. table. Currently InnoDB does not support symbolic link on Windows. */ if (srv_file_per_table + && !mysqld_embedded && (!create_info->options & HA_LEX_CREATE_TMP_TABLE)) { if ((name[1] == ':') @@ -6521,6 +6621,7 @@ ha_innobase::create( goto cleanup; } + /* Create the keys */ if (form->s->keys == 0 || primary_key_no == -1) { @@ -6969,6 +7070,24 @@ ha_innobase::rename_table( innobase_commit_low(trx); trx_free_for_mysql(trx); + /* Add a special case to handle the Duplicated Key error + and return DB_ERROR instead. + This is to avoid a possible SIGSEGV error from mysql error + handling code. Currently, mysql handles the Duplicated Key + error by re-entering the storage layer and getting dup key + info by calling get_dup_key(). This operation requires a valid + table handle ('row_prebuilt_t' structure) which could no + longer be available in the error handling stage. The suggested + solution is to report a 'table exists' error message (since + the dup key error here is due to an existing table whose name + is the one we are trying to rename to) and return the generic + error code. */ + if (error == (int) DB_DUPLICATE_KEY) { + my_error(ER_TABLE_EXISTS_ERROR, MYF(0), to); + + error = DB_ERROR; + } + error = convert_error_code_to_mysql(error, 0, NULL); DBUG_RETURN(error); @@ -7521,11 +7640,15 @@ ha_innobase::check( ret = row_check_table_for_mysql(prebuilt); - if (ret == DB_SUCCESS) { + switch (ret) { + case DB_SUCCESS: return(HA_ADMIN_OK); + case DB_INTERRUPTED: + my_error(ER_QUERY_INTERRUPTED, MYF(0)); + return(-1); + default: + return(HA_ADMIN_CORRUPT); } - - return(HA_ADMIN_CORRUPT); } /*************************************************************//** @@ -8071,8 +8194,11 @@ ha_innobase::external_lock( ulong const binlog_format= thd_binlog_format(thd); ulong const tx_isolation = thd_tx_isolation(ha_thd()); if (tx_isolation <= ISO_READ_COMMITTED - && binlog_format == BINLOG_FORMAT_STMT - && thd_binlog_filter_ok(thd)) + && binlog_format == BINLOG_FORMAT_STMT +#if MYSQL_VERSION_ID > 50140 + && thd_binlog_filter_ok(thd) +#endif /* MYSQL_VERSION_ID > 50140 */ + ) { char buf[256]; my_snprintf(buf, sizeof(buf), @@ -8185,6 +8311,23 @@ ha_innobase::external_lock( statement has ended */ if (trx->n_mysql_tables_in_use == 0) { +#ifdef EXTENDED_SLOWLOG + increment_thd_innodb_stats(thd, trx->io_reads, + trx->io_read, + trx->io_reads_wait_timer, + trx->lock_que_wait_timer, + trx->innodb_que_wait_timer, + trx->distinct_page_access); + + trx->io_reads = 0; + trx->io_read = 0; + trx->io_reads_wait_timer = 0; + trx->lock_que_wait_timer = 0; + trx->innodb_que_wait_timer = 0; + trx->distinct_page_access = 0; + if (trx->distinct_page_access_hash) + memset(trx->distinct_page_access_hash, 0, DPAH_SIZE); +#endif trx->mysql_n_tables_locked = 0; prebuilt->used_in_HANDLER = FALSE; @@ -8472,8 +8615,8 @@ innodb_mutex_show_status( rw_lock_wait_time += mutex->lspent_time; } #else /* UNIV_DEBUG */ - buf1len= (uint) my_snprintf(buf1, sizeof(buf1), "%s:%lu", - mutex->cfile_name, (ulong) mutex->cline); + buf1len= (uint) my_snprintf(buf1, sizeof(buf1), "%s", + mutex->cmutex_name); buf2len= (uint) my_snprintf(buf2, sizeof(buf2), "os_waits=%lu", mutex->count_os_wait); @@ -8498,8 +8641,8 @@ next_mutex: while (lock != NULL) { if (lock->count_os_wait && !buf_pool_is_block_lock(lock)) { - buf1len= my_snprintf(buf1, sizeof(buf1), "%s:%lu", - lock->cfile_name, (ulong) lock->cline); + buf1len= my_snprintf(buf1, sizeof(buf1), "%s", + lock->lock_name); buf2len= my_snprintf(buf2, sizeof(buf2), "os_waits=%lu", lock->count_os_wait); @@ -8720,6 +8863,7 @@ ha_innobase::store_lock( && isolation_level != TRX_ISO_SERIALIZABLE && (lock_type == TL_READ || lock_type == TL_READ_NO_INSERT) && (sql_command == SQLCOM_INSERT_SELECT + || sql_command == SQLCOM_REPLACE_SELECT || sql_command == SQLCOM_UPDATE || sql_command == SQLCOM_CREATE_TABLE)) { @@ -8727,10 +8871,11 @@ ha_innobase::store_lock( option set or this session is using READ COMMITTED isolation level and isolation level of the transaction is not set to serializable and MySQL is doing - INSERT INTO...SELECT or UPDATE ... = (SELECT ...) or - CREATE ... SELECT... without FOR UPDATE or - IN SHARE MODE in select, then we use consistent - read for select. */ + INSERT INTO...SELECT or REPLACE INTO...SELECT + or UPDATE ... = (SELECT ...) or CREATE ... + SELECT... without FOR UPDATE or IN SHARE + MODE in select, then we use consistent read + for select. */ prebuilt->select_lock_type = LOCK_NONE; prebuilt->stored_select_lock_type = LOCK_NONE; @@ -8959,8 +9104,7 @@ ha_innobase::get_auto_increment( col_max_value = innobase_get_int_col_max_value( table->next_number_field); - current = *first_value > col_max_value ? autoinc : *first_value; - + current = *first_value > col_max_value ? autoinc : *first_value; need = *nb_reserved_values * increment; /* Compute the last value in the interval */ @@ -9324,8 +9468,7 @@ innobase_xa_prepare( executing XA PREPARE and XA COMMIT commands. In this case we cannot know how many minutes or hours will be between XA PREPARE and XA COMMIT, and we don't want - to block for undefined period of time. - */ + to block for undefined period of time. */ pthread_mutex_lock(&prepare_commit_mutex); trx->active_trans = 2; } @@ -9577,6 +9720,20 @@ ha_innobase::check_if_incompatible_data( DBUG_RETURN(COMPATIBLE_DATA_NO); } + /* Renaming column asynchronizes dictionary between mysqld and InnoDB... + If not synchronized, treat as COMPATIBLE_DATA_NO + until the bug http://bugs.mysql.com/47621 is fixed officialily */ + { + uint i; + for (i = 0; i < table->s->fields; i++) { + if (table->field[i]->flags & FIELD_IN_ADD_INDEX + && innobase_strcasecmp(table->field[i]->field_name, + dict_table_get_col_name(prebuilt->table, i))) { + return(COMPATIBLE_DATA_NO); + } + } + } + /* Check if a column participating in a foreign key is being renamed. There is no mechanism for updating InnoDB foreign key definitions. */ if (foreign_key_column_is_being_renamed(prebuilt, table)) { @@ -9685,25 +9842,24 @@ innobase_file_format_check_on_off( /************************************************************//** Validate the file format check config parameters, as a side effect it sets the srv_check_file_format_at_startup variable. -@return true if valid config value */ +@return the format_id if valid config value, otherwise, return -1 */ static -bool -innobase_file_format_check_validate( +int +innobase_file_format_validate_and_set( /*================================*/ const char* format_check) /*!< in: parameter value */ { uint format_id; - bool ret = true; format_id = innobase_file_format_name_lookup(format_check); if (format_id < DICT_TF_FORMAT_MAX + 1) { srv_check_file_format_at_startup = format_id; + + return((int) format_id); } else { - ret = false; + return(-1); } - - return(ret); } /*************************************************************//** @@ -9722,7 +9878,6 @@ innodb_file_format_name_validate( struct st_mysql_value* value) /*!< in: incoming string */ { const char* file_format_input; - char* file_format_input_strdup; char buff[STRING_BUFFER_USUAL_SIZE]; int len = sizeof(buff); @@ -9739,18 +9894,12 @@ innodb_file_format_name_validate( if (format_id <= DICT_TF_FORMAT_MAX) { - /* Copy out from stack-allocated memory (which will not - survive return from this function). The memory will be - freed in innodb_file_format_check_update(). */ - file_format_input_strdup = thd_strmake(thd, file_format_input, len); - - *static_cast<char**>(save) = file_format_input_strdup; + /* Save a pointer to the name in the + 'file_format_name_map' constant array. */ + *static_cast<const char**>(save) = + trx_sys_file_format_id_to_name(format_id); - if (file_format_input_strdup == NULL) { - return(1); - } else { - return(0); - } + return(0); } } @@ -9810,9 +9959,9 @@ innodb_file_format_check_validate( struct st_mysql_value* value) /*!< in: incoming string */ { const char* file_format_input; - char* file_format_input_strdup; char buff[STRING_BUFFER_USUAL_SIZE]; int len = sizeof(buff); + int format_id; ut_a(save != NULL); ut_a(value != NULL); @@ -9825,33 +9974,35 @@ innodb_file_format_check_validate( message if they did so. */ if (innobase_file_format_check_on_off(file_format_input)) { - sql_print_warning( + push_warning_printf(thd, + MYSQL_ERROR::WARN_LEVEL_WARN, + ER_WRONG_ARGUMENTS, "InnoDB: invalid innodb_file_format_check " "value; on/off can only be set at startup or " "in the configuration file"); - } else if (innobase_file_format_check_validate( - file_format_input)) { + } else { + format_id = innobase_file_format_validate_and_set( + file_format_input); - /* Copy out from stack-allocated memory (which will not - survive return from this function). The memory will be - freed in innodb_file_format_check_update(). */ - file_format_input_strdup = thd_strmake(thd, file_format_input, len); + if (format_id >= 0) { + /* Save a pointer to the name in the + 'file_format_name_map' constant array. */ + *static_cast<const char**>(save) = + trx_sys_file_format_id_to_name( + (uint)format_id); - *static_cast<char**>(save) = file_format_input_strdup; + return(0); - if (file_format_input_strdup == NULL) { - return(1); } else { - return(0); + push_warning_printf(thd, + MYSQL_ERROR::WARN_LEVEL_WARN, + ER_WRONG_ARGUMENTS, + "InnoDB: invalid innodb_file_format_check " + "value; can be any format up to %s " + "or its equivalent numeric id", + trx_sys_file_format_id_to_name( + DICT_TF_FORMAT_MAX)); } - - } else { - sql_print_warning( - "InnoDB: invalid innodb_file_format_check " - "value; can be any format up to %s " - "or its equivalent numeric id", - trx_sys_file_format_id_to_name( - DICT_TF_FORMAT_MAX)); } } @@ -9882,6 +10033,7 @@ innodb_file_format_check_update( ut_a(var_ptr != NULL); format_name_in = *static_cast<const char*const*>(save); + if (!format_name_in) { return; @@ -9931,6 +10083,25 @@ innodb_adaptive_hash_index_update( } } +/****************************************************************//** +Update the system variable innodb_old_blocks_pct using the "saved" +value. This function is registered as a callback with MySQL. */ +static +void +innodb_old_blocks_pct_update( +/*=========================*/ + THD* thd, /*!< in: thread handle */ + struct st_mysql_sys_var* var, /*!< in: pointer to + system variable */ + void* var_ptr,/*!< out: where the + formal string goes */ + const void* save) /*!< in: immediate result + from check function */ +{ + innobase_old_blocks_pct = buf_LRU_old_ratio_update( + *static_cast<const uint*>(save), TRUE); +} + /*************************************************************//** Check if it is a valid value of innodb_change_buffering. This function is registered as a callback with MySQL. @@ -10077,13 +10248,18 @@ static MYSQL_SYSVAR_BOOL(extra_undoslots, innobase_extra_undoslots, static MYSQL_SYSVAR_BOOL(fast_recovery, innobase_fast_recovery, PLUGIN_VAR_NOCMDARG | PLUGIN_VAR_READONLY, "Enable to use speed hack of recovery avoiding flush list sorting.", - NULL, NULL, FALSE); + NULL, NULL, TRUE); -static MYSQL_SYSVAR_BOOL(use_purge_thread, innobase_use_purge_thread, +static MYSQL_SYSVAR_BOOL(recovery_stats, innobase_recovery_stats, PLUGIN_VAR_NOCMDARG | PLUGIN_VAR_READONLY, - "Enable to use purge devoted thread.", + "Output statistics of recovery process after it.", NULL, NULL, FALSE); +static MYSQL_SYSVAR_ULONG(use_purge_thread, srv_use_purge_thread, + PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY, + "Number of purge devoted threads. #### over 1 is EXPERIMENTAL ####", + NULL, NULL, 1, 0, 64, 0); + static MYSQL_SYSVAR_BOOL(overwrite_relay_log_info, innobase_overwrite_relay_log_info, PLUGIN_VAR_NOCMDARG | PLUGIN_VAR_READONLY, "During InnoDB crash recovery on slave overwrite relay-log.info " @@ -10124,12 +10300,15 @@ static MYSQL_SYSVAR_STR(file_format, innobase_file_format_name, innodb_file_format_name_validate, innodb_file_format_name_update, "Antelope"); +/* If a new file format is introduced, the file format +name needs to be updated accordingly. Please refer to +file_format_name_map[] defined in trx0sys.c for the next +file format name. */ static MYSQL_SYSVAR_STR(file_format_check, innobase_file_format_check, PLUGIN_VAR_OPCMDARG, "The highest file format in the tablespace.", innodb_file_format_check_validate, - innodb_file_format_check_update, - "on"); + innodb_file_format_check_update, "Barracuda"); static MYSQL_SYSVAR_ULONG(flush_log_at_trx_commit, srv_flush_log_at_trx_commit, PLUGIN_VAR_OPCMDARG, @@ -10138,7 +10317,7 @@ static MYSQL_SYSVAR_ULONG(flush_log_at_trx_commit, srv_flush_log_at_trx_commit, " or 2 (write at commit, flush once per second).", NULL, NULL, 1, 0, 2, 0); -static MYSQL_SYSVAR_STR(flush_method, innobase_unix_file_flush_method, +static MYSQL_SYSVAR_STR(flush_method, innobase_file_flush_method, PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY, "With which method to flush data.", NULL, NULL, NULL); @@ -10179,7 +10358,7 @@ static MYSQL_SYSVAR_ULONG(max_dirty_pages_pct, srv_max_buf_pool_modified_pct, static MYSQL_SYSVAR_BOOL(adaptive_flushing, srv_adaptive_flushing, PLUGIN_VAR_NOCMDARG, "Attempt flushing dirty pages to avoid IO bursts at checkpoints.", - NULL, NULL, TRUE); + NULL, NULL, FALSE); static MYSQL_SYSVAR_ULONG(max_purge_lag, srv_max_purge_lag, PLUGIN_VAR_RQCMDARG, @@ -10275,7 +10454,7 @@ static MYSQL_SYSVAR_ULONG(concurrency_tickets, srv_n_free_tickets_to_enter, NULL, NULL, 500L, 1L, ~0L, 0); static MYSQL_SYSVAR_LONG(file_io_threads, innobase_file_io_threads, - PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY, + PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY | PLUGIN_VAR_NOSYSVAR, "Number of file I/O threads in InnoDB.", NULL, NULL, 4, 4, 64, 0); @@ -10314,6 +10493,18 @@ static MYSQL_SYSVAR_LONG(mirrored_log_groups, innobase_mirrored_log_groups, "Number of identical copies of log groups we keep for the database. Currently this should be set to 1.", NULL, NULL, 1, 1, 10, 0); +static MYSQL_SYSVAR_UINT(old_blocks_pct, innobase_old_blocks_pct, + PLUGIN_VAR_RQCMDARG, + "Percentage of the buffer pool to reserve for 'old' blocks.", + NULL, innodb_old_blocks_pct_update, 100 * 3 / 8, 5, 95, 0); + +static MYSQL_SYSVAR_UINT(old_blocks_time, buf_LRU_old_threshold_ms, + PLUGIN_VAR_RQCMDARG, + "Move blocks to the 'new' end of the buffer pool if the first access" + " was at least this many milliseconds ago." + " The timeout is disabled if 0 (the default).", + NULL, NULL, 0, 0, UINT_MAX32, 0); + static MYSQL_SYSVAR_LONG(open_files, innobase_open_files, PLUGIN_VAR_RQCMDARG | PLUGIN_VAR_READONLY, "How many files at the maximum InnoDB keeps open at the same time.", @@ -10392,13 +10583,18 @@ static MYSQL_SYSVAR_LONGLONG(ibuf_max_size, srv_ibuf_max_size, static MYSQL_SYSVAR_ULONG(ibuf_active_contract, srv_ibuf_active_contract, PLUGIN_VAR_RQCMDARG, "Enable/Disable active_contract of insert buffer. 0:disable 1:enable", - NULL, NULL, 0, 0, 1, 0); + NULL, NULL, 1, 0, 1, 0); static MYSQL_SYSVAR_ULONG(ibuf_accel_rate, srv_ibuf_accel_rate, PLUGIN_VAR_RQCMDARG, "Tunes amount of insert buffer processing of background, in addition to innodb_io_capacity. (in percentage)", NULL, NULL, 100, 100, 999999999, 0); +static MYSQL_SYSVAR_ULONG(checkpoint_age_target, srv_checkpoint_age_target, + PLUGIN_VAR_RQCMDARG, + "Control soft limit of checkpoint age. (0 : not control)", + NULL, NULL, 0, 0, ~0UL, 0); + static MYSQL_SYSVAR_ULONG(flush_neighbor_pages, srv_flush_neighbor_pages, PLUGIN_VAR_RQCMDARG, "Enable/Disable flushing also neighbor pages. 0:disable 1:enable", @@ -10434,7 +10630,7 @@ TYPELIB read_ahead_typelib= }; static MYSQL_SYSVAR_ENUM(read_ahead, srv_read_ahead, PLUGIN_VAR_RQCMDARG, - "Control read ahead activity. (none, random, [linear], both)", + "Control read ahead activity (none, random, [linear], both). [from 1.0.5: random read ahead is ignored]", NULL, innodb_read_ahead_update, 2, &read_ahead_typelib); static @@ -10465,8 +10661,8 @@ TYPELIB adaptive_checkpoint_typelib= }; static MYSQL_SYSVAR_ENUM(adaptive_checkpoint, srv_adaptive_checkpoint, PLUGIN_VAR_RQCMDARG, - "Enable/Disable flushing along modified age. ([none], reflex, estimate)", - NULL, innodb_adaptive_checkpoint_update, 0, &adaptive_checkpoint_typelib); + "Enable/Disable flushing along modified age. (none, reflex, [estimate])", + NULL, innodb_adaptive_checkpoint_update, 2, &adaptive_checkpoint_typelib); static MYSQL_SYSVAR_ULONG(enable_unsafe_group_commit, srv_enable_unsafe_group_commit, PLUGIN_VAR_RQCMDARG, @@ -10488,6 +10684,11 @@ static MYSQL_SYSVAR_ULONG(dict_size_limit, srv_dict_size_limit, "Limit the allocated memory for dictionary cache. (0: unlimited)", NULL, NULL, 0, 0, LONG_MAX, 0); +static MYSQL_SYSVAR_ULONG(relax_table_creation, srv_relax_table_creation, + PLUGIN_VAR_RQCMDARG, + "Relax limitation of column size at table creation as builtin InnoDB.", + NULL, NULL, 0, 0, 1, 0); + static struct st_mysql_sys_var* innobase_system_variables[]= { MYSQL_SYSVAR(additional_mem_pool_size), MYSQL_SYSVAR(autoextend_increment), @@ -10500,6 +10701,7 @@ static struct st_mysql_sys_var* innobase_system_variables[]= { MYSQL_SYSVAR(doublewrite), MYSQL_SYSVAR(extra_undoslots), MYSQL_SYSVAR(fast_recovery), + MYSQL_SYSVAR(recovery_stats), MYSQL_SYSVAR(fast_shutdown), MYSQL_SYSVAR(file_io_threads), MYSQL_SYSVAR(read_io_threads), @@ -10524,6 +10726,8 @@ static struct st_mysql_sys_var* innobase_system_variables[]= { MYSQL_SYSVAR(adaptive_flushing), MYSQL_SYSVAR(max_purge_lag), MYSQL_SYSVAR(mirrored_log_groups), + MYSQL_SYSVAR(old_blocks_pct), + MYSQL_SYSVAR(old_blocks_time), MYSQL_SYSVAR(open_files), MYSQL_SYSVAR(overwrite_relay_log_info), MYSQL_SYSVAR(rollback_on_timeout), @@ -10550,6 +10754,7 @@ static struct st_mysql_sys_var* innobase_system_variables[]= { MYSQL_SYSVAR(ibuf_max_size), MYSQL_SYSVAR(ibuf_active_contract), MYSQL_SYSVAR(ibuf_accel_rate), + MYSQL_SYSVAR(checkpoint_age_target), MYSQL_SYSVAR(flush_neighbor_pages), MYSQL_SYSVAR(read_ahead), MYSQL_SYSVAR(adaptive_checkpoint), @@ -10562,6 +10767,7 @@ static struct st_mysql_sys_var* innobase_system_variables[]= { MYSQL_SYSVAR(read_ahead_threshold), MYSQL_SYSVAR(io_capacity), MYSQL_SYSVAR(use_purge_thread), + MYSQL_SYSVAR(relax_table_creation), NULL }; @@ -10593,6 +10799,7 @@ i_s_innodb_cmpmem, i_s_innodb_cmpmem_reset, i_s_innodb_table_stats, i_s_innodb_index_stats, +i_s_innodb_admin_command, i_s_innodb_patches mysql_declare_plugin_end; diff --git a/storage/xtradb/handler/ha_innodb.h b/storage/xtradb/handler/ha_innodb.h index b88aac2ae32..31e88ed8530 100644 --- a/storage/xtradb/handler/ha_innodb.h +++ b/storage/xtradb/handler/ha_innodb.h @@ -258,12 +258,14 @@ int thd_binlog_format(const MYSQL_THD thd); */ void thd_mark_transaction_to_rollback(MYSQL_THD thd, bool all); +#if MYSQL_VERSION_ID > 50140 /** Check if binary logging is filtered for thread's current db. @param thd Thread handle @retval 1 the query is not filtered, 0 otherwise. */ bool thd_binlog_filter_ok(const MYSQL_THD thd); +#endif /* MYSQL_VERSION_ID > 50140 */ } typedef struct trx_struct trx_t; @@ -289,6 +291,8 @@ trx_t* innobase_trx_allocate( /*==================*/ MYSQL_THD thd); /*!< in: user thread handle */ + + /*********************************************************************//** This function checks each index name for a table against reserved system default primary index name 'GEN_CLUST_INDEX'. If a name diff --git a/storage/xtradb/handler/handler0alter.cc b/storage/xtradb/handler/handler0alter.cc index a523fe406ed..e060d88b3b8 100644 --- a/storage/xtradb/handler/handler0alter.cc +++ b/storage/xtradb/handler/handler0alter.cc @@ -35,7 +35,6 @@ extern "C" { } #include "ha_innodb.h" -#include "handler0vars.h" /*************************************************************//** Copies an InnoDB column to a MySQL field. This function is @@ -629,7 +628,7 @@ ha_innobase::add_index( ulint num_created = 0; ibool dict_locked = FALSE; ulint new_primary; - ulint error; + int error; DBUG_ENTER("ha_innobase::add_index"); ut_a(table); @@ -668,7 +667,7 @@ ha_innobase::add_index( if (UNIV_UNLIKELY(error)) { err_exit: mem_heap_free(heap); - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx_free_for_mysql(trx); trx_commit_for_mysql(prebuilt->trx); DBUG_RETURN(error); @@ -766,10 +765,11 @@ err_exit: ut_ad(error == DB_SUCCESS); /* Commit the data dictionary transaction in order to release - the table locks on the system tables. Unfortunately, this - means that if MySQL crashes while creating a new primary key - inside row_merge_build_indexes(), indexed_table will not be - dropped on crash recovery. Thus, it will become orphaned. */ + the table locks on the system tables. This means that if + MySQL crashes while creating a new primary key inside + row_merge_build_indexes(), indexed_table will not be dropped + by trx_rollback_active(). It will have to be recovered or + dropped by the database administrator. */ trx_commit_for_mysql(trx); row_mysql_unlock_data_dictionary(trx); @@ -806,7 +806,7 @@ error_handling: alter table t drop index b, add index (b); The fix will have to parse the SQL and note that the index - being added has the same name as the the one being dropped and + being added has the same name as the one being dropped and ignore that in the dup index check.*/ //dict_table_check_for_dup_indexes(prebuilt->table); #endif @@ -868,6 +868,7 @@ error_handling: indexed_table->n_mysql_handles_opened++; error = row_merge_drop_table(trx, innodb_table); + innodb_table = indexed_table; goto convert_error; case DB_TOO_BIG_RECORD: @@ -882,7 +883,9 @@ error: /* fall through */ default: if (new_primary) { - row_merge_drop_table(trx, indexed_table); + if (indexed_table != innodb_table) { + row_merge_drop_table(trx, indexed_table); + } } else { if (!dict_locked) { row_mysql_lock_data_dictionary(trx); diff --git a/storage/xtradb/handler/handler0vars.h b/storage/xtradb/handler/handler0vars.h deleted file mode 100644 index bed4f61ab26..00000000000 --- a/storage/xtradb/handler/handler0vars.h +++ /dev/null @@ -1,73 +0,0 @@ -/***************************************************************************** - -Copyright (c) 2008, 2009, Innobase Oy. All Rights Reserved. - -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., 59 Temple -Place, Suite 330, Boston, MA 02111-1307 USA - -*****************************************************************************/ - -/*******************************************************************//** -@file handler/handler0vars.h -This file contains accessor functions for dynamic plugin on Windows. -***********************************************************************/ - -#if defined __WIN__ && defined MYSQL_DYNAMIC_PLUGIN -/*******************************************************************//** -This is a list of externals that can not be resolved by delay loading. -They have to be resolved indirectly via their addresses in the .map file. -All of them are external variables. */ -extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_bin; -extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_latin1; -extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_filename; -extern MYSQL_PLUGIN_IMPORT CHARSET_INFO* system_charset_info; -extern MYSQL_PLUGIN_IMPORT CHARSET_INFO* default_charset_info; -//extern MYSQL_PLUGIN_IMPORT CHARSET_INFO** all_charsets; -extern MYSQL_PLUGIN_IMPORT system_variables global_system_variables; -//extern MYSQL_PLUGIN_IMPORT char* mysql_real_data_home; -extern MYSQL_PLUGIN_IMPORT char* mysql_data_home; -//extern MYSQL_PLUGIN_IMPORT char** tx_isolation_names; -//extern MYSQL_PLUGIN_IMPORT char** binlog_format_names; -//extern MYSQL_PLUGIN_IMPORT char reg_ext; -extern MYSQL_PLUGIN_IMPORT pthread_mutex_t LOCK_thread_count; -extern MYSQL_PLUGIN_IMPORT key_map key_map_full; -extern MYSQL_PLUGIN_IMPORT MY_TMPDIR mysql_tmpdir_list; -extern MYSQL_PLUGIN_IMPORT bool mysqld_embedded; -extern MYSQL_PLUGIN_IMPORT uint lower_case_table_names; -extern MYSQL_PLUGIN_IMPORT ulong specialflag; -extern MYSQL_PLUGIN_IMPORT int my_umask; - -extern MYSQL_PLUGIN_IMPORT char *relay_log_info_file; - -/* -#define my_charset_bin (*wdl_my_charset_bin) -#define my_charset_latin1 (*wdl_my_charset_latin1) -#define my_charset_filename (*wdl_my_charset_filename) -#define system_charset_info (*wdl_system_charset_info) -#define default_charset_info (*wdl_default_charset_info) -#define all_charsets (wdl_all_charsets) -#define global_system_variables (*wdl_global_system_variables) -#define mysql_real_data_home (wdl_mysql_real_data_home) -#define mysql_data_home (*wdl_mysql_data_home) -#define tx_isolation_names (wdl_tx_isolation_names) -#define binlog_format_names (wdl_binlog_format_names) -#define reg_ext (wdl_reg_ext) -#define LOCK_thread_count (*wdl_LOCK_thread_count) -#define key_map_full (*wdl_key_map_full) -#define mysql_tmpdir_list (*wdl_mysql_tmpdir_list) -#define mysqld_embedded (*wdl_mysqld_embedded) -*/ -//#define lower_case_table_names (*wdl_lower_case_table_names) -//#define specialflag (*wdl_specialflag) -//#define my_umask (*wdl_my_umask) - -#endif diff --git a/storage/xtradb/handler/i_s.cc b/storage/xtradb/handler/i_s.cc index a46583561cc..f282b141282 100644 --- a/storage/xtradb/handler/i_s.cc +++ b/storage/xtradb/handler/i_s.cc @@ -47,6 +47,7 @@ extern "C" { #include "trx0rseg.h" /* for trx_rseg_struct */ #include "trx0sys.h" /* for trx_sys */ #include "dict0dict.h" /* for dict_sys */ +#include "buf0lru.h" /* for XTRA_LRU_[DUMP/RESTORE] */ /* from buf0buf.c */ struct buf_chunk_struct{ ulint mem_size; /* allocated size of the chunk */ @@ -56,7 +57,6 @@ struct buf_chunk_struct{ buf_block_t* blocks; /* array of buffer control blocks */ }; } -#include "handler0vars.h" static const char plugin_author[] = "Innobase Oy"; @@ -84,14 +84,16 @@ do { \ #define STRUCT_FLD(name, value) value #endif -static const ST_FIELD_INFO END_OF_ST_FIELD_INFO = - {STRUCT_FLD(field_name, NULL), - STRUCT_FLD(field_length, 0), - STRUCT_FLD(field_type, MYSQL_TYPE_NULL), - STRUCT_FLD(value, 0), - STRUCT_FLD(field_flags, 0), - STRUCT_FLD(old_name, ""), - STRUCT_FLD(open_method, SKIP_OPEN_TABLE)}; +/* Don't use a static const variable here, as some C++ compilers (notably +HPUX aCC: HP ANSI C++ B3910B A.03.65) can't handle it. */ +#define END_OF_ST_FIELD_INFO \ + {STRUCT_FLD(field_name, NULL), \ + STRUCT_FLD(field_length, 0), \ + STRUCT_FLD(field_type, MYSQL_TYPE_NULL), \ + STRUCT_FLD(value, 0), \ + STRUCT_FLD(field_flags, 0), \ + STRUCT_FLD(old_name, ""), \ + STRUCT_FLD(open_method, SKIP_OPEN_TABLE)} /* Use the following types mapping: @@ -511,7 +513,7 @@ static ST_FIELD_INFO i_s_innodb_buffer_pool_pages_index_fields_info[] = STRUCT_FLD(old_name, ""), STRUCT_FLD(open_method, SKIP_OPEN_TABLE)}, - {STRUCT_FLD(field_name, "accessed"), + {STRUCT_FLD(field_name, "access_time"), STRUCT_FLD(field_length, MY_INT64_NUM_DECIMAL_DIGITS), STRUCT_FLD(field_type, MYSQL_TYPE_LONGLONG), STRUCT_FLD(value, 0), @@ -728,7 +730,7 @@ i_s_innodb_buffer_pool_pages_fill( field_store_string(table->field[0], page_type); table->field[1]->store(block->page.space); table->field[2]->store(block->page.offset); - table->field[3]->store(block->page.LRU_position); + table->field[3]->store(0); table->field[4]->store(block->page.buf_fix_count); table->field[5]->store(block->page.flush_type); @@ -817,11 +819,11 @@ i_s_innodb_buffer_pool_pages_index_fill( table->field[5]->store(page_get_n_recs(frame)); table->field[6]->store(page_get_data_size(frame)); table->field[7]->store(block->is_hashed); - table->field[8]->store(block->page.accessed); + table->field[8]->store(block->page.access_time); table->field[9]->store(block->page.newest_modification != 0); table->field[10]->store(block->page.oldest_modification != 0); table->field[11]->store(block->page.old); - table->field[12]->store(block->page.LRU_position); + table->field[12]->store(0); table->field[13]->store(block->page.buf_fix_count); table->field[14]->store(block->page.flush_type); @@ -915,7 +917,7 @@ i_s_innodb_buffer_pool_pages_blob_fill( table->field[4]->store(block->page.offset); } - table->field[5]->store(block->page.LRU_position); + table->field[5]->store(0); table->field[6]->store(block->page.buf_fix_count); table->field[7]->store(block->page.flush_type); @@ -2953,3 +2955,170 @@ UNIV_INTERN struct st_mysql_plugin i_s_innodb_index_stats = STRUCT_FLD(system_vars, NULL), STRUCT_FLD(__reserved1, NULL) }; + +/*********************************************************************** +*/ +static ST_FIELD_INFO i_s_innodb_admin_command_info[] = +{ + {STRUCT_FLD(field_name, "result_message"), + STRUCT_FLD(field_length, 1024), + STRUCT_FLD(field_type, MYSQL_TYPE_STRING), + STRUCT_FLD(value, 0), + STRUCT_FLD(field_flags, 0), + STRUCT_FLD(old_name, ""), + STRUCT_FLD(open_method, SKIP_OPEN_TABLE)}, + + END_OF_ST_FIELD_INFO +}; + +#ifndef INNODB_COMPATIBILITY_HOOKS +#error InnoDB needs MySQL to be built with #define INNODB_COMPATIBILITY_HOOKS +#endif + +extern "C" { +char **thd_query(MYSQL_THD thd); +} + +static +int +i_s_innodb_admin_command_fill( +/*==========================*/ + THD* thd, + TABLE_LIST* tables, + COND* cond) +{ + TABLE* i_s_table = (TABLE *) tables->table; + CHARSET_INFO *cs= system_charset_info; + char** query_str; + char* ptr; + char quote = '\0'; + char* command_head = "XTRA_"; + + DBUG_ENTER("i_s_innodb_admin_command_fill"); + + /* deny access to non-superusers */ + if (check_global_access(thd, PROCESS_ACL)) { + DBUG_RETURN(0); + } + + if(thd_sql_command(thd) != SQLCOM_SELECT) { + field_store_string(i_s_table->field[0], + "SELECT command is only accepted."); + goto end_func; + } + + query_str = thd_query(thd); + ptr = *query_str; + + for (; *ptr; ptr++) { + if (*ptr == quote) { + quote = '\0'; + } else if (quote) { + } else if (*ptr == '`' || *ptr == '"') { + quote = *ptr; + } else { + long i; + for (i = 0; command_head[i]; i++) { + if (toupper((int)(unsigned char)(ptr[i])) + != toupper((int)(unsigned char) + (command_head[i]))) { + goto nomatch; + } + } + break; +nomatch: + ; + } + } + + if (!*ptr) { + field_store_string(i_s_table->field[0], + "No XTRA_* command in the SQL statement." + " Please add /*!XTRA_xxxx*/ to the SQL."); + goto end_func; + } + + if (!strncasecmp("XTRA_HELLO", ptr, 10)) { + /* This is example command XTRA_HELLO */ + + ut_print_timestamp(stderr); + fprintf(stderr, " InnoDB: administration command test for XtraDB" + " 'XTRA_HELLO' was detected.\n"); + + field_store_string(i_s_table->field[0], + "Hello!"); + goto end_func; + } + else if (!strncasecmp("XTRA_LRU_DUMP", ptr, 13)) { + ut_print_timestamp(stderr); + fprintf(stderr, " InnoDB: administration command 'XTRA_LRU_DUMP'" + " was detected.\n"); + + if (buf_LRU_file_dump()) { + field_store_string(i_s_table->field[0], + "XTRA_LRU_DUMP was succeeded."); + } else { + field_store_string(i_s_table->field[0], + "XTRA_LRU_DUMP was failed."); + } + + goto end_func; + } + else if (!strncasecmp("XTRA_LRU_RESTORE", ptr, 16)) { + ut_print_timestamp(stderr); + fprintf(stderr, " InnoDB: administration command 'XTRA_LRU_RESTORE'" + " was detected.\n"); + + if (buf_LRU_file_restore()) { + field_store_string(i_s_table->field[0], + "XTRA_LRU_RESTORE was succeeded."); + } else { + field_store_string(i_s_table->field[0], + "XTRA_LRU_RESTORE was failed."); + } + + goto end_func; + } + + field_store_string(i_s_table->field[0], + "Undefined XTRA_* command."); + goto end_func; + +end_func: + if (schema_table_store_record(thd, i_s_table)) { + DBUG_RETURN(1); + } else { + DBUG_RETURN(0); + } +} + +static +int +i_s_innodb_admin_command_init( +/*==========================*/ + void* p) +{ + DBUG_ENTER("i_s_innodb_admin_command_init"); + ST_SCHEMA_TABLE* schema = (ST_SCHEMA_TABLE*) p; + + schema->fields_info = i_s_innodb_admin_command_info; + schema->fill_table = i_s_innodb_admin_command_fill; + + DBUG_RETURN(0); +} + +UNIV_INTERN struct st_mysql_plugin i_s_innodb_admin_command = +{ + STRUCT_FLD(type, MYSQL_INFORMATION_SCHEMA_PLUGIN), + STRUCT_FLD(info, &i_s_info), + STRUCT_FLD(name, "XTRADB_ADMIN_COMMAND"), + STRUCT_FLD(author, plugin_author), + STRUCT_FLD(descr, "XtraDB specific command acceptor"), + STRUCT_FLD(license, PLUGIN_LICENSE_GPL), + STRUCT_FLD(init, i_s_innodb_admin_command_init), + STRUCT_FLD(deinit, i_s_common_deinit), + STRUCT_FLD(version, 0x0100 /* 1.0 */), + STRUCT_FLD(status_vars, NULL), + STRUCT_FLD(system_vars, NULL), + STRUCT_FLD(__reserved1, NULL) +}; diff --git a/storage/xtradb/handler/i_s.h b/storage/xtradb/handler/i_s.h index 7bb03460285..6d0f426acc3 100644 --- a/storage/xtradb/handler/i_s.h +++ b/storage/xtradb/handler/i_s.h @@ -40,5 +40,6 @@ extern struct st_mysql_plugin i_s_innodb_patches; extern struct st_mysql_plugin i_s_innodb_rseg; extern struct st_mysql_plugin i_s_innodb_table_stats; extern struct st_mysql_plugin i_s_innodb_index_stats; +extern struct st_mysql_plugin i_s_innodb_admin_command; #endif /* i_s_h */ diff --git a/storage/xtradb/handler/innodb_patch_info.h b/storage/xtradb/handler/innodb_patch_info.h index f8a728d0d27..253f6360aa0 100644 --- a/storage/xtradb/handler/innodb_patch_info.h +++ b/storage/xtradb/handler/innodb_patch_info.h @@ -38,5 +38,10 @@ struct innodb_enhancement { {"innodb_stats","Additional features about InnoDB statistics/optimizer","","http://www.percona.com/docs/wiki/percona-xtradb"}, {"innodb_recovery_patches","Bugfixes and adjustments about recovery process","","http://www.percona.com/docs/wiki/percona-xtradb"}, {"innodb_purge_thread","Enable to use purge devoted thread","","http://www.percona.com/docs/wiki/percona-xtradb"}, +{"innodb_admin_command_base","XtraDB specific command interface through i_s","","http://www.percona.com/docs/wiki/percona-xtradb"}, +{"innodb_show_lock_name","Show mutex/lock name instead of crated file/line","","http://www.percona.com/docs/wiki/percona-xtradb"}, +{"innodb_extend_slow","Extended statistics in slow.log","It is InnoDB-part only. It needs to patch also to mysqld.","http://www.percona.com/docs/wiki/percona-xtradb"}, +{"innodb_relax_table_creation","Relax limitation of column size at table creation as builtin InnoDB.","","http://www.percona.com/docs/wiki/percona-xtradb"}, +{"innodb_lru_dump_restore","Dump and restore command for content of buffer pool","","http://www.percona.com/docs/wiki/percona-xtradb"}, {NULL, NULL, NULL, NULL} }; diff --git a/storage/xtradb/handler/win_delay_loader.cc b/storage/xtradb/handler/win_delay_loader.cc deleted file mode 100644 index 9b92f6a9cf2..00000000000 --- a/storage/xtradb/handler/win_delay_loader.cc +++ /dev/null @@ -1,1024 +0,0 @@ -/***************************************************************************** - -Copyright (c) 2008, 2009, Innobase Oy. All Rights Reserved. - -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., 59 Temple -Place, Suite 330, Boston, MA 02111-1307 USA - -*****************************************************************************/ - -/*******************************************************************//** -@file handler/win_delay_loader.cc -This file contains functions that implement the delay loader on Windows. - -This is a customized version of delay loader with limited functionalities. -It does not support: - -* (manual) unloading -* multiple delay loaded DLLs -* multiple loading of the same DLL - -This delay loader is used only by the InnoDB plugin. Other components (DLLs) -can still use the default delay loader, provided by MSVC. - -Several acronyms used by Microsoft: - * IAT: import address table - * INT: import name table - * RVA: Relative Virtual Address - -See http://msdn.microsoft.com/en-us/magazine/bb985992.aspx for details of -PE format. -***********************************************************************/ -#if defined (__WIN__) && defined (MYSQL_DYNAMIC_PLUGIN) -# define WIN32_LEAN_AND_MEAN -# include <windows.h> -# include <delayimp.h> -# include <mysql_priv.h> - -extern "C" { -# include "univ.i" -# include "hash0hash.h" -} - -/*******************************************************************//** -This following contains a list of externals that can not be resolved by -delay loading. They have to be resolved indirectly via their addresses -in the .map file. All of them are external variables. */ -CHARSET_INFO* wdl_my_charset_bin; -CHARSET_INFO* wdl_my_charset_latin1; -CHARSET_INFO* wdl_my_charset_filename; -CHARSET_INFO** wdl_system_charset_info; -CHARSET_INFO** wdl_default_charset_info; -CHARSET_INFO** wdl_all_charsets; -system_variables* wdl_global_system_variables; -char* wdl_mysql_real_data_home; -char** wdl_mysql_data_home; -char** wdl_tx_isolation_names; -char** wdl_binlog_format_names; -char* wdl_reg_ext; -pthread_mutex_t* wdl_LOCK_thread_count; -key_map* wdl_key_map_full; -MY_TMPDIR* wdl_mysql_tmpdir_list; -bool* wdl_mysqld_embedded; -uint* wdl_lower_case_table_names; -ulong* wdl_specialflag; -int* wdl_my_umask; - -/*******************************************************************//** -The preferred load-address defined in PE (portable executable format). */ -#if defined(_M_IA64) -#pragma section(".base", long, read) -extern "C" -__declspec(allocate(".base")) -const IMAGE_DOS_HEADER __ImageBase; -#else -extern "C" -const IMAGE_DOS_HEADER __ImageBase; -#endif - -/*******************************************************************//** -A template function for converting a relative address (RVA) to an -absolute address (VA). This is due to the pointers in the delay -descriptor (ImgDelayDescr in delayimp.h) have been changed from -VAs to RVAs to work on both 32- and 64-bit platforms. -@return absolute virtual address */ -template <class X> -X PFromRva( -/*=======*/ - RVA rva) /*!< in: relative virtual address */ -{ - return X(PBYTE(&__ImageBase) + rva); -} - -/*******************************************************************//** -Convert to the old format for convenience. The structure as well as its -element names follow the definition of ImgDelayDescr in delayimp.h. */ -struct InternalImgDelayDescr -{ - DWORD grAttrs; /*!< attributes */ - LPCSTR szName; /*!< pointer to dll name */ - HMODULE* phmod; /*!< address of module handle */ - PImgThunkData pIAT; /*!< address of the IAT */ - PCImgThunkData pINT; /*!< address of the INT */ - PCImgThunkData pBoundIAT; /*!< address of the optional bound IAT */ - PCImgThunkData pUnloadIAT; /*!< address of optional copy of - original IAT */ - DWORD dwTimeStamp; /*!< 0 if not bound, - otherwise date/time stamp of DLL - bound to (Old BIND) */ -}; - -typedef struct map_hash_chain_struct map_hash_chain_t; - -struct map_hash_chain_struct { - char* symbol; /*!< pointer to a symbol */ - ulint value; /*!< address of the symbol */ - map_hash_chain_t* next; /*!< pointer to the next cell - in the same folder. */ - map_hash_chain_t* chain; /*!< a linear chain used for - cleanup. */ -}; - -static HMODULE my_hmod = 0; -static struct hash_table_struct* m_htbl = NULL ; -static map_hash_chain_t* chain_header = NULL; -static ibool wdl_init = FALSE; -const ulint MAP_HASH_CELLS_NUM = 10000; - -#ifndef DBUG_OFF -/*******************************************************************//** -In the dynamic plugin, it is required to call the following dbug functions -in the server: - _db_pargs_ - _db_doprnt_ - _db_enter_ - _db_return_ - _db_dump_ - -The plugin will get those function pointers during the initialization. */ -typedef void (__cdecl* pfn_db_enter_)( - const char* _func_, - const char* _file_, - uint _line_, - const char** _sfunc_, - const char** _sfile_, - uint* _slevel_, - char***); - -typedef void (__cdecl* pfn_db_return_)( - uint _line_, - const char** _sfunc_, - const char** _sfile_, - uint* _slevel_); - -typedef void (__cdecl* pfn_db_pargs_)( - uint _line_, - const char* keyword); - -typedef void (__cdecl* pfn_db_doprnt_)( - const char* format, - ...); - -typedef void (__cdecl* pfn_db_dump_)( - uint _line_, - const char* keyword, - const unsigned char* memory, - size_t length); - -static pfn_db_enter_ wdl_db_enter_; -static pfn_db_return_ wdl_db_return_; -static pfn_db_pargs_ wdl_db_pargs_; -static pfn_db_doprnt_ wdl_db_doprnt_; -static pfn_db_dump_ wdl_db_dump_; -#endif /* !DBUG_OFF */ - -/*************************************************************//** -Creates a hash table with >= n array cells. The actual number of cells is -chosen to be a prime number slightly bigger than n. - -This is the same function as hash_create in hash0hash.c, except the -memory allocation. This function is invoked before the engine is -initialized, and buffer pools are not ready yet. -@return own: created hash table */ -static -hash_table_t* -wdl_hash_create( -/*============*/ - ulint n) /*!< in: number of array cells */ -{ - hash_cell_t* array; - ulint prime; - hash_table_t* table; - - prime = ut_find_prime(n); - - table = (hash_table_t*) malloc(sizeof(hash_table_t)); - if (table == NULL) { - return(NULL); - } - - array = (hash_cell_t*) malloc(sizeof(hash_cell_t) * prime); - if (array == NULL) { - free(table); - return(NULL); - } - - table->array = array; - table->n_cells = prime; - table->n_mutexes = 0; - table->mutexes = NULL; - table->heaps = NULL; - table->heap = NULL; - table->magic_n = HASH_TABLE_MAGIC_N; - - /* Initialize the cell array */ - hash_table_clear(table); - - return(table); -} - -/*************************************************************//** -Frees a hash table. */ -static -void -wdl_hash_table_free( -/*================*/ - hash_table_t* table) /*!< in, own: hash table */ -{ - ut_a(table != NULL); - ut_a(table->mutexes == NULL); - - free(table->array); - free(table); -} - -/*******************************************************************//** -Function for calculating the count of imports given the base of the IAT. -@return number of imports */ -static -ulint -wdl_import_count( -/*=============*/ - PCImgThunkData pitd_base) /*!< in: base of the IAT */ -{ - ulint ret = 0; - PCImgThunkData pitd = pitd_base; - - while (pitd->u1.Function) { - pitd++; - ret++; - } - - return(ret); -} - -/*******************************************************************//** -Read Mapfile to a hashtable for faster access -@return TRUE if the mapfile is loaded successfully. */ -static -ibool -wdl_load_mapfile( -/*=============*/ - const char* filename) /*!< in: name of the mapfile. */ -{ - FILE* fp; - const size_t nSize = 256; - char tmp_buf[nSize]; - char* func_name; - char* func_addr; - ulint load_addr = 0; - ibool valid_load_addr = FALSE; -#ifdef _WIN64 - const char* tmp_string = " Preferred load address is %16llx"; -#else - const char* tmp_string = " Preferred load address is %08x"; -#endif - - fp = fopen(filename, "r"); - if (fp == NULL) { - - return(FALSE); - } - - /* Check whether to create the hashtable */ - if (m_htbl == NULL) { - - m_htbl = wdl_hash_create(MAP_HASH_CELLS_NUM); - - if (m_htbl == NULL) { - - fclose(fp); - return(FALSE); - } - } - - /* Search start of symbol list and get the preferred load address */ - while (fgets(tmp_buf, sizeof(tmp_buf), fp)) { - - if (sscanf(tmp_buf, tmp_string, &load_addr) == 1) { - - valid_load_addr = TRUE; - } - - if (strstr(tmp_buf, "Rva+Base") != NULL) { - - break; - } - } - - if (valid_load_addr == FALSE) { - - /* No "Preferred load address", the map file is wrong. */ - fclose(fp); - return(FALSE); - } - - /* Read symbol list */ - while (fgets(tmp_buf, sizeof(tmp_buf), fp)) - { - map_hash_chain_t* map_cell; - ulint map_fold; - - if (*tmp_buf == 0) { - - continue; - } - - func_name = strtok(tmp_buf, " "); - func_name = strtok(NULL, " "); - func_addr = strtok(NULL, " "); - - if (func_name && func_addr) { - - ut_snprintf(tmp_buf, nSize, "0x%s", func_addr); - if (*func_name == '_') { - - func_name++; - } - - map_cell = (map_hash_chain_t*) - malloc(sizeof(map_hash_chain_t)); - if (map_cell == NULL) { - return(FALSE); - } - - /* Chain all cells together */ - map_cell->chain = chain_header; - chain_header = map_cell; - - map_cell->symbol = strdup(func_name); - map_cell->value = (ulint) _strtoui64(tmp_buf, NULL, 0) - - load_addr; - map_fold = ut_fold_string(map_cell->symbol); - - HASH_INSERT(map_hash_chain_t, - next, - m_htbl, - map_fold, - map_cell); - } - } - - fclose(fp); - - return(TRUE); -} - -/*************************************************************//** -Cleanup.during DLL unload */ -static -void -wdl_cleanup(void) -/*=============*/ -{ - while (chain_header != NULL) { - map_hash_chain_t* tmp; - - tmp = chain_header->chain; - free(chain_header->symbol); - free(chain_header); - chain_header = tmp; - } - - if (m_htbl != NULL) { - - wdl_hash_table_free(m_htbl); - } -} - -/*******************************************************************//** -Load the mapfile mysqld.map. -@return the module handle */ -static -HMODULE -wdl_get_mysqld_mapfile(void) -/*========================*/ -{ - char file_name[MAX_PATH]; - char* ext; - ulint err; - - if (my_hmod == 0) { - - size_t nSize = MAX_PATH - strlen(".map") -1; - - /* First find out the name of current executable */ - my_hmod = GetModuleHandle(NULL); - if (my_hmod == 0) { - - return(my_hmod); - } - - err = GetModuleFileName(my_hmod, file_name, nSize); - if (err == 0) { - - my_hmod = 0; - return(my_hmod); - } - - ext = strrchr(file_name, '.'); - if (ext != NULL) { - - *ext = 0; - strcat(file_name, ".map"); - - err = wdl_load_mapfile(file_name); - if (err == 0) { - - my_hmod = 0; - } - } else { - - my_hmod = 0; - } - } - - return(my_hmod); -} - -/*******************************************************************//** -Retrieves the address of an exported function. It follows the convention -of GetProcAddress(). -@return address of exported function. */ -static -FARPROC -wdl_get_procaddr_from_map( -/*======================*/ - HANDLE m_handle, /*!< in: module handle */ - const char* import_proc) /*!< in: procedure name */ -{ - map_hash_chain_t* hash_chain; - ulint map_fold; - - map_fold = ut_fold_string(import_proc); - HASH_SEARCH( - next, - m_htbl, - map_fold, - map_hash_chain_t*, - hash_chain, - , - (ut_strcmp(hash_chain->symbol, import_proc) == 0)); - - if (hash_chain == NULL) { - -#ifdef _WIN64 - /* On Win64, the leading '_' may not be taken out. In this - case, search again without the leading '_'. */ - if (*import_proc == '_') { - - import_proc++; - } - - map_fold = ut_fold_string(import_proc); - HASH_SEARCH( - next, - m_htbl, - map_fold, - map_hash_chain_t*, - hash_chain, - , - (ut_strcmp(hash_chain->symbol, import_proc) == 0)); - - if (hash_chain == NULL) { -#endif - if (wdl_init == TRUE) { - - sql_print_error( - "InnoDB: the procedure pointer of %s" - " is not found.", - import_proc); - } - - return(0); -#ifdef _WIN64 - } -#endif - } - - return((FARPROC) ((ulint) m_handle + hash_chain->value)); -} - -/*******************************************************************//** -Retrieves the address of an exported variable. -Note: It does not follow the Windows call convention FARPROC. -@return address of exported variable. */ -static -void* -wdl_get_varaddr_from_map( -/*=====================*/ - HANDLE m_handle, /*!< in: module handle */ - const char* import_variable) /*!< in: variable name */ -{ - map_hash_chain_t* hash_chain; - ulint map_fold; - - map_fold = ut_fold_string(import_variable); - HASH_SEARCH( - next, - m_htbl, - map_fold, - map_hash_chain_t*, - hash_chain, - , - (ut_strcmp(hash_chain->symbol, import_variable) == 0)); - - if (hash_chain == NULL) { - -#ifdef _WIN64 - /* On Win64, the leading '_' may not be taken out. In this - case, search again without the leading '_'. */ - if (*import_variable == '_') { - - import_variable++; - } - - map_fold = ut_fold_string(import_variable); - HASH_SEARCH( - next, - m_htbl, - map_fold, - map_hash_chain_t*, - hash_chain, - , - (ut_strcmp(hash_chain->symbol, import_variable) == 0)); - - if (hash_chain == NULL) { -#endif - if (wdl_init == TRUE) { - - sql_print_error( - "InnoDB: the variable address of %s" - " is not found.", - import_variable); - } - - return(0); -#ifdef _WIN64 - } -#endif - } - - return((void*) ((ulint) m_handle + hash_chain->value)); -} - -/*******************************************************************//** -Bind all unresolved external variables from the MySQL executable. -@return TRUE if successful */ -static -bool -wdl_get_external_variables(void) -/*============================*/ -{ - HMODULE hmod = wdl_get_mysqld_mapfile(); - - if (hmod == 0) { - - return(FALSE); - } - -#define GET_SYM(sym, var, type) \ - var = (type*) wdl_get_varaddr_from_map(hmod, sym); \ - if (var == NULL) return(FALSE) -#ifdef _WIN64 -#define GET_SYM2(sym1, sym2, var, type) \ - var = (type*) wdl_get_varaddr_from_map(hmod, sym1); \ - if (var == NULL) return(FALSE) -#else -#define GET_SYM2(sym1, sym2, var, type) \ - var = (type*) wdl_get_varaddr_from_map(hmod, sym2); \ - if (var == NULL) return(FALSE) -#endif // (_WIN64) -#define GET_C_SYM(sym, type) GET_SYM(#sym, wdl_##sym, type) -#define GET_PROC_ADDR(sym) \ - wdl##sym = (pfn##sym) wdl_get_procaddr_from_map(hmod, #sym) - - GET_C_SYM(my_charset_bin, CHARSET_INFO); - GET_C_SYM(my_charset_latin1, CHARSET_INFO); - GET_C_SYM(my_charset_filename, CHARSET_INFO); - GET_C_SYM(default_charset_info, CHARSET_INFO*); - GET_C_SYM(all_charsets, CHARSET_INFO*); - GET_C_SYM(my_umask, int); - - GET_SYM("?global_system_variables@@3Usystem_variables@@A", - wdl_global_system_variables, struct system_variables); - GET_SYM("?mysql_real_data_home@@3PADA", - wdl_mysql_real_data_home, char); - GET_SYM("?reg_ext@@3PADA", wdl_reg_ext, char); - GET_SYM("?LOCK_thread_count@@3U_RTL_CRITICAL_SECTION@@A", - wdl_LOCK_thread_count, pthread_mutex_t); - GET_SYM("?key_map_full@@3V?$Bitmap@$0EA@@@A", - wdl_key_map_full, key_map); - GET_SYM("?mysql_tmpdir_list@@3Ust_my_tmpdir@@A", - wdl_mysql_tmpdir_list, MY_TMPDIR); - GET_SYM("?mysqld_embedded@@3_NA", - wdl_mysqld_embedded, bool); - GET_SYM("?lower_case_table_names@@3IA", - wdl_lower_case_table_names, uint); - GET_SYM("?specialflag@@3KA", wdl_specialflag, ulong); - - GET_SYM2("?system_charset_info@@3PEAUcharset_info_st@@EA", - "?system_charset_info@@3PAUcharset_info_st@@A", - wdl_system_charset_info, CHARSET_INFO*); - GET_SYM2("?mysql_data_home@@3PEADEA", - "?mysql_data_home@@3PADA", - wdl_mysql_data_home, char*); - GET_SYM2("?tx_isolation_names@@3PAPEBDA", - "?tx_isolation_names@@3PAPBDA", - wdl_tx_isolation_names, char*); - GET_SYM2("?binlog_format_names@@3PAPEBDA", - "?binlog_format_names@@3PAPBDA", - wdl_binlog_format_names, char*); - -#ifndef DBUG_OFF - GET_PROC_ADDR(_db_enter_); - GET_PROC_ADDR(_db_return_); - GET_PROC_ADDR(_db_pargs_); - GET_PROC_ADDR(_db_doprnt_); - GET_PROC_ADDR(_db_dump_); - - /* If any of the dbug functions is not available, just make them - all invalid. This is the case when working with a non-debug - version of the server. */ - if (wdl_db_enter_ == NULL || wdl_db_return_ == NULL - || wdl_db_pargs_ == NULL || wdl_db_doprnt_ == NULL - || wdl_db_dump_ == NULL) { - - wdl_db_enter_ = NULL; - wdl_db_return_ = NULL; - wdl_db_pargs_ = NULL; - wdl_db_doprnt_ = NULL; - wdl_db_dump_ = NULL; - } -#endif /* !DBUG_OFF */ - - wdl_init = TRUE; - return(TRUE); - -#undef GET_SYM -#undef GET_SYM2 -#undef GET_C_SYM -#undef GET_PROC_ADDR -} - -/*******************************************************************//** -The DLL Delayed Loading Helper Function for resolving externals. - -The function may fail due to one of the three reasons: - -* Invalid parameter, which happens if the attributes in pidd aren't - specified correctly. -* Failed to load the map file mysqld.map. -* Failed to find an external name in the map file mysqld.map. - -Note: this function is called by run-time as well as __HrLoadAllImportsForDll. -So, it has to follow Windows call convention. -@return the address of the imported function */ -extern "C" -FARPROC WINAPI -__delayLoadHelper2( -/*===============*/ - PCImgDelayDescr pidd, /*!< in: a const pointer to a - ImgDelayDescr, see delayimp.h. */ - FARPROC* iat_entry) /*!< in/out: A pointer to the slot in - the delay load import address table - to be updated with the address of the - imported function. */ -{ - ulint iIAT, iINT; - HMODULE hmod; - PCImgThunkData pitd; - FARPROC fun = NULL; - - /* Set up data used for the hook procs */ - InternalImgDelayDescr idd = { - pidd->grAttrs, - PFromRva<LPCSTR>(pidd->rvaDLLName), - PFromRva<HMODULE*>(pidd->rvaHmod), - PFromRva<PImgThunkData>(pidd->rvaIAT), - PFromRva<PCImgThunkData>(pidd->rvaINT), - PFromRva<PCImgThunkData>(pidd->rvaBoundIAT), - PFromRva<PCImgThunkData>(pidd->rvaUnloadIAT), - pidd->dwTimeStamp - }; - - DelayLoadInfo dli = { - sizeof(DelayLoadInfo), - pidd, - iat_entry, - idd.szName, - {0}, - 0, - 0, - 0 - }; - - /* Check the Delay Load Attributes, log an error of invalid - parameter, which happens if the attributes in pidd are not - specified correctly. */ - if ((idd.grAttrs & dlattrRva) == 0) { - - sql_print_error("InnoDB: invalid parameter for delay loader."); - return(0); - } - - hmod = *idd.phmod; - - /* Calculate the index for the IAT entry in the import address table. - The INT entries are ordered the same as the IAT entries so the - calculation can be done on the IAT side. */ - iIAT = (PCImgThunkData) iat_entry - idd.pIAT; - iINT = iIAT; - - pitd = &(idd.pINT[iINT]); - - dli.dlp.fImportByName = !IMAGE_SNAP_BY_ORDINAL(pitd->u1.Ordinal); - - if (dli.dlp.fImportByName) { - - dli.dlp.szProcName = (LPCSTR) (PFromRva<PIMAGE_IMPORT_BY_NAME> - ((RVA) ((UINT_PTR) pitd->u1.AddressOfData))->Name); - } else { - - dli.dlp.dwOrdinal = (ulint) IMAGE_ORDINAL(pitd->u1.Ordinal); - } - - /* Now, load the mapfile, if it has not been done yet */ - if (hmod == 0) { - - hmod = wdl_get_mysqld_mapfile(); - } - - if (hmod == 0) { - /* LoadLibrary failed. */ - PDelayLoadInfo rgpdli[1] = {&dli}; - - dli.dwLastError = ::GetLastError(); - - sql_print_error( - "InnoDB: failed to load mysqld.map with error %d.", - dli.dwLastError); - - return(0); - } - - /* Store the library handle. */ - idd.phmod = &hmod; - - /* Go for the procedure now. */ - dli.hmodCur = hmod; - - if (pidd->rvaBoundIAT && pidd->dwTimeStamp) { - - /* Bound imports exist, check the timestamp from the target - image */ - PIMAGE_NT_HEADERS pinh; - - pinh = (PIMAGE_NT_HEADERS) ((byte*) hmod - + ((PIMAGE_DOS_HEADER) hmod)->e_lfanew); - - if (pinh->Signature == IMAGE_NT_SIGNATURE - && pinh->FileHeader.TimeDateStamp == idd.dwTimeStamp - && (DWORD) hmod == pinh->OptionalHeader.ImageBase) { - - /* We have a decent address in the bound IAT. */ - fun = (FARPROC) (UINT_PTR) - idd.pBoundIAT[iIAT].u1.Function; - - if (fun) { - - *iat_entry = fun; - return(fun); - } - } - } - - fun = wdl_get_procaddr_from_map(hmod, dli.dlp.szProcName); - - if (fun == 0) { - - return(0); - } - - *iat_entry = fun; - return(fun); -} - -/*******************************************************************//** -Unload a DLL that was delay loaded. This function is called by run-time. -@return TRUE is returned if the DLL is found and the IAT matches the -original one. */ -extern "C" -BOOL WINAPI -__FUnloadDelayLoadedDLL2( -/*=====================*/ - LPCSTR module_name) /*!< in: DLL name */ -{ - return(TRUE); -} - -/**************************************************************//** -Load all imports from a DLL that was specified with the /delayload linker -option. -Note: this function is called by run-time. So, it has to follow Windows call -convention. -@return S_OK if the DLL matches, otherwise ERROR_MOD_NOT_FOUND is returned. */ -extern "C" -HRESULT WINAPI -__HrLoadAllImportsForDll( -/*=====================*/ - LPCSTR module_name) /*!< in: DLL name */ -{ - PIMAGE_NT_HEADERS img; - PCImgDelayDescr pidd; - IMAGE_DATA_DIRECTORY* image_data; - LPCSTR current_module; - HRESULT ret = ERROR_MOD_NOT_FOUND; - HMODULE hmod = (HMODULE) &__ImageBase; - - img = (PIMAGE_NT_HEADERS) ((byte*) hmod - + ((PIMAGE_DOS_HEADER) hmod)->e_lfanew); - image_data = - &img->OptionalHeader.DataDirectory[IMAGE_DIRECTORY_ENTRY_DELAY_IMPORT]; - - /* Scan the delay load IAT/INT for the DLL */ - if (image_data->Size) { - - pidd = PFromRva<PCImgDelayDescr>(image_data->VirtualAddress); - - /* Check all of the listed DLLs we want to load. */ - while (pidd->rvaDLLName) { - - current_module = PFromRva<LPCSTR>(pidd->rvaDLLName); - - if (stricmp(module_name, current_module) == 0) { - - /* Found it, break out with pidd and - current_module set appropriately */ - break; - } - - /* To the next delay import descriptor */ - pidd++; - } - - if (pidd->rvaDLLName) { - - /* Found a matching DLL, now process it. */ - FARPROC* iat_entry; - size_t count; - - iat_entry = PFromRva<FARPROC*>(pidd->rvaIAT); - count = wdl_import_count((PCImgThunkData) iat_entry); - - /* now load all the imports from the DLL */ - while (count > 0) { - - /* No need to check the return value */ - __delayLoadHelper2(pidd, iat_entry); - iat_entry++; - count--; - } - - ret = S_OK; - } - } - - return ret; -} - -/**************************************************************//** -The main function of a DLL -@return TRUE if the call succeeds */ -BOOL -WINAPI -DllMain( -/*====*/ - HINSTANCE hinstDLL, /*!< in: handle to the DLL module */ - DWORD fdwReason, /*!< Reason code that indicates why the - DLL entry-point function is being - called.*/ - LPVOID lpvReserved) /*!< in: additional parameter based on - fdwReason */ -{ - BOOL success = TRUE; - - switch (fdwReason) { - - case DLL_PROCESS_ATTACH: - success = wdl_get_external_variables(); - break; - - case DLL_PROCESS_DETACH: - wdl_cleanup(); - break; - } - - return(success); -} - -#ifndef DBUG_OFF -/**************************************************************//** -Process entry point to user function. It makes the call to _db_enter_ -in mysqld.exe. The DBUG functions are defined in my_dbug.h. */ -extern "C" UNIV_INTERN -void -_db_enter_( - const char* _func_, /*!< in: current function name */ - const char* _file_, /*!< in: current file name */ - uint _line_, /*!< in: current source line number */ - const char** _sfunc_, /*!< out: previous _func_ */ - const char** _sfile_, /*!< out: previous _file_ */ - uint* _slevel_, /*!< out: previous nesting level */ - char*** _sframep_) /*!< out: previous frame pointer */ -{ - if (wdl_db_enter_ != NULL) { - - wdl_db_enter_(_func_, _file_, _line_, _sfunc_, _sfile_, - _slevel_, _sframep_); - } -} - -/**************************************************************//** -Process exit from user function. It makes the call to _db_return_() -in the server. */ -extern "C" UNIV_INTERN -void -_db_return_( - uint _line_, /*!< in: current source line number */ - const char** _sfunc_, /*!< out: previous _func_ */ - const char** _sfile_, /*!< out: previous _file_ */ - uint* _slevel_) /*!< out: previous level */ -{ - if (wdl_db_return_ != NULL) { - - wdl_db_return_(_line_, _sfunc_, _sfile_, _slevel_); - } -} - -/**************************************************************//** -Log arguments for subsequent use. It makes the call to _db_pargs_() -in the server. */ -extern "C" UNIV_INTERN -void -_db_pargs_( - uint _line_, /*!< in: current source line number */ - const char* keyword) /*!< in: keyword for current macro */ -{ - if (wdl_db_pargs_ != NULL) { - - wdl_db_pargs_(_line_, keyword); - } -} - -/**************************************************************//** -Handle print of debug lines. It saves the text into a buffer first, -then makes the call to _db_doprnt_() in the server. The text is -truncated to the size of buffer. */ -extern "C" UNIV_INTERN -void -_db_doprnt_( - const char* format, /*!< in: the format string */ - ...) /*!< in: list of arguments */ -{ - va_list argp; - char buffer[512]; - - if (wdl_db_doprnt_ != NULL) { - - va_start(argp, format); - /* it is ok to ignore the trunction. */ - _vsnprintf(buffer, sizeof(buffer), format, argp); - wdl_db_doprnt_(buffer); - va_end(argp); - } -} - -/**************************************************************//** -Dump a string in hex. It makes the call to _db_dump_() in the server. */ -extern "C" UNIV_INTERN -void -_db_dump_( - uint _line_, /*!< in: current source line - number */ - const char* keyword, /*!< in: keyword list */ - const unsigned char* memory, /*!< in: memory to dump */ - size_t length) /*!< in: bytes to dump */ -{ - if (wdl_db_dump_ != NULL) { - - wdl_db_dump_(_line_, keyword, memory, length); - } -} - -#endif /* !DBUG_OFF */ -#endif /* defined (__WIN__) && defined (MYSQL_DYNAMIC_PLUGIN) */ diff --git a/storage/xtradb/ibuf/ibuf0ibuf.c b/storage/xtradb/ibuf/ibuf0ibuf.c index a32efbd061f..94f3751dd04 100644 --- a/storage/xtradb/ibuf/ibuf0ibuf.c +++ b/storage/xtradb/ibuf/ibuf0ibuf.c @@ -390,6 +390,27 @@ ibuf_count_set( #endif /******************************************************************//** +Closes insert buffer and frees the data structures. */ +UNIV_INTERN +void +ibuf_close(void) +/*============*/ +{ + mutex_free(&ibuf_pessimistic_insert_mutex); + memset(&ibuf_pessimistic_insert_mutex, + 0x0, sizeof(ibuf_pessimistic_insert_mutex)); + + mutex_free(&ibuf_mutex); + memset(&ibuf_mutex, 0x0, sizeof(ibuf_mutex)); + + mutex_free(&ibuf_bitmap_mutex); + memset(&ibuf_bitmap_mutex, 0x0, sizeof(ibuf_mutex)); + + mem_free(ibuf); + ibuf = NULL; +} + +/******************************************************************//** Updates the size information of the ibuf, assuming the segment size has not changed. */ static diff --git a/storage/xtradb/include/btr0cur.h b/storage/xtradb/include/btr0cur.h index b2d43ae3254..480a3877e54 100644 --- a/storage/xtradb/include/btr0cur.h +++ b/storage/xtradb/include/btr0cur.h @@ -618,7 +618,7 @@ enum btr_cur_method { hash_node, and might be necessary to update */ BTR_CUR_BINARY, /*!< success using the binary search */ - BTR_CUR_INSERT_TO_IBUF, /*!< performed the intended insert to + BTR_CUR_INSERT_TO_IBUF /*!< performed the intended insert to the insert buffer */ }; diff --git a/storage/xtradb/include/btr0sea.h b/storage/xtradb/include/btr0sea.h index 5c50829e874..7f4842d0897 100644 --- a/storage/xtradb/include/btr0sea.h +++ b/storage/xtradb/include/btr0sea.h @@ -41,6 +41,12 @@ void btr_search_sys_create( /*==================*/ ulint hash_size); /*!< in: hash index hash table size */ +/*****************************************************************//** +Frees the adaptive search system at a database shutdown. */ +UNIV_INTERN +void +btr_search_sys_free(void); +/*=====================*/ /********************************************************************//** Disable the adaptive hash search system and empty the index. */ diff --git a/storage/xtradb/include/buf0buf.h b/storage/xtradb/include/buf0buf.h index 8063f9ec2c8..ddd08c49a12 100644 --- a/storage/xtradb/include/buf0buf.h +++ b/storage/xtradb/include/buf0buf.h @@ -346,7 +346,7 @@ buf_page_release( mtr_t* mtr); /*!< in: mtr */ /********************************************************************//** Moves a page to the start of the buffer pool LRU list. This high-level -function can be used to prevent an important page from from slipping out of +function can be used to prevent an important page from slipping out of the buffer pool. */ UNIV_INTERN void @@ -707,15 +707,6 @@ buf_page_belongs_to_unzip_LRU( /*==========================*/ const buf_page_t* bpage) /*!< in: pointer to control block */ __attribute__((pure)); -/*********************************************************************//** -Determine the approximate LRU list position of a block. -@return LRU list position */ -UNIV_INLINE -ulint -buf_page_get_LRU_position( -/*======================*/ - const buf_page_t* bpage) /*!< in: control block */ - __attribute__((pure)); /*********************************************************************//** Gets the mutex of a block. @@ -825,14 +816,14 @@ buf_page_set_old( buf_page_t* bpage, /*!< in/out: control block */ ibool old); /*!< in: old */ /*********************************************************************//** -Determine if a block has been accessed in the buffer pool. -@return TRUE if accessed */ +Determine the time of first access of a block in the buffer pool. +@return ut_time_ms() at the time of first access, 0 if not accessed */ UNIV_INLINE -ibool +unsigned buf_page_is_accessed( /*=================*/ const buf_page_t* bpage) /*!< in: control block */ - __attribute__((pure)); + __attribute__((nonnull, pure)); /*********************************************************************//** Flag a block accessed. */ UNIV_INLINE @@ -840,7 +831,8 @@ void buf_page_set_accessed( /*==================*/ buf_page_t* bpage, /*!< in/out: control block */ - ibool accessed); /*!< in: accessed */ + ulint time_ms) /*!< in: ut_time_ms() */ + __attribute__((nonnull)); /*********************************************************************//** Gets the buf_block_t handle of a buffered file block if an uncompressed page frame exists, or NULL. @@ -1026,14 +1018,6 @@ buf_block_hash_get( /*===============*/ ulint space, /*!< in: space id */ ulint offset);/*!< in: offset of the page within space */ -/*******************************************************************//** -Increments the pool clock by one and returns its new value. Remember that -in the 32 bit version the clock wraps around at 4 billion! -@return new clock value */ -UNIV_INLINE -ulint -buf_pool_clock_tic(void); -/*====================*/ /*********************************************************************//** Gets the current length of the free list of buffer blocks. @return length of the free list */ @@ -1073,16 +1057,10 @@ struct buf_page_struct{ flushed to disk, this tells the flush_type. @see enum buf_flush */ - unsigned accessed:1; /*!< TRUE if the page has been accessed - while in the buffer pool: read-ahead - may read in pages which have not been - accessed yet; a thread is allowed to - read this for heuristic purposes - without holding any mutex or latch */ unsigned io_fix:2; /*!< type of pending I/O operation; also protected by buf_pool_mutex @see enum buf_io_fix */ - unsigned buf_fix_count:24;/*!< count of how manyfold this block + unsigned buf_fix_count:25;/*!< count of how manyfold this block is currently bufferfixed */ /* @} */ #endif /* !UNIV_HOTBACKUP */ @@ -1112,7 +1090,16 @@ struct buf_page_struct{ - BUF_BLOCK_FILE_PAGE: flush_list - BUF_BLOCK_ZIP_DIRTY: flush_list - BUF_BLOCK_ZIP_PAGE: zip_clean - - BUF_BLOCK_ZIP_FREE: zip_free[] */ + - BUF_BLOCK_ZIP_FREE: zip_free[] + + The contents of the list node + is undefined if !in_flush_list + && state == BUF_BLOCK_FILE_PAGE, + or if state is one of + BUF_BLOCK_MEMORY, + BUF_BLOCK_REMOVE_HASH or + BUF_BLOCK_READY_IN_USE. */ + /* resplit for optimistic use */ UT_LIST_NODE_T(buf_page_t) free; UT_LIST_NODE_T(buf_page_t) flush_list; @@ -1155,18 +1142,8 @@ struct buf_page_struct{ debugging */ //#endif /* UNIV_DEBUG */ unsigned old:1; /*!< TRUE if the block is in the old - blocks in the LRU list */ - unsigned LRU_position:31;/*!< value which monotonically - decreases (or may stay - constant if old==TRUE) toward - the end of the LRU list, if - buf_pool->ulint_clock has not - wrapped around: NOTE that this - value can only be used in - heuristic algorithms, because - of the possibility of a - wrap-around! */ - unsigned freed_page_clock:32;/*!< the value of + blocks in buf_pool->LRU_old */ + unsigned freed_page_clock:31;/*!< the value of buf_pool->freed_page_clock when this block was the last time put to the head of the @@ -1174,6 +1151,9 @@ struct buf_page_struct{ to read this for heuristic purposes without holding any mutex or latch */ + unsigned access_time:32; /*!< time of first access, or + 0 if the block was never accessed + in the buffer pool */ /* @} */ # ifdef UNIV_DEBUG_FILE_ACCESSES ibool file_page_was_freed; @@ -1318,6 +1298,31 @@ Compute the hash fold value for blocks in buf_pool->zip_hash. */ #define BUF_POOL_ZIP_FOLD_BPAGE(b) BUF_POOL_ZIP_FOLD((buf_block_t*) (b)) /* @} */ +/** @brief The buffer pool statistics structure. */ +struct buf_pool_stat_struct{ + ulint n_page_gets; /*!< number of page gets performed; + also successful searches through + the adaptive hash index are + counted as page gets; this field + is NOT protected by the buffer + pool mutex */ + ulint n_pages_read; /*!< number read operations */ + ulint n_pages_written;/*!< number write operations */ + ulint n_pages_created;/*!< number of pages created + in the pool with no read */ + ulint n_ra_pages_read;/*!< number of pages read in + as part of read ahead */ + ulint n_ra_pages_evicted;/*!< number of read ahead + pages that are evicted without + being accessed */ + ulint n_pages_made_young; /*!< number of pages made young, in + calls to buf_LRU_make_block_young() */ + ulint n_pages_not_made_young; /*!< number of pages not made + young because the first access + was not long enough ago, in + buf_page_peek_if_too_old() */ +}; + /** @brief The buffer pool structure. NOTE! The definition appears here only for other modules of this @@ -1342,28 +1347,16 @@ struct buf_pool_struct{ ulint n_pend_reads; /*!< number of pending read operations */ ulint n_pend_unzip; /*!< number of pending decompressions */ - time_t last_printout_time; /*!< when buf_print was last time + time_t last_printout_time; + /*!< when buf_print_io was last time called */ - ulint n_pages_read; /*!< number read operations */ - ulint n_pages_written;/*!< number write operations */ - ulint n_pages_created;/*!< number of pages created - in the pool with no read */ - ulint n_page_gets; /*!< number of page gets performed; - also successful searches through - the adaptive hash index are - counted as page gets; this field - is NOT protected by the buffer - pool mutex */ - ulint n_page_gets_old;/*!< n_page_gets when buf_print was - last time called: used to calculate - hit rate */ - ulint n_pages_read_old;/*!< n_pages_read when buf_print was - last time called */ - ulint n_pages_written_old;/*!< number write operations */ - ulint n_pages_created_old;/*!< number of pages created in - the pool with no read */ + buf_pool_stat_t stat; /*!< current statistics */ + buf_pool_stat_t old_stat; /*!< old statistics */ + /* @} */ + /** @name Page flushing algorithm fields */ + /* @{ */ UT_LIST_BASE_NODE_T(buf_page_t) flush_list; @@ -1379,10 +1372,6 @@ struct buf_pool_struct{ /*!< this is in the set state when there is no flush batch of the given type running */ - ulint ulint_clock; /*!< a sequence number used to count - time. NOTE! This counter wraps - around at 4 billion (if ulint == - 32 bits)! */ ulint freed_page_clock;/*!< a sequence number used to count the number of buffer blocks removed from the end of @@ -1406,17 +1395,18 @@ struct buf_pool_struct{ block list */ UT_LIST_BASE_NODE_T(buf_page_t) LRU; /*!< base node of the LRU list */ - buf_page_t* LRU_old; /*!< pointer to the about 3/8 oldest - blocks in the LRU list; NULL if LRU - length less than BUF_LRU_OLD_MIN_LEN; + buf_page_t* LRU_old; /*!< pointer to the about + buf_LRU_old_ratio/BUF_LRU_OLD_RATIO_DIV + oldest blocks in the LRU list; + NULL if LRU length less than + BUF_LRU_OLD_MIN_LEN; NOTE: when LRU_old != NULL, its length should always equal LRU_old_len */ ulint LRU_old_len; /*!< length of the LRU list from the block to which LRU_old points onward, including that block; see buf0lru.c for the restrictions - on this value; not defined if - LRU_old == NULL; + on this value; 0 if LRU_old == NULL; NOTE: LRU_old_len must be adjusted whenever LRU_old shrinks or grows! */ diff --git a/storage/xtradb/include/buf0buf.ic b/storage/xtradb/include/buf0buf.ic index 1a2b2556839..e48794fc753 100644 --- a/storage/xtradb/include/buf0buf.ic +++ b/storage/xtradb/include/buf0buf.ic @@ -72,9 +72,30 @@ buf_page_peek_if_too_old( /*=====================*/ const buf_page_t* bpage) /*!< in: block to make younger */ { - return(buf_pool->freed_page_clock - >= buf_page_get_freed_page_clock(bpage) - + 1 + (buf_pool->curr_size / 4)); + if (UNIV_UNLIKELY(buf_pool->freed_page_clock == 0)) { + /* If eviction has not started yet, do not update the + statistics or move blocks in the LRU list. This is + either the warm-up phase or an in-memory workload. */ + return(FALSE); + } else if (buf_LRU_old_threshold_ms && bpage->old) { + unsigned access_time = buf_page_is_accessed(bpage); + + if (access_time > 0 + && (ut_time_ms() - access_time) + >= buf_LRU_old_threshold_ms) { + return(TRUE); + } + + buf_pool->stat.n_pages_not_made_young++; + return(FALSE); + } else { + /* FIXME: bpage->freed_page_clock is 31 bits */ + return((buf_pool->freed_page_clock & ((1UL << 31) - 1)) + > ((ulint) bpage->freed_page_clock + + (buf_pool->curr_size + * (BUF_LRU_OLD_RATIO_DIV - buf_LRU_old_ratio) + / (BUF_LRU_OLD_RATIO_DIV * 4)))); + } } /*********************************************************************//** @@ -125,23 +146,6 @@ try_again: return(lsn); } - -/*******************************************************************//** -Increments the buf_pool clock by one and returns its new value. Remember -that in the 32 bit version the clock wraps around at 4 billion! -@return new clock value */ -UNIV_INLINE -ulint -buf_pool_clock_tic(void) -/*====================*/ -{ - //ut_ad(buf_pool_mutex_own()); - ut_ad(mutex_own(&LRU_list_mutex)); - - buf_pool->ulint_clock++; - - return(buf_pool->ulint_clock); -} #endif /* !UNIV_HOTBACKUP */ /*********************************************************************//** @@ -288,21 +292,6 @@ buf_page_belongs_to_unzip_LRU( } /*********************************************************************//** -Determine the approximate LRU list position of a block. -@return LRU list position */ -UNIV_INLINE -ulint -buf_page_get_LRU_position( -/*======================*/ - const buf_page_t* bpage) /*!< in: control block */ -{ - ut_ad(buf_page_in_file(bpage)); - //ut_ad(buf_pool_mutex_own()); /* This is used in optimistic */ - - return(bpage->LRU_position); -} - -/*********************************************************************//** Gets the mutex of a block. @return pointer to mutex protecting bpage */ UNIV_INLINE @@ -508,10 +497,19 @@ buf_page_set_old( ut_ad(bpage->in_LRU_list); #ifdef UNIV_LRU_DEBUG - if (UT_LIST_GET_PREV(LRU, bpage) && UT_LIST_GET_NEXT(LRU, bpage) - && UT_LIST_GET_PREV(LRU, bpage)->old - == UT_LIST_GET_NEXT(LRU, bpage)->old) { - ut_a(UT_LIST_GET_PREV(LRU, bpage)->old == old); + ut_a((buf_pool->LRU_old_len == 0) == (buf_pool->LRU_old == NULL)); + /* If a block is flagged "old", the LRU_old list must exist. */ + ut_a(!old || buf_pool->LRU_old); + + if (UT_LIST_GET_PREV(LRU, bpage) && UT_LIST_GET_NEXT(LRU, bpage)) { + const buf_page_t* prev = UT_LIST_GET_PREV(LRU, bpage); + const buf_page_t* next = UT_LIST_GET_NEXT(LRU, bpage); + if (prev->old == next->old) { + ut_a(prev->old == old); + } else { + ut_a(!prev->old); + ut_a(buf_pool->LRU_old == (old ? bpage : next)); + } } #endif /* UNIV_LRU_DEBUG */ @@ -519,17 +517,17 @@ buf_page_set_old( } /*********************************************************************//** -Determine if a block has been accessed in the buffer pool. -@return TRUE if accessed */ +Determine the time of first access of a block in the buffer pool. +@return ut_time_ms() at the time of first access, 0 if not accessed */ UNIV_INLINE -ibool +unsigned buf_page_is_accessed( /*=================*/ const buf_page_t* bpage) /*!< in: control block */ { ut_ad(buf_page_in_file(bpage)); - return(bpage->accessed); + return(bpage->access_time); } /*********************************************************************//** @@ -539,12 +537,16 @@ void buf_page_set_accessed( /*==================*/ buf_page_t* bpage, /*!< in/out: control block */ - ibool accessed) /*!< in: accessed */ + ulint time_ms) /*!< in: ut_time_ms() */ { ut_a(buf_page_in_file(bpage)); + //ut_ad(buf_pool_mutex_own()); ut_ad(mutex_own(buf_page_get_mutex(bpage))); - bpage->accessed = accessed; + if (!bpage->access_time) { + /* Make this the time of the first access. */ + bpage->access_time = time_ms; + } } /*********************************************************************//** @@ -825,15 +827,15 @@ buf_page_get_newest_modification( ib_uint64_t lsn; mutex_t* block_mutex = buf_page_get_mutex_enter(bpage); - ut_a(block_mutex); - - if (buf_page_in_file(bpage)) { + if (block_mutex && buf_page_in_file(bpage)) { lsn = bpage->newest_modification; } else { lsn = 0; } - mutex_exit(block_mutex); + if (block_mutex) { + mutex_exit(block_mutex); + } return(lsn); } diff --git a/storage/xtradb/include/buf0lru.h b/storage/xtradb/include/buf0lru.h index cde88824b64..0a7d01c95cf 100644 --- a/storage/xtradb/include/buf0lru.h +++ b/storage/xtradb/include/buf0lru.h @@ -69,7 +69,7 @@ These are low-level functions #########################################################################*/ /** Minimum LRU list length for which the LRU_old pointer is defined */ -#define BUF_LRU_OLD_MIN_LEN 80 +#define BUF_LRU_OLD_MIN_LEN 512 /* 8 megabytes of 16k pages */ /** Maximum LRU list search length in buf_flush_LRU_recommendation() */ #define BUF_LRU_FREE_SEARCH_LEN (5 + 2 * BUF_READ_AHEAD_AREA) @@ -84,15 +84,6 @@ void buf_LRU_invalidate_tablespace( /*==========================*/ ulint id); /*!< in: space id */ -/******************************************************************//** -Gets the minimum LRU_position field for the blocks in an initial segment -(determined by BUF_LRU_INITIAL_RATIO) of the LRU list. The limit is not -guaranteed to be precise, because the ulint_clock may wrap around. -@return the limit; zero if could not determine it */ -UNIV_INTERN -ulint -buf_LRU_get_recent_limit(void); -/*==========================*/ /********************************************************************//** Insert a compressed block into buf_pool->zip_clean in the LRU order. */ UNIV_INTERN @@ -203,6 +194,18 @@ void buf_LRU_make_block_old( /*===================*/ buf_page_t* bpage); /*!< in: control block */ +/**********************************************************************//** +Updates buf_LRU_old_ratio. +@return updated old_pct */ +UNIV_INTERN +uint +buf_LRU_old_ratio_update( +/*=====================*/ + uint old_pct,/*!< in: Reserve this percentage of + the buffer pool for "old" blocks. */ + ibool adjust);/*!< in: TRUE=adjust the LRU list; + FALSE=just assign buf_LRU_old_ratio + during the initialization of InnoDB */ /********************************************************************//** Update the historical stats that we are collecting for LRU eviction policy at the end of each interval. */ @@ -210,6 +213,18 @@ UNIV_INTERN void buf_LRU_stat_update(void); /*=====================*/ +/********************************************************************//** +Dump the LRU page list to the specific file. */ +UNIV_INTERN +ibool +buf_LRU_file_dump(void); +/*===================*/ +/********************************************************************//** +Read the pages based on the specific file.*/ +UNIV_INTERN +ibool +buf_LRU_file_restore(void); +/*======================*/ #if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG /**********************************************************************//** @@ -229,6 +244,35 @@ buf_LRU_print(void); /*===============*/ #endif /* UNIV_DEBUG_PRINT || UNIV_DEBUG || UNIV_BUF_DEBUG */ +/** @name Heuristics for detecting index scan @{ */ +/** Reserve this much/BUF_LRU_OLD_RATIO_DIV of the buffer pool for +"old" blocks. Protected by buf_pool_mutex. */ +extern uint buf_LRU_old_ratio; +/** The denominator of buf_LRU_old_ratio. */ +#define BUF_LRU_OLD_RATIO_DIV 1024 +/** Maximum value of buf_LRU_old_ratio. +@see buf_LRU_old_adjust_len +@see buf_LRU_old_ratio_update */ +#define BUF_LRU_OLD_RATIO_MAX BUF_LRU_OLD_RATIO_DIV +/** Minimum value of buf_LRU_old_ratio. +@see buf_LRU_old_adjust_len +@see buf_LRU_old_ratio_update +The minimum must exceed +(BUF_LRU_OLD_TOLERANCE + 5) * BUF_LRU_OLD_RATIO_DIV / BUF_LRU_OLD_MIN_LEN. */ +#define BUF_LRU_OLD_RATIO_MIN 51 + +#if BUF_LRU_OLD_RATIO_MIN >= BUF_LRU_OLD_RATIO_MAX +# error "BUF_LRU_OLD_RATIO_MIN >= BUF_LRU_OLD_RATIO_MAX" +#endif +#if BUF_LRU_OLD_RATIO_MAX > BUF_LRU_OLD_RATIO_DIV +# error "BUF_LRU_OLD_RATIO_MAX > BUF_LRU_OLD_RATIO_DIV" +#endif + +/** Move blocks to "new" LRU list only if the first access was at +least this many milliseconds ago. Not protected by any mutex or latch. */ +extern uint buf_LRU_old_threshold_ms; +/* @} */ + /** @brief Statistics for selecting the LRU list for eviction. These statistics are not 'of' LRU but 'for' LRU. We keep count of I/O diff --git a/storage/xtradb/include/buf0rea.h b/storage/xtradb/include/buf0rea.h index b4d25e6fde0..71f62ff7b49 100644 --- a/storage/xtradb/include/buf0rea.h +++ b/storage/xtradb/include/buf0rea.h @@ -27,28 +27,59 @@ Created 11/5/1995 Heikki Tuuri #define buf0rea_h #include "univ.i" +#include "trx0types.h" #include "buf0types.h" /********************************************************************//** +Low-level function which reads a page asynchronously from a file to the +buffer buf_pool if it is not already there, in which case does nothing. +Sets the io_fix flag and sets an exclusive lock on the buffer frame. The +flag is cleared and the x-lock released by an i/o-handler thread. +@return 1 if a read request was queued, 0 if the page already resided +in buf_pool, or if the page is in the doublewrite buffer blocks in +which case it is never read into the pool, or if the tablespace does +not exist or is being dropped +@return 1 if read request is issued. 0 if it is not */ +UNIV_INTERN +ulint +buf_read_page_low( +/*==============*/ + ulint* err, /*!< out: DB_SUCCESS or DB_TABLESPACE_DELETED if we are + trying to read from a non-existent tablespace, or a + tablespace which is just now being dropped */ + ibool sync, /*!< in: TRUE if synchronous aio is desired */ + ulint mode, /*!< in: BUF_READ_IBUF_PAGES_ONLY, ..., + ORed to OS_AIO_SIMULATED_WAKE_LATER (see below + at read-ahead functions) */ + ulint space, /*!< in: space id */ + ulint zip_size,/*!< in: compressed page size, or 0 */ + ibool unzip, /*!< in: TRUE=request uncompressed page */ + ib_int64_t tablespace_version, /*!< in: if the space memory object has + this timestamp different from what we are giving here, + treat the tablespace as dropped; this is a timestamp we + use to stop dangling page reads from a tablespace + which we have DISCARDed + IMPORTed back */ + ulint offset, /*!< in: page number */ + trx_t* trx); +/********************************************************************//** High-level function which reads a page asynchronously from a file to the buffer buf_pool if it is not already there. Sets the io_fix flag and sets an exclusive lock on the buffer frame. The flag is cleared and the x-lock -released by the i/o-handler thread. Does a random read-ahead if it seems -sensible. -@return number of page read requests issued: this can be greater than -1 if read-ahead occurred */ +released by the i/o-handler thread. +@return TRUE if page has been read in, FALSE in case of failure */ UNIV_INTERN -ulint +ibool buf_read_page( /*==========*/ ulint space, /*!< in: space id */ ulint zip_size,/*!< in: compressed page size in bytes, or 0 */ - ulint offset);/*!< in: page number */ + ulint offset, /*!< in: page number */ + trx_t* trx); /********************************************************************//** Applies linear read-ahead if in the buf_pool the page is a border page of a linear read-ahead area and all the pages in the area have been accessed. Does not read any page if the read-ahead mechanism is not activated. Note -that the the algorithm looks at the 'natural' adjacent successor and +that the algorithm looks at the 'natural' adjacent successor and predecessor of the page, which on the leaf level of a B-tree are the next and previous page in the chain of leaves. To know these, the page specified in (space, offset) must already be present in the buf_pool. Thus, the @@ -74,8 +105,9 @@ buf_read_ahead_linear( /*==================*/ ulint space, /*!< in: space id */ ulint zip_size,/*!< in: compressed page size in bytes, or 0 */ - ulint offset);/*!< in: page number of a page; NOTE: the current thread + ulint offset, /*!< in: page number of a page; NOTE: the current thread must want access to this page (see NOTE 3 above) */ + trx_t* trx); /********************************************************************//** Issues read requests for pages which the ibuf module wants to read in, in order to contract the insert buffer tree. Technically, this function is like diff --git a/storage/xtradb/include/buf0types.h b/storage/xtradb/include/buf0types.h index e7167d716a0..bfae6477135 100644 --- a/storage/xtradb/include/buf0types.h +++ b/storage/xtradb/include/buf0types.h @@ -34,6 +34,8 @@ typedef struct buf_block_struct buf_block_t; typedef struct buf_chunk_struct buf_chunk_t; /** Buffer pool comprising buf_chunk_t */ typedef struct buf_pool_struct buf_pool_t; +/** Buffer pool statistics struct */ +typedef struct buf_pool_stat_struct buf_pool_stat_t; /** A buffer frame. @see page_t */ typedef byte buf_frame_t; diff --git a/storage/xtradb/include/db0err.h b/storage/xtradb/include/db0err.h index 23898583b72..747e9b5364e 100644 --- a/storage/xtradb/include/db0err.h +++ b/storage/xtradb/include/db0err.h @@ -32,6 +32,7 @@ enum db_err { /* The following are error codes */ DB_ERROR, + DB_INTERRUPTED, DB_OUT_OF_MEMORY, DB_OUT_OF_FILE_SPACE, DB_LOCK_WAIT, diff --git a/storage/xtradb/include/dict0crea.h b/storage/xtradb/include/dict0crea.h index 3107d771d88..cce1246b789 100644 --- a/storage/xtradb/include/dict0crea.h +++ b/storage/xtradb/include/dict0crea.h @@ -110,7 +110,7 @@ dict_create_or_check_foreign_constraint_tables(void); Adds foreign key definitions to data dictionary tables in the database. We look at table->foreign_list, and also generate names to constraints that were not named by the user. A generated constraint has a name of the format -databasename/tablename_ibfk_<number>, where the numbers start from 1, and are +databasename/tablename_ibfk_NUMBER, where the numbers start from 1, and are given locally for this table, that is, the number is not global, as in the old format constraints < 4.0.18 it used to be. @return error code or DB_SUCCESS */ diff --git a/storage/xtradb/include/dict0dict.h b/storage/xtradb/include/dict0dict.h index 0842e307cda..3d49b6a98a4 100644 --- a/storage/xtradb/include/dict0dict.h +++ b/storage/xtradb/include/dict0dict.h @@ -712,7 +712,7 @@ dict_index_find_on_id_low( dulint id); /*!< in: index id */ /**********************************************************************//** Adds an index to the dictionary cache. -@return DB_SUCCESS or error code */ +@return DB_SUCCESS, DB_TOO_BIG_RECORD, or DB_CORRUPTION */ UNIV_INTERN ulint dict_index_add_to_cache( @@ -1157,6 +1157,13 @@ void dict_ind_init(void); /*===============*/ +/**********************************************************************//** +Closes the data dictionary module. */ +UNIV_INTERN +void +dict_close(void); +/*============*/ + #ifndef UNIV_NONINL #include "dict0dict.ic" #endif diff --git a/storage/xtradb/include/dict0mem.h b/storage/xtradb/include/dict0mem.h index 1ee906fbf57..2d001111938 100644 --- a/storage/xtradb/include/dict0mem.h +++ b/storage/xtradb/include/dict0mem.h @@ -317,7 +317,7 @@ struct dict_foreign_struct{ char* id; /*!< id of the constraint as a null-terminated string */ unsigned n_fields:10; /*!< number of indexes' first fields - for which the the foreign key + for which the foreign key constraint is defined: we allow the indexes to contain more fields than mentioned in the constraint, as long diff --git a/storage/xtradb/include/fil0fil.h b/storage/xtradb/include/fil0fil.h index 0470d533dec..805948f5b5d 100644 --- a/storage/xtradb/include/fil0fil.h +++ b/storage/xtradb/include/fil0fil.h @@ -224,15 +224,6 @@ fil_space_create( 0 for uncompressed tablespaces */ ulint purpose);/*!< in: FIL_TABLESPACE, or FIL_LOG if log */ /*******************************************************************//** -Frees a space object from a the tablespace memory cache. Closes the files in -the chain but does not delete them. -@return TRUE if success */ -UNIV_INTERN -ibool -fil_space_free( -/*===========*/ - ulint id); /*!< in: space id */ -/*******************************************************************//** Returns the size of the space in pages. The tablespace must be cached in the memory cache. @return space size, 0 if space not found */ @@ -278,6 +269,12 @@ fil_init( ulint hash_size, /*!< in: hash table size */ ulint max_n_open); /*!< in: max number of open files */ /*******************************************************************//** +Initializes the tablespace memory cache. */ +UNIV_INTERN +void +fil_close(void); +/*===========*/ +/*******************************************************************//** Opens all log files and system tablespace data files. They stay open until the database server shutdown. This should be called at a server startup after the space objects for the log and the system tablespace have been created. The @@ -614,9 +611,12 @@ fil_space_get_n_reserved_extents( Reads or writes data. This operation is asynchronous (aio). @return DB_SUCCESS, or DB_TABLESPACE_DELETED if we are trying to do i/o on a tablespace which does not exist */ +#define fil_io(type, sync, space_id, zip_size, block_offset, byte_offset, len, buf, message) \ + _fil_io(type, sync, space_id, zip_size, block_offset, byte_offset, len, buf, message, NULL) + UNIV_INTERN ulint -fil_io( +_fil_io( /*===*/ ulint type, /*!< in: OS_FILE_READ or OS_FILE_WRITE, ORed to OS_FILE_LOG, if a log i/o @@ -641,8 +641,25 @@ fil_io( void* buf, /*!< in/out: buffer where to store read data or from where to write; in aio this must be appropriately aligned */ - void* message); /*!< in: message for aio handler if non-sync + void* message, /*!< in: message for aio handler if non-sync aio used, else ignored */ + trx_t* trx); +/********************************************************************//** +Confirm whether the parameters are valid or not */ +UNIV_INTERN +ibool +fil_area_is_exist( +/*==============*/ + ulint space_id, /*!< in: space id */ + ulint zip_size, /*!< in: compressed page size in bytes; + 0 for uncompressed pages */ + ulint block_offset, /*!< in: offset in number of blocks */ + ulint byte_offset, /*!< in: remainder of offset in bytes; in + aio this must be divisible by the OS block + size */ + ulint len); /*!< in: how many bytes to read or write; this + must not cross a file boundary; in aio this + must be a block size multiple */ /**********************************************************************//** Waits for an aio operation to complete. This function is used to write the handler for completed requests. The aio array of pending requests is divided diff --git a/storage/xtradb/include/fsp0fsp.h b/storage/xtradb/include/fsp0fsp.h index 5f7dc58eedc..7abd3914eda 100644 --- a/storage/xtradb/include/fsp0fsp.h +++ b/storage/xtradb/include/fsp0fsp.h @@ -42,7 +42,7 @@ fsp_init(void); /*==========*/ /**********************************************************************//** Gets the current free limit of the system tablespace. The free limit -means the place of the first page which has never been put to the the +means the place of the first page which has never been put to the free list for allocation. The space above that address is initialized to zero. Sets also the global variable log_fsp_current_free_limit. @return free limit in megabytes */ diff --git a/storage/xtradb/include/ibuf0ibuf.h b/storage/xtradb/include/ibuf0ibuf.h index 21330997df3..8aa21fb9d95 100644 --- a/storage/xtradb/include/ibuf0ibuf.h +++ b/storage/xtradb/include/ibuf0ibuf.h @@ -356,6 +356,12 @@ void ibuf_print( /*=======*/ FILE* file); /*!< in: file where to print */ +/******************************************************************//** +Closes insert buffer and frees the data structures. */ +UNIV_INTERN +void +ibuf_close(void); +/*============*/ #define IBUF_HEADER_PAGE_NO FSP_IBUF_HEADER_PAGE_NO #define IBUF_TREE_ROOT_PAGE_NO FSP_IBUF_TREE_ROOT_PAGE_NO diff --git a/storage/xtradb/include/lock0lock.h b/storage/xtradb/include/lock0lock.h index fa5db831d4f..82e4c9bd976 100644 --- a/storage/xtradb/include/lock0lock.h +++ b/storage/xtradb/include/lock0lock.h @@ -59,6 +59,12 @@ lock_sys_create( /*============*/ ulint n_cells); /*!< in: number of slots in lock hash table */ /*********************************************************************//** +Closes the lock system at database shutdown. */ +UNIV_INTERN +void +lock_sys_close(void); +/*================*/ +/*********************************************************************//** Checks if some transaction has an implicit x-lock on a record in a clustered index. @return transaction which has the x-lock, or NULL */ @@ -630,6 +636,14 @@ lock_number_of_rows_locked( /*=======================*/ trx_t* trx); /*!< in: transaction */ /*******************************************************************//** +Check if a transaction holds any autoinc locks. +@return TRUE if the transaction holds any AUTOINC locks. */ +UNIV_INTERN +ibool +lock_trx_holds_autoinc_locks( +/*=========================*/ + const trx_t* trx); /*!< in: transaction */ +/*******************************************************************//** Release all the transaction's autoinc locks. */ UNIV_INTERN void diff --git a/storage/xtradb/include/log0log.h b/storage/xtradb/include/log0log.h index 059f548a085..135aeb69e2d 100644 --- a/storage/xtradb/include/log0log.h +++ b/storage/xtradb/include/log0log.h @@ -118,10 +118,9 @@ UNIV_INLINE ib_uint64_t log_reserve_and_write_fast( /*=======================*/ - byte* str, /*!< in: string */ + const void* str, /*!< in: string */ ulint len, /*!< in: string length */ - ib_uint64_t* start_lsn,/*!< out: start lsn of the log record */ - ibool* success);/*!< out: TRUE if success */ + ib_uint64_t* start_lsn);/*!< out: start lsn of the log record */ /***********************************************************************//** Releases the log mutex. */ UNIV_INLINE @@ -283,7 +282,7 @@ log_make_checkpoint_at( later lsn, if IB_ULONGLONG_MAX, makes a checkpoint at the latest lsn */ ibool write_always); /*!< in: the function normally checks if - the the new checkpoint would have a + the new checkpoint would have a greater lsn than the previous one: if not, then no physical write is done; by setting this parameter TRUE, a @@ -573,6 +572,18 @@ UNIV_INTERN void log_refresh_stats(void); /*===================*/ +/********************************************************** +Shutdown the log system but do not release all the memory. */ +UNIV_INTERN +void +log_shutdown(void); +/*==============*/ +/********************************************************** +Free the log system data structures. */ +UNIV_INTERN +void +log_mem_free(void); +/*==============*/ extern log_t* log_sys; @@ -585,7 +596,7 @@ extern log_t* log_sys; #define LOG_RECOVER 98887331 /* The counting of lsn's starts from this value: this must be non-zero */ -#define LOG_START_LSN ((ib_uint64_t) (16 * OS_FILE_LOG_BLOCK_SIZE)) +#define LOG_START_LSN ((ib_uint64_t) (16 * OS_FILE_LOG_BLOCK_SIZE)) #define LOG_BUFFER_SIZE (srv_log_buffer_size * UNIV_PAGE_SIZE) #define LOG_ARCHIVE_BUF_SIZE (srv_log_buffer_size * UNIV_PAGE_SIZE / 4) @@ -722,9 +733,12 @@ struct log_group_struct{ ulint lsn_offset; /*!< the offset of the above lsn */ ulint n_pending_writes;/*!< number of currently pending flush writes for this log group */ + byte** file_header_bufs_ptr;/*!< unaligned buffers */ byte** file_header_bufs;/*!< buffers for each file header in the group */ +#ifdef UNIV_LOG_ARCHIVE /*-----------------------------*/ + byte** archive_file_header_bufs_ptr;/*!< unaligned buffers */ byte** archive_file_header_bufs;/*!< buffers for each file header in the group */ ulint archive_space_id;/*!< file space which @@ -743,10 +757,12 @@ struct log_group_struct{ completion function then sets the new value to ..._file_no */ ulint next_archived_offset; /*!< like the preceding field */ +#endif /* UNIV_LOG_ARCHIVE */ /*-----------------------------*/ ib_uint64_t scanned_lsn; /*!< used only in recovery: recovery scan succeeded up to this lsn in this log group */ + byte* checkpoint_buf_ptr;/*!< unaligned checkpoint header */ byte* checkpoint_buf; /*!< checkpoint header is written from this buffer to the group */ UT_LIST_NODE_T(log_group_t) @@ -764,6 +780,7 @@ struct log_struct{ #ifndef UNIV_HOTBACKUP mutex_t mutex; /*!< mutex protecting the log */ #endif /* !UNIV_HOTBACKUP */ + byte* buf_ptr; /* unaligned log buffer */ byte* buf; /*!< log buffer */ ulint buf_size; /*!< log buffer size in bytes */ ulint max_buf_free; /*!< recommended maximum value of @@ -900,6 +917,7 @@ struct log_struct{ should wait for this without owning the log mutex */ #endif /* !UNIV_HOTBACKUP */ + byte* checkpoint_buf_ptr;/* unaligned checkpoint header */ byte* checkpoint_buf; /*!< checkpoint header is read to this buffer */ /* @} */ diff --git a/storage/xtradb/include/log0log.ic b/storage/xtradb/include/log0log.ic index d071985982a..36d151a3064 100644 --- a/storage/xtradb/include/log0log.ic +++ b/storage/xtradb/include/log0log.ic @@ -27,6 +27,7 @@ Created 12/9/1995 Heikki Tuuri #include "mach0data.h" #include "mtr0mtr.h" +#ifdef UNIV_LOG_DEBUG /******************************************************//** Checks by parsing that the catenated log segment for a single mtr is consistent. */ @@ -34,11 +35,12 @@ UNIV_INTERN ibool log_check_log_recs( /*===============*/ - byte* buf, /*!< in: pointer to the start of + const byte* buf, /*!< in: pointer to the start of the log segment in the log_sys->buf log buffer */ ulint len, /*!< in: segment length in bytes */ ib_uint64_t buf_start_lsn); /*!< in: buffer start lsn */ +#endif /* UNIV_LOG_DEBUG */ /************************************************************//** Gets a log block flush bit. @@ -305,55 +307,76 @@ UNIV_INLINE ib_uint64_t log_reserve_and_write_fast( /*=======================*/ - byte* str, /*!< in: string */ + const void* str, /*!< in: string */ ulint len, /*!< in: string length */ - ib_uint64_t* start_lsn,/*!< out: start lsn of the log record */ - ibool* success)/*!< out: TRUE if success */ + ib_uint64_t* start_lsn)/*!< out: start lsn of the log record */ { - log_t* log = log_sys; ulint data_len; - ib_uint64_t lsn; +#ifdef UNIV_LOG_LSN_DEBUG + /* length of the LSN pseudo-record */ + ulint lsn_len = 1 + + mach_get_compressed_size(log_sys->lsn >> 32) + + mach_get_compressed_size(log_sys->lsn & 0xFFFFFFFFUL); +#endif /* UNIV_LOG_LSN_DEBUG */ - *success = TRUE; + mutex_enter(&log_sys->mutex); - mutex_enter(&(log->mutex)); - - data_len = len + log->buf_free % OS_FILE_LOG_BLOCK_SIZE; + data_len = len +#ifdef UNIV_LOG_LSN_DEBUG + + lsn_len +#endif /* UNIV_LOG_LSN_DEBUG */ + + log_sys->buf_free % OS_FILE_LOG_BLOCK_SIZE; if (data_len >= OS_FILE_LOG_BLOCK_SIZE - LOG_BLOCK_TRL_SIZE) { /* The string does not fit within the current log block or the log block would become full */ - *success = FALSE; - - mutex_exit(&(log->mutex)); + mutex_exit(&log_sys->mutex); return(0); } - *start_lsn = log->lsn; - - ut_memcpy(log->buf + log->buf_free, str, len); + *start_lsn = log_sys->lsn; + +#ifdef UNIV_LOG_LSN_DEBUG + { + /* Write the LSN pseudo-record. */ + byte* b = &log_sys->buf[log_sys->buf_free]; + *b++ = MLOG_LSN | (MLOG_SINGLE_REC_FLAG & *(const byte*) str); + /* Write the LSN in two parts, + as a pseudo page number and space id. */ + b += mach_write_compressed(b, log_sys->lsn >> 32); + b += mach_write_compressed(b, log_sys->lsn & 0xFFFFFFFFUL); + ut_a(b - lsn_len == &log_sys->buf[log_sys->buf_free]); + + memcpy(b, str, len); + len += lsn_len; + } +#else /* UNIV_LOG_LSN_DEBUG */ + memcpy(log_sys->buf + log_sys->buf_free, str, len); +#endif /* UNIV_LOG_LSN_DEBUG */ - log_block_set_data_len((byte*) ut_align_down(log->buf + log->buf_free, + log_block_set_data_len((byte*) ut_align_down(log_sys->buf + + log_sys->buf_free, OS_FILE_LOG_BLOCK_SIZE), data_len); #ifdef UNIV_LOG_DEBUG - log->old_buf_free = log->buf_free; - log->old_lsn = log->lsn; + log_sys->old_buf_free = log_sys->buf_free; + log_sys->old_lsn = log_sys->lsn; #endif - log->buf_free += len; + log_sys->buf_free += len; - ut_ad(log->buf_free <= log->buf_size); + ut_ad(log_sys->buf_free <= log_sys->buf_size); - lsn = log->lsn += len; + log_sys->lsn += len; #ifdef UNIV_LOG_DEBUG - log_check_log_recs(log->buf + log->old_buf_free, - log->buf_free - log->old_buf_free, log->old_lsn); + log_check_log_recs(log_sys->buf + log_sys->old_buf_free, + log_sys->buf_free - log_sys->old_buf_free, + log_sys->old_lsn); #endif - return(lsn); + return(log_sys->lsn); } /***********************************************************************//** diff --git a/storage/xtradb/include/log0recv.h b/storage/xtradb/include/log0recv.h index 8468c213bdb..ac6b19a3f6a 100644 --- a/storage/xtradb/include/log0recv.h +++ b/storage/xtradb/include/log0recv.h @@ -239,6 +239,18 @@ UNIV_INTERN void recv_sys_create(void); /*=================*/ +/**********************************************************//** +Release recovery system mutexes. */ +UNIV_INTERN +void +recv_sys_close(void); +/*================*/ +/********************************************************//** +Frees the recovery system memory. */ +UNIV_INTERN +void +recv_sys_mem_free(void); +/*===================*/ /********************************************************//** Inits the recovery system for a recovery operation. */ UNIV_INTERN @@ -246,6 +258,12 @@ void recv_sys_init( /*==========*/ ulint available_memory); /*!< in: available memory in bytes */ +/********************************************************//** +Reset the state of the recovery system variables. */ +UNIV_INTERN +void +recv_sys_var_init(void); +/*===================*/ /*******************************************************************//** Empties the hash table of stored log records, applying them to appropriate pages. */ @@ -412,6 +430,39 @@ struct recv_sys_struct{ hash_table_t* addr_hash;/*!< hash table of file addresses of pages */ ulint n_addrs;/*!< number of not processed hashed file addresses in the hash table */ + +/* If you modified the following defines at original file, + You should also modify them. */ +/* defined in os0file.c */ +#define OS_AIO_MERGE_N_CONSECUTIVE 64 +/* defined in log0recv.c */ +#define RECV_READ_AHEAD_AREA 32 + time_t stats_recv_start_time; + ulint stats_recv_turns; + + ulint stats_read_requested_pages; + ulint stats_read_in_area[RECV_READ_AHEAD_AREA]; + + ulint stats_read_io_pages; + ulint stats_read_io_consecutive[OS_AIO_MERGE_N_CONSECUTIVE]; + ulint stats_write_io_pages; + ulint stats_write_io_consecutive[OS_AIO_MERGE_N_CONSECUTIVE]; + + ulint stats_doublewrite_check_pages; + ulint stats_doublewrite_overwrite_pages; + + ulint stats_recover_pages_with_read; + ulint stats_recover_pages_without_read; + + ulint stats_log_recs; + ulint stats_log_len_sum; + + ulint stats_applied_log_recs; + ulint stats_applied_log_len_sum; + ulint stats_pages_already_new; + + ib_uint64_t stats_oldest_modified_lsn; + ib_uint64_t stats_newest_modified_lsn; }; /** The recovery system */ @@ -433,6 +484,11 @@ are allowed yet: the variable name is misleading. */ extern ibool recv_no_ibuf_operations; /** TRUE when recv_init_crash_recovery() has been called. */ extern ibool recv_needed_recovery; +#ifdef UNIV_DEBUG +/** TRUE if writing to the redo log (mtr_commit) is forbidden. +Protected by log_sys->mutex. */ +extern ibool recv_no_log_write; +#endif /* UNIV_DEBUG */ /** TRUE if buf_page_is_corrupted() should check if the log sequence number (FIL_PAGE_LSN) is in the future. Initially FALSE, and set by diff --git a/storage/xtradb/include/mem0mem.h b/storage/xtradb/include/mem0mem.h index a092b024219..98f8748e529 100644 --- a/storage/xtradb/include/mem0mem.h +++ b/storage/xtradb/include/mem0mem.h @@ -82,6 +82,13 @@ void mem_init( /*=====*/ ulint size); /*!< in: common pool size in bytes */ +/******************************************************************//** +Closes the memory system. */ +UNIV_INTERN +void +mem_close(void); +/*===========*/ + /**************************************************************//** Use this macro instead of the corresponding function! Macro for memory heap creation. */ diff --git a/storage/xtradb/include/mem0pool.h b/storage/xtradb/include/mem0pool.h index 18f988241d6..5e93bf88a47 100644 --- a/storage/xtradb/include/mem0pool.h +++ b/storage/xtradb/include/mem0pool.h @@ -62,6 +62,13 @@ mem_pool_create( /*============*/ ulint size); /*!< in: pool size in bytes */ /********************************************************************//** +Frees a memory pool. */ +UNIV_INTERN +void +mem_pool_free( +/*==========*/ + mem_pool_t* pool); /*!< in, own: memory pool */ +/********************************************************************//** Allocates memory from a pool. NOTE: This low-level function should only be used in mem0mem.*! @return own: allocated memory buffer */ diff --git a/storage/xtradb/include/mtr0mtr.h b/storage/xtradb/include/mtr0mtr.h index 69a2c03f4cb..bc3f1951be9 100644 --- a/storage/xtradb/include/mtr0mtr.h +++ b/storage/xtradb/include/mtr0mtr.h @@ -106,6 +106,9 @@ For 1 - 8 bytes, the flag value must give the length also! @{ */ #define MLOG_IBUF_BITMAP_INIT ((byte)27) /*!< initialize an ibuf bitmap page */ /*#define MLOG_FULL_PAGE ((byte)28) full contents of a page */ +#ifdef UNIV_LOG_LSN_DEBUG +# define MLOG_LSN ((byte)28) /* current LSN */ +#endif #define MLOG_INIT_FILE_PAGE ((byte)29) /*!< this means that a file page is taken into use and the prior @@ -118,7 +121,7 @@ For 1 - 8 bytes, the flag value must give the length also! @{ */ #define MLOG_WRITE_STRING ((byte)30) /*!< write a string to a page */ #define MLOG_MULTI_REC_END ((byte)31) /*!< if a single mtr writes - log records for several pages, + several log records, this log record ends the sequence of these records */ #define MLOG_DUMMY_RECORD ((byte)32) /*!< dummy log record used to diff --git a/storage/xtradb/include/os0file.h b/storage/xtradb/include/os0file.h index d8d2f0e5d9e..3eff5216867 100644 --- a/storage/xtradb/include/os0file.h +++ b/storage/xtradb/include/os0file.h @@ -53,6 +53,7 @@ Created 10/21/1995 Heikki Tuuri #define os0file_h #include "univ.i" +#include "trx0types.h" #ifndef __WIN__ #include <dirent.h> @@ -157,6 +158,8 @@ log. */ to become available again */ #define OS_FILE_SHARING_VIOLATION 76 #define OS_FILE_ERROR_NOT_SPECIFIED 77 +#define OS_FILE_INSUFFICIENT_RESOURCE 78 +#define OS_FILE_OPERATION_ABORTED 79 /* @} */ /** Types for aio operations @{ */ @@ -497,9 +500,12 @@ os_file_get_last_error( /*******************************************************************//** Requests a synchronous read operation. @return TRUE if request was successful, FALSE if fail */ +#define os_file_read(file, buf, offset, offset_high, n) \ + _os_file_read(file, buf, offset, offset_high, n, NULL) + UNIV_INTERN ibool -os_file_read( +_os_file_read( /*=========*/ os_file_t file, /*!< in: handle to a file */ void* buf, /*!< in: buffer where to read */ @@ -507,7 +513,8 @@ os_file_read( offset where to read */ ulint offset_high,/*!< in: most significant 32 bits of offset */ - ulint n); /*!< in: number of bytes to read */ + ulint n, /*!< in: number of bytes to read */ + trx_t* trx); /*******************************************************************//** Rewind file to its start, read at most size - 1 bytes from it to str, and NUL-terminate str. All errors are silently ignored. This function is @@ -619,6 +626,13 @@ os_aio_init( ulint n_write_segs, /*<! in: number of writer threads */ ulint n_slots_sync); /*<! in: number of slots in the sync aio array */ +/*********************************************************************** +Frees the asynchronous io system. */ +UNIV_INTERN +void +os_aio_free(void); +/*=============*/ + /*******************************************************************//** Requests an asynchronous i/o operation. @return TRUE if request was queued successfully, FALSE if fail */ @@ -654,10 +668,11 @@ os_aio( (can be used to identify a completed aio operation); ignored if mode is OS_AIO_SYNC */ - void* message2);/*!< in: message for the aio handler + void* message2,/*!< in: message for the aio handler (can be used to identify a completed aio operation); ignored if mode is OS_AIO_SYNC */ + trx_t* trx); /************************************************************************//** Wakes up all async i/o threads so that they know to exit themselves in shutdown. */ diff --git a/storage/xtradb/include/os0sync.h b/storage/xtradb/include/os0sync.h index 0e0b32e7036..0c22162b900 100644 --- a/storage/xtradb/include/os0sync.h +++ b/storage/xtradb/include/os0sync.h @@ -285,44 +285,74 @@ os_fast_mutex_free( /**********************************************************//** Atomic compare-and-swap and increment for InnoDB. */ -#ifdef HAVE_GCC_ATOMIC_BUILTINS +#if defined(HAVE_IB_GCC_ATOMIC_BUILTINS) + +#define HAVE_ATOMIC_BUILTINS + /**********************************************************//** Returns true if swapped, ptr is pointer to target, old_val is value to compare to, new_val is the value to swap in. */ + # define os_compare_and_swap(ptr, old_val, new_val) \ __sync_bool_compare_and_swap(ptr, old_val, new_val) + # define os_compare_and_swap_ulint(ptr, old_val, new_val) \ os_compare_and_swap(ptr, old_val, new_val) + # define os_compare_and_swap_lint(ptr, old_val, new_val) \ os_compare_and_swap(ptr, old_val, new_val) -# define os_compare_and_swap_thread_id(ptr, old_val, new_val) \ + +# ifdef HAVE_IB_ATOMIC_PTHREAD_T_GCC +# define os_compare_and_swap_thread_id(ptr, old_val, new_val) \ os_compare_and_swap(ptr, old_val, new_val) +# define INNODB_RW_LOCKS_USE_ATOMICS +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes and rw_locks use GCC atomic builtins" +# else /* HAVE_IB_ATOMIC_PTHREAD_T_GCC */ +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes use GCC atomic builtins, rw_locks do not" +# endif /* HAVE_IB_ATOMIC_PTHREAD_T_GCC */ + /**********************************************************//** Returns the resulting value, ptr is pointer to target, amount is the amount of increment. */ + # define os_atomic_increment(ptr, amount) \ __sync_add_and_fetch(ptr, amount) + # define os_atomic_increment_lint(ptr, amount) \ os_atomic_increment(ptr, amount) + # define os_atomic_increment_ulint(ptr, amount) \ os_atomic_increment(ptr, amount) + /**********************************************************//** Returns the old value of *ptr, atomically sets *ptr to new_val */ + # define os_atomic_test_and_set_byte(ptr, new_val) \ __sync_lock_test_and_set(ptr, new_val) + +#elif defined(HAVE_IB_SOLARIS_ATOMICS) + +#define HAVE_ATOMIC_BUILTINS + /* If not compiling with GCC or GCC doesn't support the atomic intrinsics and running on Solaris >= 10 use Solaris atomics */ -#elif defined(HAVE_SOLARIS_ATOMICS) + #include <atomic.h> + /**********************************************************//** Returns true if swapped, ptr is pointer to target, old_val is value to compare to, new_val is the value to swap in. */ + # define os_compare_and_swap_ulint(ptr, old_val, new_val) \ (atomic_cas_ulong(ptr, old_val, new_val) == old_val) + # define os_compare_and_swap_lint(ptr, old_val, new_val) \ ((lint)atomic_cas_ulong((ulong_t*) ptr, old_val, new_val) == old_val) -# ifdef INNODB_RW_LOCKS_USE_ATOMICS -# if SIZEOF_PTHREAD_T == 4 + +# ifdef HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS +# if SIZEOF_PTHREAD_T == 4 # define os_compare_and_swap_thread_id(ptr, old_val, new_val) \ ((pthread_t)atomic_cas_32(ptr, old_val, new_val) == old_val) # elif SIZEOF_PTHREAD_T == 8 @@ -331,21 +361,35 @@ compare to, new_val is the value to swap in. */ # else # error "SIZEOF_PTHREAD_T != 4 or 8" # endif /* SIZEOF_PTHREAD_T CHECK */ -# endif /* INNODB_RW_LOCKS_USE_ATOMICS */ +# define INNODB_RW_LOCKS_USE_ATOMICS +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes and rw_locks use Solaris atomic functions" +# else /* HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS */ +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes use Solaris atomic functions, rw_locks do not" +# endif /* HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS */ /**********************************************************//** Returns the resulting value, ptr is pointer to target, amount is the amount of increment. */ + # define os_atomic_increment_lint(ptr, amount) \ atomic_add_long_nv((ulong_t*) ptr, amount) + # define os_atomic_increment_ulint(ptr, amount) \ atomic_add_long_nv(ptr, amount) + /**********************************************************//** Returns the old value of *ptr, atomically sets *ptr to new_val */ + # define os_atomic_test_and_set_byte(ptr, new_val) \ atomic_swap_uchar(ptr, new_val) -/* On Windows, use Windows atomics / interlocked */ + #elif defined(HAVE_WINDOWS_ATOMICS) + +#define HAVE_ATOMIC_BUILTINS + +/* On Windows, use Windows atomics / interlocked */ # ifdef _WIN64 # define win_cmp_and_xchg InterlockedCompareExchange64 # define win_xchg_and_add InterlockedExchangeAdd64 @@ -353,31 +397,46 @@ Returns the old value of *ptr, atomically sets *ptr to new_val */ # define win_cmp_and_xchg InterlockedCompareExchange # define win_xchg_and_add InterlockedExchangeAdd # endif + /**********************************************************//** Returns true if swapped, ptr is pointer to target, old_val is value to compare to, new_val is the value to swap in. */ + # define os_compare_and_swap_ulint(ptr, old_val, new_val) \ (win_cmp_and_xchg(ptr, new_val, old_val) == old_val) + # define os_compare_and_swap_lint(ptr, old_val, new_val) \ (win_cmp_and_xchg(ptr, new_val, old_val) == old_val) -# ifdef INNODB_RW_LOCKS_USE_ATOMICS -# define os_compare_and_swap_thread_id(ptr, old_val, new_val) \ + +/* windows thread objects can always be passed to windows atomic functions */ +# define os_compare_and_swap_thread_id(ptr, old_val, new_val) \ (InterlockedCompareExchange(ptr, new_val, old_val) == old_val) -# endif /* INNODB_RW_LOCKS_USE_ATOMICS */ +# define INNODB_RW_LOCKS_USE_ATOMICS +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes and rw_locks use Windows interlocked functions" + /**********************************************************//** Returns the resulting value, ptr is pointer to target, amount is the amount of increment. */ + # define os_atomic_increment_lint(ptr, amount) \ (win_xchg_and_add(ptr, amount) + amount) + # define os_atomic_increment_ulint(ptr, amount) \ ((ulint) (win_xchg_and_add(ptr, amount) + amount)) + /**********************************************************//** Returns the old value of *ptr, atomically sets *ptr to new_val. InterlockedExchange() operates on LONG, and the LONG will be clobbered */ + # define os_atomic_test_and_set_byte(ptr, new_val) \ ((byte) InterlockedExchange(ptr, new_val)) -#endif /* HAVE_GCC_ATOMIC_BUILTINS */ + +#else +# define IB_ATOMICS_STARTUP_MSG \ + "Mutexes and rw_locks use InnoDB's own implementation" +#endif #ifndef UNIV_NONINL #include "os0sync.ic" diff --git a/storage/xtradb/include/page0page.h b/storage/xtradb/include/page0page.h index a4fe069d022..3899499fb6a 100644 --- a/storage/xtradb/include/page0page.h +++ b/storage/xtradb/include/page0page.h @@ -76,8 +76,11 @@ typedef byte page_header_t; header which are set in a page create */ /*----*/ #define PAGE_LEVEL 26 /* level of the node in an index tree; the - leaf level is the level 0 */ -#define PAGE_INDEX_ID 28 /* index id where the page belongs */ + leaf level is the level 0. This field should + not be written to after page creation. */ +#define PAGE_INDEX_ID 28 /* index id where the page belongs. + This field should not be written to after + page creation. */ #define PAGE_BTR_SEG_LEAF 36 /* file segment header for the leaf pages in a B-tree: defined only on the root page of a B-tree, but not in the root of an ibuf tree */ diff --git a/storage/xtradb/include/page0page.ic b/storage/xtradb/include/page0page.ic index 318ec1cc1f2..8f794410f20 100644 --- a/storage/xtradb/include/page0page.ic +++ b/storage/xtradb/include/page0page.ic @@ -907,7 +907,7 @@ page_get_data_size( /************************************************************//** Allocates a block of memory from the free list of an index page. */ -UNIV_INTERN +UNIV_INLINE void page_mem_alloc_free( /*================*/ diff --git a/storage/xtradb/include/page0zip.h b/storage/xtradb/include/page0zip.h index 9aaa066306b..574809e5227 100644 --- a/storage/xtradb/include/page0zip.h +++ b/storage/xtradb/include/page0zip.h @@ -127,8 +127,12 @@ page_zip_decompress( /*================*/ page_zip_des_t* page_zip,/*!< in: data, ssize; out: m_start, m_end, m_nonempty, n_blobs */ - page_t* page) /*!< out: uncompressed page, may be trashed */ - __attribute__((nonnull)); + page_t* page, /*!< out: uncompressed page, may be trashed */ + ibool all) /*!< in: TRUE=decompress the whole page; + FALSE=verify but do not copy some + page header fields that should not change + after page creation */ + __attribute__((nonnull(1,2))); #ifdef UNIV_DEBUG /**********************************************************************//** @@ -385,8 +389,8 @@ IMPORTANT: if page_zip_reorganize() is invoked on a leaf page of a non-clustered index, the caller must update the insert buffer free bits in the same mini-transaction in such a way that the modification will be redo-logged. -@return TRUE on success, FALSE on failure; page and page_zip will be -left intact on failure. */ +@return TRUE on success, FALSE on failure; page_zip will be left +intact on failure, but page will be overwritten. */ UNIV_INTERN ibool page_zip_reorganize( diff --git a/storage/xtradb/include/pars0pars.h b/storage/xtradb/include/pars0pars.h index a7de7f2292e..fe5d76ebbb0 100644 --- a/storage/xtradb/include/pars0pars.h +++ b/storage/xtradb/include/pars0pars.h @@ -583,6 +583,12 @@ pars_info_get_bound_id( pars_info_t* info, /*!< in: info struct */ const char* name); /*!< in: bound id name to find */ +/******************************************************************//** +Release any resources used by the lexer. */ +UNIV_INTERN +void +pars_lexer_close(void); +/*==================*/ /** Extra information supplied for pars_sql(). */ struct pars_info_struct { diff --git a/storage/xtradb/include/rem0cmp.h b/storage/xtradb/include/rem0cmp.h index 9d2950bbb01..421308af49b 100644 --- a/storage/xtradb/include/rem0cmp.h +++ b/storage/xtradb/include/rem0cmp.h @@ -89,7 +89,7 @@ cmp_dfield_dfield( /*************************************************************//** This function is used to compare a data tuple to a physical record. Only dtuple->n_fields_cmp first fields are taken into account for -the the data tuple! If we denote by n = n_fields_cmp, then rec must +the data tuple! If we denote by n = n_fields_cmp, then rec must have either m >= n fields, or it must differ from dtuple in some of the m fields rec has. If rec has an externally stored field we do not compare it but return with value 0 if such a comparison should be diff --git a/storage/xtradb/include/rem0rec.ic b/storage/xtradb/include/rem0rec.ic index 9fe736f9b0b..8e5bd9a7fcd 100644 --- a/storage/xtradb/include/rem0rec.ic +++ b/storage/xtradb/include/rem0rec.ic @@ -65,7 +65,7 @@ most significant bytes and bits are written below less significant. - offset_of_this_record) mod 64Ki, where mod is the modulo as a non-negative number; - we can calculate the the offset of the next + we can calculate the offset of the next record with the formula: relative_offset + offset_of_this_record mod UNIV_PAGE_SIZE diff --git a/storage/xtradb/include/row0ins.h b/storage/xtradb/include/row0ins.h index 530622e6225..9f93565ddb7 100644 --- a/storage/xtradb/include/row0ins.h +++ b/storage/xtradb/include/row0ins.h @@ -45,7 +45,7 @@ row_ins_check_foreign_constraint( /*=============================*/ ibool check_ref,/*!< in: TRUE If we want to check that the referenced table is ok, FALSE if we - want to to check the foreign key table */ + want to check the foreign key table */ dict_foreign_t* foreign,/*!< in: foreign constraint; NOTE that the tables mentioned in it must be in the dictionary cache if they exist at all */ diff --git a/storage/xtradb/include/row0mysql.h b/storage/xtradb/include/row0mysql.h index 97028622505..b05241f00f8 100644 --- a/storage/xtradb/include/row0mysql.h +++ b/storage/xtradb/include/row0mysql.h @@ -177,7 +177,9 @@ row_update_prebuilt_trx( in MySQL handle */ trx_t* trx); /*!< in: transaction handle */ /*********************************************************************//** -Unlocks AUTO_INC type locks that were possibly reserved by a trx. */ +Unlocks AUTO_INC type locks that were possibly reserved by a trx. This +function should be called at the the end of an SQL statement, by the +connection thread that owns the transaction (trx->mysql_thd). */ UNIV_INTERN void row_unlock_table_autoinc_for_mysql( @@ -754,8 +756,6 @@ struct row_prebuilt_struct { store it here so that we can return it to MySQL */ /*----------------------*/ - UT_LIST_NODE_T(row_prebuilt_t) prebuilts; - /*!< list node of table->prebuilts */ ulint magic_n2; /*!< this should be the same as magic_n */ }; diff --git a/storage/xtradb/include/srv0srv.h b/storage/xtradb/include/srv0srv.h index d98e2ad5013..dc136d89b8c 100644 --- a/storage/xtradb/include/srv0srv.h +++ b/storage/xtradb/include/srv0srv.h @@ -80,6 +80,9 @@ at a time */ #define SRV_AUTO_EXTEND_INCREMENT \ (srv_auto_extend_increment * ((1024 * 1024) / UNIV_PAGE_SIZE)) +/* prototypes for new functions added to ha_innodb.cc */ +ibool innobase_get_slow_log(); + /* This is set to TRUE if the MySQL user has set it in MySQL */ extern ibool srv_lower_case_table_names; @@ -133,8 +136,9 @@ extern ulint* srv_data_file_is_raw_partition; extern ibool srv_extra_undoslots; extern ibool srv_fast_recovery; +extern ibool srv_recovery_stats; -extern ibool srv_use_purge_thread; +extern ulint srv_use_purge_thread; extern ibool srv_auto_extend_last_data_file; extern ulint srv_last_file_size_max; @@ -235,12 +239,14 @@ extern ulong srv_replication_delay; extern long long srv_ibuf_max_size; extern ulong srv_ibuf_active_contract; extern ulong srv_ibuf_accel_rate; +extern ulint srv_checkpoint_age_target; extern ulong srv_flush_neighbor_pages; extern ulong srv_enable_unsafe_group_commit; extern ulong srv_read_ahead; extern ulong srv_adaptive_checkpoint; extern ulong srv_expand_import; +extern ulint srv_relax_table_creation; extern ulong srv_extra_rsegments; extern ulong srv_dict_size_limit; @@ -345,10 +351,6 @@ extern ulint srv_buf_pool_flushed; /** Number of buffer pool reads that led to the reading of a disk page */ extern ulint srv_buf_pool_reads; -/** Number of sequential read-aheads */ -extern ulint srv_read_ahead_seq; -/** Number of random read-aheads */ -extern ulint srv_read_ahead_rnd; /** Status variables to be passed to MySQL */ typedef struct export_var_struct export_struc; @@ -428,6 +430,7 @@ enum srv_thread_type { SRV_INSERT, /**< thread flushing the insert buffer to disk */ #endif SRV_PURGE, /* thread purging undo records */ + SRV_PURGE_WORKER, /* thread purging undo records */ SRV_MASTER /**< the master thread, (whose type number must be biggest) */ }; @@ -446,7 +449,7 @@ void srv_init(void); /*==========*/ /*********************************************************************//** -Frees the OS fast mutex created in srv_boot(). */ +Frees the data structures created in srv_init(). */ UNIV_INTERN void srv_free(void); @@ -509,6 +512,13 @@ srv_purge_thread( /*=============*/ void* arg); /* in: a dummy parameter required by os_thread_create */ +/************************************************************************* +The undo purge thread. */ +UNIV_INTERN +os_thread_ret_t +srv_purge_worker_thread( +/*====================*/ + void* arg); /*******************************************************************//** Tells the Innobase server that there has been activity in the database and wakes up the master thread if it is suspended (not sleeping). Used @@ -645,13 +655,13 @@ struct export_var_struct{ #ifdef UNIV_DEBUG ulint innodb_buffer_pool_pages_latched; /*!< Latched pages */ #endif /* UNIV_DEBUG */ - ulint innodb_buffer_pool_read_requests; /*!< buf_pool->n_page_gets */ + ulint innodb_buffer_pool_read_requests; /*!< buf_pool->stat.n_page_gets */ ulint innodb_buffer_pool_reads; /*!< srv_buf_pool_reads */ ulint innodb_buffer_pool_wait_free; /*!< srv_buf_pool_wait_free */ ulint innodb_buffer_pool_pages_flushed; /*!< srv_buf_pool_flushed */ ulint innodb_buffer_pool_write_requests;/*!< srv_buf_pool_write_requests */ - ulint innodb_buffer_pool_read_ahead_seq;/*!< srv_read_ahead_seq */ - ulint innodb_buffer_pool_read_ahead_rnd;/*!< srv_read_ahead_rnd */ + ulint innodb_buffer_pool_read_ahead; /*!< srv_read_ahead */ + ulint innodb_buffer_pool_read_ahead_evicted;/*!< srv_read_ahead evicted*/ ulint innodb_dblwr_pages_written; /*!< srv_dblwr_pages_written */ ulint innodb_dblwr_writes; /*!< srv_dblwr_writes */ ibool innodb_have_atomic_builtins; /*!< HAVE_ATOMIC_BUILTINS */ @@ -663,9 +673,9 @@ struct export_var_struct{ ulint innodb_os_log_pending_writes; /*!< srv_os_log_pending_writes */ ulint innodb_os_log_pending_fsyncs; /*!< fil_n_pending_log_flushes */ ulint innodb_page_size; /*!< UNIV_PAGE_SIZE */ - ulint innodb_pages_created; /*!< buf_pool->n_pages_created */ - ulint innodb_pages_read; /*!< buf_pool->n_pages_read */ - ulint innodb_pages_written; /*!< buf_pool->n_pages_written */ + ulint innodb_pages_created; /*!< buf_pool->stat.n_pages_created */ + ulint innodb_pages_read; /*!< buf_pool->stat.n_pages_read */ + ulint innodb_pages_written; /*!< buf_pool->stat.n_pages_written */ ulint innodb_row_lock_waits; /*!< srv_n_lock_wait_count */ ulint innodb_row_lock_current_waits; /*!< srv_n_lock_wait_current_count */ ib_int64_t innodb_row_lock_time; /*!< srv_n_lock_wait_time diff --git a/storage/xtradb/include/sync0rw.h b/storage/xtradb/include/sync0rw.h index aedfd5f3f86..85fa014d77a 100644 --- a/storage/xtradb/include/sync0rw.h +++ b/storage/xtradb/include/sync0rw.h @@ -120,7 +120,7 @@ is necessary only if the memory block containing it is freed. */ # endif /* UNIV_SYNC_DEBUG */ #else /* UNIV_DEBUG */ # define rw_lock_create(L, level) \ - rw_lock_create_func((L), __FILE__, __LINE__) + rw_lock_create_func((L), #L, NULL, 0) #endif /* UNIV_DEBUG */ /******************************************************************//** @@ -137,8 +137,8 @@ rw_lock_create_func( # ifdef UNIV_SYNC_DEBUG ulint level, /*!< in: level */ # endif /* UNIV_SYNC_DEBUG */ - const char* cmutex_name, /*!< in: mutex name */ #endif /* UNIV_DEBUG */ + const char* cmutex_name, /*!< in: mutex name */ const char* cfile_name, /*!< in: file name where created */ ulint cline); /*!< in: file line where created */ /******************************************************************//** @@ -540,7 +540,8 @@ struct rw_lock_struct { ulint level; /*!< Level in the global latching order. */ #endif /* UNIV_SYNC_DEBUG */ ulint count_os_wait; /*!< Count of os_waits. May not be accurate */ - const char* cfile_name;/*!< File name where lock created */ + //const char* cfile_name;/*!< File name where lock created */ + const char* lock_name;/*!< lock name */ /* last s-lock file/line is not guaranteed to be correct */ const char* last_s_file_name;/*!< File name where last s-locked */ const char* last_x_file_name;/*!< File name where last x-locked */ @@ -551,7 +552,7 @@ struct rw_lock_struct { are at the start of this struct, thus we can peek this field without causing much memory bus traffic */ - unsigned cline:14; /*!< Line where created */ + //unsigned cline:14; /*!< Line where created */ unsigned last_s_line:14; /*!< Line number where last time s-locked */ unsigned last_x_line:14; /*!< Line number where last time x-locked */ ulint magic_n; /*!< RW_LOCK_MAGIC_N */ diff --git a/storage/xtradb/include/sync0sync.h b/storage/xtradb/include/sync0sync.h index 966bcced722..c653e44b5bd 100644 --- a/storage/xtradb/include/sync0sync.h +++ b/storage/xtradb/include/sync0sync.h @@ -80,7 +80,7 @@ necessary only if the memory block containing it is freed. */ # endif #else # define mutex_create(M, level) \ - mutex_create_func((M), __FILE__, __LINE__) + mutex_create_func((M), #M, NULL, 0) #endif /******************************************************************//** @@ -93,8 +93,8 @@ void mutex_create_func( /*==============*/ mutex_t* mutex, /*!< in: pointer to memory */ -#ifdef UNIV_DEBUG const char* cmutex_name, /*!< in: mutex name */ +#ifdef UNIV_DEBUG # ifdef UNIV_SYNC_DEBUG ulint level, /*!< in: level */ # endif /* UNIV_SYNC_DEBUG */ @@ -513,7 +513,7 @@ struct mutex_struct { os_fast_mutex; /*!< We use this OS mutex in place of lock_word when atomic operations are not enabled */ #endif - ulint waiters; /*!< This ulint is set to 1 if there are (or + volatile ulint waiters; /*!< This ulint is set to 1 if there are (or may be) threads waiting in the global wait array for this mutex to be released. Otherwise, this is 0. */ @@ -524,9 +524,9 @@ struct mutex_struct { ulint line; /*!< Line where the mutex was locked */ ulint level; /*!< Level in the global latching order */ #endif /* UNIV_SYNC_DEBUG */ +#ifdef UNIV_DEBUG const char* cfile_name;/*!< File name where mutex created */ ulint cline; /*!< Line where created */ -#ifdef UNIV_DEBUG os_thread_id_t thread_id; /*!< The thread id of the thread which locked the mutex. */ ulint magic_n; /*!< MUTEX_MAGIC_N */ @@ -541,9 +541,9 @@ struct mutex_struct { ulong count_os_yield; /*!< count of os_wait */ ulonglong lspent_time; /*!< mutex os_wait timer msec */ ulonglong lmax_spent_time;/*!< mutex os_wait timer msec */ - const char* cmutex_name; /*!< mutex name */ ulint mutex_type; /*!< 0=usual mutex, 1=rw_lock mutex */ #endif /* UNIV_DEBUG */ + const char* cmutex_name; /*!< mutex name */ }; /** The global array of wait cells for implementation of the databases own diff --git a/storage/xtradb/include/thr0loc.h b/storage/xtradb/include/thr0loc.h index 34c5232238d..293d1ebd57f 100644 --- a/storage/xtradb/include/thr0loc.h +++ b/storage/xtradb/include/thr0loc.h @@ -39,6 +39,12 @@ UNIV_INTERN void thr_local_init(void); /*================*/ + /****************************************************************//** +Close the thread local storage module. */ +UNIV_INTERN +void +thr_local_close(void); +/*=================*/ /*******************************************************************//** Creates a local storage struct for the calling new thread. */ UNIV_INTERN diff --git a/storage/xtradb/include/trx0i_s.h b/storage/xtradb/include/trx0i_s.h index 9bf032de9f9..7bd4e1b88c8 100644 --- a/storage/xtradb/include/trx0i_s.h +++ b/storage/xtradb/include/trx0i_s.h @@ -141,6 +141,13 @@ void trx_i_s_cache_init( /*===============*/ trx_i_s_cache_t* cache); /*!< out: cache to init */ +/*******************************************************************//** +Free the INFORMATION SCHEMA trx related cache. */ +UNIV_INTERN +void +trx_i_s_cache_free( +/*===============*/ + trx_i_s_cache_t* cache); /*!< in/out: cache to free */ /*******************************************************************//** Issue a shared/read lock on the tables cache. */ diff --git a/storage/xtradb/include/trx0purge.h b/storage/xtradb/include/trx0purge.h index 7812ad7eb92..ae5bc6f90be 100644 --- a/storage/xtradb/include/trx0purge.h +++ b/storage/xtradb/include/trx0purge.h @@ -71,6 +71,12 @@ void trx_purge_sys_create(void); /*======================*/ /********************************************************************//** +Frees the global purge system control structure. */ +UNIV_INTERN +void +trx_purge_sys_close(void); +/*======================*/ +/************************************************************************ Adds the update undo log as the first log in the history list. Removes the update undo log segment from the rseg slot if it is too big for reuse. */ UNIV_INTERN @@ -108,6 +114,25 @@ UNIV_INTERN ulint trx_purge(void); /*===========*/ +/********************************************************************** +This function runs a purge worker batch */ +UNIV_INTERN +void +trx_purge_worker( +/*=============*/ + ulint worker_id); +/********************************************************************** +This function waits the event for worker batch */ +UNIV_INTERN +void +trx_purge_worker_wait(void); +/*========================*/ +/********************************************************************** +This function wakes the waiting worker batch */ +UNIV_INTERN +void +trx_purge_worker_wake(void); +/*========================*/ /******************************************************************//** Prints information of the purge system to stderr. */ UNIV_INTERN @@ -125,6 +150,11 @@ struct trx_purge_struct{ of the trx system and it never ends */ que_t* query; /*!< The query graph which will do the parallelized purge operation */ + ulint n_worker; + os_event_t worker_event; + sess_t** sess_arr; + trx_t** trx_arr; + que_t** query_arr; rw_lock_t latch; /*!< The latch protecting the purge view. A purge operation must acquire an x-latch here for the instant at which diff --git a/storage/xtradb/include/trx0rec.h b/storage/xtradb/include/trx0rec.h index 0ae82c33afe..a6e56e963c6 100644 --- a/storage/xtradb/include/trx0rec.h +++ b/storage/xtradb/include/trx0rec.h @@ -44,8 +44,8 @@ UNIV_INLINE trx_undo_rec_t* trx_undo_rec_copy( /*==============*/ - trx_undo_rec_t* undo_rec, /*!< in: undo log record */ - mem_heap_t* heap); /*!< in: heap where copied */ + const trx_undo_rec_t* undo_rec, /*!< in: undo log record */ + mem_heap_t* heap); /*!< in: heap where copied */ /**********************************************************************//** Reads the undo log record type. @return record type */ diff --git a/storage/xtradb/include/trx0rec.ic b/storage/xtradb/include/trx0rec.ic index 037b5d4f6cf..e7e41d6d9f6 100644 --- a/storage/xtradb/include/trx0rec.ic +++ b/storage/xtradb/include/trx0rec.ic @@ -100,8 +100,8 @@ UNIV_INLINE trx_undo_rec_t* trx_undo_rec_copy( /*==============*/ - trx_undo_rec_t* undo_rec, /*!< in: undo log record */ - mem_heap_t* heap) /*!< in: heap where copied */ + const trx_undo_rec_t* undo_rec, /*!< in: undo log record */ + mem_heap_t* heap) /*!< in: heap where copied */ { ulint len; diff --git a/storage/xtradb/include/trx0roll.h b/storage/xtradb/include/trx0roll.h index ddca9e9e4ef..1dee5655c8c 100644 --- a/storage/xtradb/include/trx0roll.h +++ b/storage/xtradb/include/trx0roll.h @@ -133,6 +133,17 @@ trx_rollback( Rollback or clean up any incomplete transactions which were encountered in crash recovery. If the transaction already was committed, then we clean up a possible insert undo log. If the +transaction was not yet committed, then we roll it back. */ +UNIV_INTERN +void +trx_rollback_or_clean_recovered( +/*============================*/ + ibool all); /*!< in: FALSE=roll back dictionary transactions; + TRUE=roll back all non-PREPARED transactions */ +/*******************************************************************//** +Rollback or clean up any incomplete transactions which were +encountered in crash recovery. If the transaction already was +committed, then we clean up a possible insert undo log. If the transaction was not yet committed, then we roll it back. Note: this is done in a background thread. @return a dummy parameter */ @@ -208,9 +219,9 @@ int trx_general_rollback_for_mysql( /*===========================*/ trx_t* trx, /*!< in: transaction handle */ - ibool partial,/*!< in: TRUE if partial rollback requested */ trx_savept_t* savept);/*!< in: pointer to savepoint undo number, if - partial rollback requested */ + partial rollback requested, or NULL for + complete rollback */ /*******************************************************************//** Rolls back a transaction back to a named savepoint. Modifications after the savepoint are undone but InnoDB does NOT release the corresponding locks diff --git a/storage/xtradb/include/trx0rseg.h b/storage/xtradb/include/trx0rseg.h index f0f7a47279e..0d7dc60329f 100644 --- a/storage/xtradb/include/trx0rseg.h +++ b/storage/xtradb/include/trx0rseg.h @@ -125,6 +125,13 @@ trx_rseg_create( ulint max_size, /*!< in: max size in pages */ ulint* id, /*!< out: rseg id */ mtr_t* mtr); /*!< in: mtr */ +/*************************************************************************** +Free's an instance of the rollback segment in memory. */ +UNIV_INTERN +void +trx_rseg_mem_free( +/*==============*/ + trx_rseg_t* rseg); /* in, own: instance to free */ /* Real max value may be 4076 in usual. But reserve 4 slot for safety or etc... */ diff --git a/storage/xtradb/include/trx0sys.h b/storage/xtradb/include/trx0sys.h index 1f3f11926f8..8789d575bca 100644 --- a/storage/xtradb/include/trx0sys.h +++ b/storage/xtradb/include/trx0sys.h @@ -344,6 +344,12 @@ void trx_sys_file_format_tag_init(void); /*==============================*/ /*****************************************************************//** +Shutdown/Close the transaction system. */ +UNIV_INTERN +void +trx_sys_close(void); +/*===============*/ +/*****************************************************************//** Get the name representation of the file format from its id. @return pointer to the name */ UNIV_INTERN diff --git a/storage/xtradb/include/trx0sys.ic b/storage/xtradb/include/trx0sys.ic index 1c7c732751b..820d31d0692 100644 --- a/storage/xtradb/include/trx0sys.ic +++ b/storage/xtradb/include/trx0sys.ic @@ -34,11 +34,11 @@ typedef byte trx_sysf_rseg_t; /* Rollback segment specification slot offsets */ /*-------------------------------------------------------------*/ -#define TRX_SYS_RSEG_SPACE 0 /* space where the the segment +#define TRX_SYS_RSEG_SPACE 0 /* space where the segment header is placed; starting with MySQL/InnoDB 5.1.7, this is UNIV_UNDEFINED if the slot is unused */ -#define TRX_SYS_RSEG_PAGE_NO 4 /* page number where the the segment +#define TRX_SYS_RSEG_PAGE_NO 4 /* page number where the segment header is placed; this is FIL_NULL if the slot is unused */ /*-------------------------------------------------------------*/ diff --git a/storage/xtradb/include/trx0trx.h b/storage/xtradb/include/trx0trx.h index da031d9d5d3..3b845e498d0 100644 --- a/storage/xtradb/include/trx0trx.h +++ b/storage/xtradb/include/trx0trx.h @@ -179,7 +179,7 @@ trx_commit_off_kernel( /****************************************************************//** Cleans up a transaction at database startup. The cleanup is needed if the transaction already got to the middle of a commit when the database -crashed, andf we cannot roll it back. */ +crashed, and we cannot roll it back. */ UNIV_INTERN void trx_cleanup_at_db_startup( @@ -360,7 +360,7 @@ enum trx_dict_op { operation modes in crash recovery. */ TRX_DICT_OP_TABLE = 1, /** The transaction is creating or dropping an index in an - existing table. In crash recovery, the the data dictionary + existing table. In crash recovery, the data dictionary must be locked, but the table must not be dropped. */ TRX_DICT_OP_INDEX = 2 }; @@ -729,6 +729,17 @@ struct trx_struct{ /*------------------------------*/ char detailed_error[256]; /*!< detailed error message for last error, or empty. */ + /*------------------------------*/ + ulint io_reads; + ib_uint64_t io_read; + ulint io_reads_wait_timer; + ib_uint64_t lock_que_wait_ustarted; + ulint lock_que_wait_timer; + ulint innodb_que_wait_timer; + ulint distinct_page_access; +#define DPAH_SIZE 8192 + byte* distinct_page_access_hash; + ibool take_stats; }; #define TRX_MAX_N_THREADS 32 /* maximum number of diff --git a/storage/xtradb/include/trx0types.h b/storage/xtradb/include/trx0types.h index 08cc9622d02..24cf57d53d5 100644 --- a/storage/xtradb/include/trx0types.h +++ b/storage/xtradb/include/trx0types.h @@ -70,7 +70,7 @@ typedef struct trx_named_savept_struct trx_named_savept_t; enum trx_rb_ctx { RB_NONE = 0, /*!< no rollback */ RB_NORMAL, /*!< normal rollback */ - RB_RECOVERY, /*!< rolling back an incomplete transaction, + RB_RECOVERY /*!< rolling back an incomplete transaction, in crash recovery */ }; diff --git a/storage/xtradb/include/trx0undo.h b/storage/xtradb/include/trx0undo.h index 4db10eaa92e..a084f2394b5 100644 --- a/storage/xtradb/include/trx0undo.h +++ b/storage/xtradb/include/trx0undo.h @@ -333,6 +333,13 @@ trx_undo_parse_discard_latest( byte* end_ptr,/*!< in: buffer end */ page_t* page, /*!< in: page or NULL */ mtr_t* mtr); /*!< in: mtr or NULL */ +/************************************************************************ +Frees an undo log memory copy. */ +UNIV_INTERN +void +trx_undo_mem_free( +/*==============*/ + trx_undo_t* undo); /* in: the undo object to be freed */ /* Types of an undo log segment */ #define TRX_UNDO_INSERT 1 /* contains undo entries for inserts */ diff --git a/storage/xtradb/include/univ.i b/storage/xtradb/include/univ.i index 5ccb7c39551..26097941c5f 100644 --- a/storage/xtradb/include/univ.i +++ b/storage/xtradb/include/univ.i @@ -46,12 +46,12 @@ Created 1/20/1994 Heikki Tuuri #define INNODB_VERSION_MAJOR 1 #define INNODB_VERSION_MINOR 0 -#define INNODB_VERSION_BUGFIX 4 -#define PERCONA_INNODB_VERSION 8 +#define INNODB_VERSION_BUGFIX 6 +#define PERCONA_INNODB_VERSION 9 /* The following is the InnoDB version as shown in SELECT plugin_version FROM information_schema.plugins; -calculated in in make_version_string() in sql/sql_show.cc like this: +calculated in make_version_string() in sql/sql_show.cc like this: "version >> 8" . "version & 0xff" because the version is shown with only one dot, we skip the last component, i.e. we show M.N.P as M.N */ @@ -59,13 +59,14 @@ component, i.e. we show M.N.P as M.N */ (INNODB_VERSION_MAJOR << 8 | INNODB_VERSION_MINOR) /* auxiliary macros to help creating the version as string */ -#define __INNODB_VERSION(a, b, c, d) (#a "." #b "." #c "-" #d) -#define _INNODB_VERSION(a, b, c, d) __INNODB_VERSION(a, b, c, d) +#define __INNODB_VERSION(a, b, c, d) (#a "." #b "." #c "-" #d) +#define _INNODB_VERSION(a, b, c, d) __INNODB_VERSION(a, b, c, d) + #define INNODB_VERSION_STR \ _INNODB_VERSION(INNODB_VERSION_MAJOR, \ INNODB_VERSION_MINOR, \ - INNODB_VERSION_BUGFIX, \ + INNODB_VERSION_BUGFIX, \ PERCONA_INNODB_VERSION) #define REFMAN "http://dev.mysql.com/doc/refman/5.1/en/" @@ -80,17 +81,25 @@ the virtual method table (vtable) in GCC 3. */ # define ha_innobase ha_innodb #endif /* MYSQL_DYNAMIC_PLUGIN */ +/* if any of the following macros is defined at this point this means +that the code from the "right" plug.in was executed and we do not +need to include ut0auxconf.h which would either define the same macros +or will be empty */ +#if !defined(HAVE_IB_GCC_ATOMIC_BUILTINS) \ + && !defined(HAVE_IB_ATOMIC_PTHREAD_T_GCC) \ + && !defined(HAVE_IB_SOLARIS_ATOMICS) \ + && !defined(HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS) \ + && !defined(SIZEOF_PTHREAD_T) \ + && !defined(HAVE_IB_PAUSE_INSTRUCTION) +# include "ut0auxconf.h" +#endif + #if (defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)) && !defined(MYSQL_SERVER) && !defined(__WIN__) # undef __WIN__ # define __WIN__ # include <windows.h> -# if defined(HAVE_WINDOWS_ATOMICS) -/* If atomics are defined we use them in InnoDB mutex implementation */ -# define HAVE_ATOMIC_BUILTINS -# endif /* HAVE_WINDOWS_ATOMICS */ - # ifdef _NT_ # define __NT__ # endif @@ -113,45 +122,17 @@ if we are compiling on Windows. */ # include <sys/mman.h> /* mmap() for os0proc.c */ # endif -# undef PACKAGE -# undef VERSION - /* Include the header file generated by GNU autoconf */ # ifndef __WIN__ -#ifndef UNIV_HOTBACKUP -# include "config.h" -#endif /* UNIV_HOTBACKUP */ +# ifndef UNIV_HOTBACKUP +# include "config.h" +# endif /* UNIV_HOTBACKUP */ # endif # ifdef HAVE_SCHED_H # include <sched.h> # endif -# if defined(HAVE_GCC_ATOMIC_BUILTINS) || defined(HAVE_SOLARIS_ATOMICS) \ - || defined(HAVE_WINDOWS_ATOMICS) -/* If atomics are defined we use them in InnoDB mutex implementation */ -# define HAVE_ATOMIC_BUILTINS -# endif /* (HAVE_GCC_ATOMIC_BUILTINS) || (HAVE_SOLARIS_ATOMICS) - || (HAVE_WINDOWS_ATOMICS) */ - -/* For InnoDB rw_locks to work with atomics we need the thread_id -to be no more than machine word wide. The following enables using -atomics for InnoDB rw_locks where these conditions are met. */ -#ifdef HAVE_ATOMIC_BUILTINS -/* if HAVE_ATOMIC_PTHREAD_T is defined at this point that means that -the code from plug.in has defined it and we do not need to include -ut0auxconf.h which would either define HAVE_ATOMIC_PTHREAD_T or will -be empty */ -# ifndef HAVE_ATOMIC_PTHREAD_T -# include "ut0auxconf.h" -# endif /* HAVE_ATOMIC_PTHREAD_T */ -/* now HAVE_ATOMIC_PTHREAD_T is eventually defined either by plug.in or -from Makefile.in->ut0auxconf.h */ -# ifdef HAVE_ATOMIC_PTHREAD_T -# define INNODB_RW_LOCKS_USE_ATOMICS -# endif /* HAVE_ATOMIC_PTHREAD_T */ -#endif /* HAVE_ATOMIC_BUILTINS */ - /* We only try to do explicit inlining of functions with gcc and Sun Studio */ @@ -198,12 +179,18 @@ command. Not tested on Windows. */ debugging without UNIV_DEBUG */ #define UNIV_DEBUG /* Enable ut_ad() assertions and disable UNIV_INLINE */ +#define UNIV_DEBUG_LOCK_VALIDATE /* Enable + ut_ad(lock_rec_validate_page()) + assertions. */ #define UNIV_DEBUG_FILE_ACCESSES /* Debug .ibd file access (field file_page_was_freed in buf_page_t) */ #define UNIV_LRU_DEBUG /* debug the buffer pool LRU */ #define UNIV_HASH_DEBUG /* debug HASH_ macros */ #define UNIV_LIST_DEBUG /* debug UT_LIST_ macros */ +#define UNIV_LOG_LSN_DEBUG /* write LSN to the redo log; +this will break redo log file compatibility, but it may be useful when +debugging redo log application problems. */ #define UNIV_MEM_DEBUG /* detect memory leaks etc */ #define UNIV_IBUF_DEBUG /* debug the insert buffer */ #define UNIV_IBUF_COUNT_DEBUG /* debug the insert buffer; @@ -253,7 +240,7 @@ by one. */ /* Linkage specifier for non-static InnoDB symbols (variables and functions) that are only referenced from within InnoDB, not from MySQL */ -#if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(UNIV_HOTBACKUP) +#if defined(__GNUC__) && (__GNUC__ >= 4) || defined(__INTEL_COMPILER) # define UNIV_INTERN __attribute__((visibility ("hidden"))) #else # define UNIV_INTERN @@ -410,7 +397,9 @@ it is read. */ /* Minimize cache-miss latency by moving data at addr into a cache before it is read or written. */ # define UNIV_PREFETCH_RW(addr) __builtin_prefetch(addr, 1, 3) -#elif defined(__SUNPRO_C) || defined(__SUNPRO_CC) +/* Sun Studio includes sun_prefetch.h as of version 5.9 */ +#elif (defined(__SUNPRO_C) && __SUNPRO_C >= 0x590) \ + || (defined(__SUNPRO_CC) && __SUNPRO_CC >= 0x590) # include <sun_prefetch.h> #if __SUNPRO_C >= 0x550 # undef UNIV_INTERN diff --git a/storage/xtradb/include/usr0sess.h b/storage/xtradb/include/usr0sess.h index 7638a0c69e2..2c288f7d455 100644 --- a/storage/xtradb/include/usr0sess.h +++ b/storage/xtradb/include/usr0sess.h @@ -44,14 +44,12 @@ sess_t* sess_open(void); /*============*/ /*********************************************************************//** -Closes a session, freeing the memory occupied by it, if it is in a state -where it should be closed. -@return TRUE if closed */ +Closes a session, freeing the memory occupied by it. */ UNIV_INTERN -ibool -sess_try_close( -/*===========*/ - sess_t* sess); /*!< in, own: session object */ +void +sess_close( +/*=======*/ + sess_t* sess); /* in, own: session object */ /* The session handle. All fields are protected by the kernel mutex */ struct sess_struct{ diff --git a/storage/xtradb/include/ut0auxconf.h b/storage/xtradb/include/ut0auxconf.h index 88fb26f1863..16bcc308392 100644 --- a/storage/xtradb/include/ut0auxconf.h +++ b/storage/xtradb/include/ut0auxconf.h @@ -1,14 +1,14 @@ /* Do not remove this file even though it is empty. This file is included in univ.i and will cause compilation failure if not present. -A custom check has been added in the generated +A custom checks have been added in the generated storage/innobase/Makefile.in that is shipped with the InnoDB Plugin -source archive. This check tries to compile a test program and if -successful then adds "#define HAVE_ATOMIC_PTHREAD_T" to this file. -This is a hack that has been developed in order to check for pthread_t -atomicity without the need to regenerate the ./configure script that is +source archive. These checks eventually define some macros and put +them in this file. +This is a hack that has been developed in order to deploy new compile +time checks without the need to regenerate the ./configure script that is distributed in the MySQL 5.1 official source archives. If by any chance Makefile.in and ./configure are regenerated and thus -the hack from Makefile.in wiped away then the "real" check from plug.in +the hack from Makefile.in wiped away then the "real" checks from plug.in will take over. */ diff --git a/storage/xtradb/include/ut0byte.h b/storage/xtradb/include/ut0byte.h index a2687e62f08..f55e2888c60 100644 --- a/storage/xtradb/include/ut0byte.h +++ b/storage/xtradb/include/ut0byte.h @@ -219,8 +219,8 @@ UNIV_INLINE void* ut_align( /*=====*/ - void* ptr, /*!< in: pointer */ - ulint align_no); /*!< in: align by this number */ + const void* ptr, /*!< in: pointer */ + ulint align_no); /*!< in: align by this number */ /*********************************************************//** The following function rounds down a pointer to the nearest aligned address. diff --git a/storage/xtradb/include/ut0byte.ic b/storage/xtradb/include/ut0byte.ic index e3beed65138..3dd51890cb4 100644 --- a/storage/xtradb/include/ut0byte.ic +++ b/storage/xtradb/include/ut0byte.ic @@ -319,8 +319,8 @@ UNIV_INLINE void* ut_align( /*=====*/ - void* ptr, /*!< in: pointer */ - ulint align_no) /*!< in: align by this number */ + const void* ptr, /*!< in: pointer */ + ulint align_no) /*!< in: align by this number */ { ut_ad(align_no > 0); ut_ad(((align_no - 1) & align_no) == 0); diff --git a/storage/xtradb/include/ut0ut.h b/storage/xtradb/include/ut0ut.h index 80094321041..197b8401428 100644 --- a/storage/xtradb/include/ut0ut.h +++ b/storage/xtradb/include/ut0ut.h @@ -34,6 +34,11 @@ Created 1/20/1994 Heikki Tuuri #define ut0ut_h #include "univ.i" + +#ifndef UNIV_HOTBACKUP +# include "os0sync.h" /* for HAVE_ATOMIC_BUILTINS */ +#endif /* UNIV_HOTBACKUP */ + #include <time.h> #ifndef MYSQL_SERVER #include <ctype.h> @@ -47,7 +52,8 @@ Created 1/20/1994 Heikki Tuuri /** Time stamp */ typedef time_t ib_time_t; -#if defined(IB_HAVE_PAUSE_INSTRUCTION) +#ifndef UNIV_HOTBACKUP +#if defined(HAVE_IB_PAUSE_INSTRUCTION) # ifdef WIN32 /* In the Win32 API, the x86 PAUSE instruction is executed by calling the YieldProcessor macro defined in WinNT.h. It is a CPU architecture- @@ -84,6 +90,7 @@ do { \ os_thread_sleep(2000 /* 2 ms */); \ } \ } while (0) +#endif /* !UNIV_HOTBACKUP */ /********************************************************//** Gets the high 32 bits in a ulint. That is makes a shift >> 32, @@ -216,6 +223,7 @@ UNIV_INTERN ib_time_t ut_time(void); /*=========*/ +#ifndef UNIV_HOTBACKUP /**********************************************************//** Returns system time. Upon successful completion, the value 0 is returned; otherwise the @@ -239,6 +247,16 @@ ullint ut_time_us( /*=======*/ ullint* tloc); /*!< out: us since epoch, if non-NULL */ +/**********************************************************//** +Returns the number of milliseconds since some epoch. The +value may wrap around. It should only be used for heuristic +purposes. +@return ms since epoch */ +UNIV_INTERN +ulint +ut_time_ms(void); +/*============*/ +#endif /* !UNIV_HOTBACKUP */ /**********************************************************//** Returns the difference of two times in seconds. diff --git a/storage/xtradb/lock/lock0lock.c b/storage/xtradb/lock/lock0lock.c index f76c27e093c..d1a37b13cd0 100644 --- a/storage/xtradb/lock/lock0lock.c +++ b/storage/xtradb/lock/lock0lock.c @@ -214,7 +214,7 @@ a waiting s-lock request on the next record? If this s-lock was placed by a read cursor moving in the ascending order in the index, we cannot do the insert immediately, because when we finally commit our transaction, the read cursor should see also the new inserted record. So we should -move the read cursor backward from the the next record for it to pass over +move the read cursor backward from the next record for it to pass over the new inserted record. This move backward may be too cumbersome to implement. If we in this situation just enqueue a second x-lock request for our transaction on the next record, then the deadlock mechanism @@ -360,10 +360,9 @@ ibool lock_rec_validate_page( /*===================*/ ulint space, /*!< in: space id */ + ulint zip_size,/*!< in: compressed page size in bytes + or 0 for uncompressed pages */ ulint page_no);/*!< in: page number */ - -/* Define the following in order to enable lock_rec_validate_page() checks. */ -# undef UNIV_DEBUG_LOCK_VALIDATE #endif /* UNIV_DEBUG */ /* The lock system */ @@ -579,6 +578,23 @@ lock_sys_create( } /*********************************************************************//** +Closes the lock system at database shutdown. */ +UNIV_INTERN +void +lock_sys_close(void) +/*================*/ +{ + if (lock_latest_err_file != NULL) { + fclose(lock_latest_err_file); + lock_latest_err_file = NULL; + } + + hash_table_free(lock_sys->rec_hash); + mem_free(lock_sys); + lock_sys = NULL; +} + +/*********************************************************************//** Gets the size of a lock struct. @return size in bytes */ UNIV_INTERN @@ -1739,6 +1755,8 @@ lock_rec_enqueue_waiting( { lock_t* lock; trx_t* trx; + ulint sec; + ulint ms; ut_ad(mutex_own(&kernel_mutex)); @@ -1797,6 +1815,10 @@ lock_rec_enqueue_waiting( trx->que_state = TRX_QUE_LOCK_WAIT; trx->was_chosen_as_deadlock_victim = FALSE; trx->wait_started = time(NULL); + if (innobase_get_slow_log() && trx->take_stats) { + ut_usectime(&sec, &ms); + trx->lock_que_wait_ustarted = (ib_uint64_t)sec * 1000000 + ms; + } ut_a(que_thr_stop(thr)); @@ -2622,6 +2644,7 @@ lock_move_reorganize_page( #ifdef UNIV_DEBUG_LOCK_VALIDATE ut_ad(lock_rec_validate_page(buf_block_get_space(block), + buf_block_get_zip_size(block), buf_block_get_page_no(block))); #endif } @@ -2711,8 +2734,10 @@ lock_move_rec_list_end( #ifdef UNIV_DEBUG_LOCK_VALIDATE ut_ad(lock_rec_validate_page(buf_block_get_space(block), + buf_block_get_zip_size(block), buf_block_get_page_no(block))); ut_ad(lock_rec_validate_page(buf_block_get_space(new_block), + buf_block_get_zip_size(block), buf_block_get_page_no(new_block))); #endif } @@ -2822,6 +2847,7 @@ lock_move_rec_list_start( #ifdef UNIV_DEBUG_LOCK_VALIDATE ut_ad(lock_rec_validate_page(buf_block_get_space(block), + buf_block_get_zip_size(block), buf_block_get_page_no(block))); #endif } @@ -3574,7 +3600,8 @@ lock_table_remove_low( and lock_grant()). Therefore it can be empty and we need to check for that. */ - if (!ib_vector_is_empty(trx->autoinc_locks)) { + if (!lock_get_wait(lock) + && !ib_vector_is_empty(trx->autoinc_locks)) { lock_t* autoinc_lock; autoinc_lock = ib_vector_pop(trx->autoinc_locks); @@ -3607,6 +3634,8 @@ lock_table_enqueue_waiting( { lock_t* lock; trx_t* trx; + ulint sec; + ulint ms; ut_ad(mutex_own(&kernel_mutex)); @@ -3647,8 +3676,10 @@ lock_table_enqueue_waiting( if (lock_deadlock_occurs(lock, trx)) { - lock_reset_lock_and_trx_wait(lock); + /* The order here is important, we don't want to + lose the state of the lock before calling remove. */ lock_table_remove_low(lock); + lock_reset_lock_and_trx_wait(lock); return(DB_DEADLOCK); } @@ -3660,6 +3691,10 @@ lock_table_enqueue_waiting( return(DB_SUCCESS); } + if (innobase_get_slow_log() && trx->take_stats) { + ut_usectime(&sec, &ms); + trx->lock_que_wait_ustarted = (ib_uint64_t)sec * 1000000 + ms; + } trx->que_state = TRX_QUE_LOCK_WAIT; trx->was_chosen_as_deadlock_victim = FALSE; trx->wait_started = time(NULL); @@ -4627,6 +4662,10 @@ lock_rec_queue_validate( next function call: we have to release lock table mutex to obey the latching order */ + /* If this thread is holding the file space latch + (fil_space_t::latch), the following check WILL break + latching order and may cause a deadlock of threads. */ + impl_trx = lock_sec_rec_some_has_impl_off_kernel( rec, index, offsets); @@ -4684,6 +4723,8 @@ ibool lock_rec_validate_page( /*===================*/ ulint space, /*!< in: space id */ + ulint zip_size,/*!< in: compressed page size in bytes + or 0 for uncompressed pages */ ulint page_no)/*!< in: page number */ { dict_index_t* index; @@ -4694,7 +4735,6 @@ lock_rec_validate_page( ulint nth_lock = 0; ulint nth_bit = 0; ulint i; - ulint zip_size; mtr_t mtr; mem_heap_t* heap = NULL; ulint offsets_[REC_OFFS_NORMAL_SIZE]; @@ -4705,7 +4745,6 @@ lock_rec_validate_page( mtr_start(&mtr); - zip_size = fil_space_get_zip_size(space); ut_ad(zip_size != ULINT_UNDEFINED); block = buf_page_get(space, zip_size, page_no, RW_X_LATCH, &mtr); buf_block_dbg_add_level(block, SYNC_NO_ORDER_CHECK); @@ -4750,6 +4789,11 @@ loop: lock_mutex_exit_kernel(); + /* If this thread is holding the file space + latch (fil_space_t::latch), the following + check WILL break the latching order and may + cause a deadlock of threads. */ + lock_rec_queue_validate(block, rec, index, offsets); lock_mutex_enter_kernel(); @@ -4840,7 +4884,9 @@ lock_validate(void) lock_mutex_exit_kernel(); - lock_rec_validate_page(space, page_no); + lock_rec_validate_page(space, + fil_space_get_zip_size(space), + page_no); lock_mutex_enter_kernel(); @@ -5364,6 +5410,20 @@ lock_release_autoinc_last_lock( } /*******************************************************************//** +Check if a transaction holds any autoinc locks. +@return TRUE if the transaction holds any AUTOINC locks. */ +UNIV_INTERN +ibool +lock_trx_holds_autoinc_locks( +/*=========================*/ + const trx_t* trx) /*!< in: transaction */ +{ + ut_a(trx->autoinc_locks != NULL); + + return(!ib_vector_is_empty(trx->autoinc_locks)); +} + +/*******************************************************************//** Release all the transaction's autoinc locks. */ UNIV_INTERN void diff --git a/storage/xtradb/log/log0log.c b/storage/xtradb/log/log0log.c index 3a89f540b77..5d54b75b725 100644 --- a/storage/xtradb/log/log0log.c +++ b/storage/xtradb/log/log0log.c @@ -241,6 +241,7 @@ log_reserve_and_open( ut_a(len < log->buf_size / 2); loop: mutex_enter(&(log->mutex)); + ut_ad(!recv_no_log_write); /* Calculate an upper limit for the space the string may take in the log buffer */ @@ -309,6 +310,7 @@ log_write_low( ut_ad(mutex_own(&(log->mutex))); part_loop: + ut_ad(!recv_no_log_write); /* Calculate a part length */ data_len = (log->buf_free % OS_FILE_LOG_BLOCK_SIZE) + str_len; @@ -362,6 +364,33 @@ part_loop: } /************************************************************//** +*/ +UNIV_INLINE +ulint +log_max_modified_age_async() +{ + if (srv_checkpoint_age_target) { + return(ut_min(log_sys->max_modified_age_async, + srv_checkpoint_age_target + - srv_checkpoint_age_target / 8)); + } else { + return(log_sys->max_modified_age_async); + } +} + +UNIV_INLINE +ulint +log_max_checkpoint_age_async() +{ + if (srv_checkpoint_age_target) { + return(ut_min(log_sys->max_checkpoint_age_async, + srv_checkpoint_age_target)); + } else { + return(log_sys->max_checkpoint_age_async); + } +} + +/************************************************************//** Closes the log. @return lsn */ UNIV_INTERN @@ -377,6 +406,7 @@ log_close(void) ib_uint64_t checkpoint_age; ut_ad(mutex_own(&(log->mutex))); + ut_ad(!recv_no_log_write); lsn = log->lsn; @@ -429,7 +459,7 @@ log_close(void) } } - if (checkpoint_age <= log->max_modified_age_async) { + if (checkpoint_age <= log_max_modified_age_async()) { goto function_exit; } @@ -437,8 +467,8 @@ log_close(void) oldest_lsn = buf_pool_get_oldest_modification(); if (!oldest_lsn - || lsn - oldest_lsn > log->max_modified_age_async - || checkpoint_age > log->max_checkpoint_age_async) { + || lsn - oldest_lsn > log_max_modified_age_async() + || checkpoint_age > log_max_checkpoint_age_async()) { log->check_flush_or_checkpoint = TRUE; } @@ -668,8 +698,6 @@ log_calc_max_ages(void) ulint archive_margin; ulint smallest_archive_margin; - ut_ad(!mutex_own(&(log_sys->mutex))); - mutex_enter(&(log_sys->mutex)); group = UT_LIST_GET_FIRST(log_sys->log_groups); @@ -770,8 +798,6 @@ void log_init(void) /*==========*/ { - byte* buf; - log_sys = mem_alloc(sizeof(log_t)); mutex_create(&log_sys->mutex, SYNC_LOG); @@ -786,8 +812,8 @@ log_init(void) ut_a(LOG_BUFFER_SIZE >= 16 * OS_FILE_LOG_BLOCK_SIZE); ut_a(LOG_BUFFER_SIZE >= 4 * UNIV_PAGE_SIZE); - buf = mem_alloc(LOG_BUFFER_SIZE + OS_FILE_LOG_BLOCK_SIZE); - log_sys->buf = ut_align(buf, OS_FILE_LOG_BLOCK_SIZE); + log_sys->buf_ptr = mem_alloc(LOG_BUFFER_SIZE + OS_FILE_LOG_BLOCK_SIZE); + log_sys->buf = ut_align(log_sys->buf_ptr, OS_FILE_LOG_BLOCK_SIZE); log_sys->buf_size = LOG_BUFFER_SIZE; @@ -832,9 +858,9 @@ log_init(void) rw_lock_create(&log_sys->checkpoint_lock, SYNC_NO_ORDER_CHECK); - log_sys->checkpoint_buf - = ut_align(mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE), - OS_FILE_LOG_BLOCK_SIZE); + log_sys->checkpoint_buf_ptr = mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE); + log_sys->checkpoint_buf = ut_align(log_sys->checkpoint_buf_ptr, + OS_FILE_LOG_BLOCK_SIZE); memset(log_sys->checkpoint_buf, '\0', OS_FILE_LOG_BLOCK_SIZE); /*----------------------------*/ @@ -917,23 +943,33 @@ log_group_init( group->lsn_offset = LOG_FILE_HDR_SIZE; group->n_pending_writes = 0; + group->file_header_bufs_ptr = mem_alloc(sizeof(byte*) * n_files); group->file_header_bufs = mem_alloc(sizeof(byte*) * n_files); #ifdef UNIV_LOG_ARCHIVE + group->archive_file_header_bufs_ptr = mem_alloc( + sizeof(byte*) * n_files); group->archive_file_header_bufs = mem_alloc(sizeof(byte*) * n_files); #endif /* UNIV_LOG_ARCHIVE */ for (i = 0; i < n_files; i++) { - *(group->file_header_bufs + i) = ut_align( - mem_alloc(LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE), + group->file_header_bufs_ptr[i] = mem_alloc( + LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE); + + group->file_header_bufs[i] = ut_align( + group->file_header_bufs_ptr[i], OS_FILE_LOG_BLOCK_SIZE); memset(*(group->file_header_bufs + i), '\0', LOG_FILE_HDR_SIZE); #ifdef UNIV_LOG_ARCHIVE - *(group->archive_file_header_bufs + i) = ut_align( - mem_alloc(LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE), + group->archive_file_header_bufs_ptr[i] = mem_alloc( + LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE); + + group->archive_file_header_bufs[i] = ut_align( + group->archive_file_header_bufs_ptr[i], OS_FILE_LOG_BLOCK_SIZE); + memset(*(group->archive_file_header_bufs + i), '\0', LOG_FILE_HDR_SIZE); #endif /* UNIV_LOG_ARCHIVE */ @@ -946,8 +982,9 @@ log_group_init( group->archived_offset = 0; #endif /* UNIV_LOG_ARCHIVE */ - group->checkpoint_buf = ut_align( - mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE), OS_FILE_LOG_BLOCK_SIZE); + group->checkpoint_buf_ptr = mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE); + group->checkpoint_buf = ut_align(group->checkpoint_buf_ptr, + OS_FILE_LOG_BLOCK_SIZE); memset(group->checkpoint_buf, '\0', OS_FILE_LOG_BLOCK_SIZE); @@ -1117,6 +1154,7 @@ log_io_complete( } mutex_enter(&(log_sys->mutex)); + ut_ad(!recv_no_log_write); ut_a(group->n_pending_writes > 0); ut_a(log_sys->n_pending_writes > 0); @@ -1148,6 +1186,7 @@ log_group_file_header_flush( ulint dest_offset; ut_ad(mutex_own(&(log_sys->mutex))); + ut_ad(!recv_no_log_write); ut_a(nth_file < group->n_files); buf = *(group->file_header_bufs + nth_file); @@ -1219,6 +1258,7 @@ log_group_write_buf( ulint i; ut_ad(mutex_own(&(log_sys->mutex))); + ut_ad(!recv_no_log_write); ut_a(len % OS_FILE_LOG_BLOCK_SIZE == 0); ut_a(((ulint) start_lsn) % OS_FILE_LOG_BLOCK_SIZE == 0); @@ -1361,6 +1401,7 @@ loop: #endif mutex_enter(&(log_sys->mutex)); + ut_ad(!recv_no_log_write); if (flush_to_disk && log_sys->flushed_to_disk_lsn >= lsn) { @@ -1974,6 +2015,7 @@ log_checkpoint( mutex_enter(&(log_sys->mutex)); + ut_ad(!recv_no_log_write); oldest_lsn = log_buf_pool_get_oldest_modification(); mutex_exit(&(log_sys->mutex)); @@ -2047,7 +2089,7 @@ log_make_checkpoint_at( later lsn, if IB_ULONGLONG_MAX, makes a checkpoint at the latest lsn */ ibool write_always) /*!< in: the function normally checks if - the the new checkpoint would have a + the new checkpoint would have a greater lsn than the previous one: if not, then no physical write is done; by setting this parameter TRUE, a @@ -2086,6 +2128,7 @@ loop: do_checkpoint = FALSE; mutex_enter(&(log->mutex)); + ut_ad(!recv_no_log_write); if (log->check_flush_or_checkpoint == FALSE) { mutex_exit(&(log->mutex)); @@ -2103,10 +2146,10 @@ loop: sync = TRUE; advance = 2 * (age - log->max_modified_age_sync); - } else if (age > log->max_modified_age_async) { + } else if (age > log_max_modified_age_async()) { /* A flush is not urgent: we do an asynchronous preflush */ - advance = age - log->max_modified_age_async; + advance = age - log_max_modified_age_async(); } else { advance = 0; } @@ -2120,7 +2163,7 @@ loop: do_checkpoint = TRUE; - } else if (checkpoint_age > log->max_checkpoint_age_async) { + } else if (checkpoint_age > log_max_checkpoint_age_async()) { /* A checkpoint is not urgent: do it asynchronously */ do_checkpoint = TRUE; @@ -3035,6 +3078,7 @@ loop: #endif /* UNIV_LOG_ARCHIVE */ mutex_enter(&(log_sys->mutex)); + ut_ad(!recv_no_log_write); if (log_sys->check_flush_or_checkpoint) { @@ -3120,6 +3164,16 @@ loop: goto loop; } + /* Check that the purge threads ended */ + if (srv_use_purge_thread + && (srv_n_threads_active[SRV_PURGE] != 0 + || srv_n_threads_active[SRV_PURGE_WORKER] != 0)) { + + mutex_exit(&kernel_mutex); + + goto loop; + } + mutex_exit(&kernel_mutex); mutex_enter(&(log_sys->mutex)); @@ -3234,6 +3288,7 @@ loop: ut_a(lsn == log_sys->lsn); } +#ifdef UNIV_LOG_DEBUG /******************************************************//** Checks by parsing that the catenated log segment for a single mtr is consistent. */ @@ -3241,7 +3296,7 @@ UNIV_INTERN ibool log_check_log_recs( /*===============*/ - byte* buf, /*!< in: pointer to the start of + const byte* buf, /*!< in: pointer to the start of the log segment in the log_sys->buf log buffer */ ulint len, /*!< in: segment length in bytes */ @@ -3249,8 +3304,8 @@ log_check_log_recs( { ib_uint64_t contiguous_lsn; ib_uint64_t scanned_lsn; - byte* start; - byte* end; + const byte* start; + const byte* end; byte* buf1; byte* scan_buf; @@ -3283,6 +3338,7 @@ log_check_log_recs( return(TRUE); } +#endif /* UNIV_LOG_DEBUG */ /******************************************************//** Peeks the current lsn. @@ -3326,10 +3382,12 @@ log_print( log_sys->last_checkpoint_lsn); fprintf(file, - "Max checkpoint age %lu\n" - "Modified age %lu\n" - "Checkpoint age %lu\n", + "Max checkpoint age %lu\n" + "Checkpoint age target %lu\n" + "Modified age %lu\n" + "Checkpoint age %lu\n", (ulong) log_sys->max_checkpoint_age, + (ulong) log_max_checkpoint_age_async(), (ulong) (log_sys->lsn - log_buf_pool_get_oldest_modification()), (ulong) (log_sys->lsn - log_sys->last_checkpoint_lsn)); @@ -3363,4 +3421,95 @@ log_refresh_stats(void) log_sys->n_log_ios_old = log_sys->n_log_ios; log_sys->last_printout_time = time(NULL); } + +/********************************************************************** +Closes a log group. */ +static +void +log_group_close( +/*===========*/ + log_group_t* group) /* in,own: log group to close */ +{ + ulint i; + + for (i = 0; i < group->n_files; i++) { + mem_free(group->file_header_bufs_ptr[i]); +#ifdef UNIV_LOG_ARCHIVE + mem_free(group->archive_file_header_bufs_ptr[i]); +#endif /* UNIV_LOG_ARCHIVE */ + } + + mem_free(group->file_header_bufs_ptr); + mem_free(group->file_header_bufs); + +#ifdef UNIV_LOG_ARCHIVE + mem_free(group->archive_file_header_bufs_ptr); + mem_free(group->archive_file_header_bufs); +#endif /* UNIV_LOG_ARCHIVE */ + + mem_free(group->checkpoint_buf_ptr); + + mem_free(group); +} + +/********************************************************** +Shutdown the log system but do not release all the memory. */ +UNIV_INTERN +void +log_shutdown(void) +/*==============*/ +{ + log_group_t* group; + + group = UT_LIST_GET_FIRST(log_sys->log_groups); + + while (UT_LIST_GET_LEN(log_sys->log_groups) > 0) { + log_group_t* prev_group = group; + + group = UT_LIST_GET_NEXT(log_groups, group); + UT_LIST_REMOVE(log_groups, log_sys->log_groups, prev_group); + + log_group_close(prev_group); + } + + mem_free(log_sys->buf_ptr); + log_sys->buf_ptr = NULL; + log_sys->buf = NULL; + mem_free(log_sys->checkpoint_buf_ptr); + log_sys->checkpoint_buf_ptr = NULL; + log_sys->checkpoint_buf = NULL; + + os_event_free(log_sys->no_flush_event); + os_event_free(log_sys->one_flushed_event); + + rw_lock_free(&log_sys->checkpoint_lock); + + mutex_free(&log_sys->mutex); + +#ifdef UNIV_LOG_ARCHIVE + rw_lock_free(&log_sys->archive_lock); + os_event_create(log_sys->archiving_on); +#endif /* UNIV_LOG_ARCHIVE */ + +#ifdef UNIV_LOG_DEBUG + recv_sys_debug_free(); +#endif + + recv_sys_close(); +} + +/********************************************************** +Free the log system data structures. */ +UNIV_INTERN +void +log_mem_free(void) +/*==============*/ +{ + if (log_sys != NULL) { + recv_sys_mem_free(); + mem_free(log_sys); + + log_sys = NULL; + } +} #endif /* !UNIV_HOTBACKUP */ diff --git a/storage/xtradb/log/log0recv.c b/storage/xtradb/log/log0recv.c index 60ca47f1207..77902b48a6f 100644 --- a/storage/xtradb/log/log0recv.c +++ b/storage/xtradb/log/log0recv.c @@ -69,20 +69,25 @@ UNIV_INTERN recv_sys_t* recv_sys = NULL; /** TRUE when applying redo log records during crash recovery; FALSE otherwise. Note that this is FALSE while a background thread is rolling back incomplete transactions. */ -UNIV_INTERN ibool recv_recovery_on = FALSE; +UNIV_INTERN ibool recv_recovery_on; #ifdef UNIV_LOG_ARCHIVE /** TRUE when applying redo log records from an archived log file */ -UNIV_INTERN ibool recv_recovery_from_backup_on = FALSE; +UNIV_INTERN ibool recv_recovery_from_backup_on; #endif /* UNIV_LOG_ARCHIVE */ #ifndef UNIV_HOTBACKUP /** TRUE when recv_init_crash_recovery() has been called. */ -UNIV_INTERN ibool recv_needed_recovery = FALSE; +UNIV_INTERN ibool recv_needed_recovery; +# ifdef UNIV_DEBUG +/** TRUE if writing to the redo log (mtr_commit) is forbidden. +Protected by log_sys->mutex. */ +UNIV_INTERN ibool recv_no_log_write = FALSE; +# endif /* UNIV_DEBUG */ /** TRUE if buf_page_is_corrupted() should check if the log sequence number (FIL_PAGE_LSN) is in the future. Initially FALSE, and set by recv_recovery_from_checkpoint_start_func(). */ -UNIV_INTERN ibool recv_lsn_checks_on = FALSE; +UNIV_INTERN ibool recv_lsn_checks_on; /** There are two conditions under which we scan the logs, the first is normal startup and the second is when we do a recovery from an @@ -92,7 +97,7 @@ startup. If we find log entries that were written after the last checkpoint we know that the server was not cleanly shutdown. We must then initialize the crash recovery environment before attempting to store these entries in the log hash table. */ -static ibool recv_log_scan_is_startup_type = FALSE; +static ibool recv_log_scan_is_startup_type; /** If the following is TRUE, the buffer pool file pages must be invalidated after recovery and no ibuf operations are allowed; this becomes TRUE if @@ -103,7 +108,7 @@ buffer pool before the pages have been recovered to the up-to-date state. TRUE means that recovery is running and no operations on the log files are allowed yet: the variable name is misleading. */ -UNIV_INTERN ibool recv_no_ibuf_operations = FALSE; +UNIV_INTERN ibool recv_no_ibuf_operations; /** TRUE when the redo log is being backed up */ # define recv_is_making_a_backup FALSE /** TRUE when recovering from a backed up redo log file */ @@ -111,30 +116,30 @@ UNIV_INTERN ibool recv_no_ibuf_operations = FALSE; #else /* !UNIV_HOTBACKUP */ # define recv_needed_recovery FALSE /** TRUE when the redo log is being backed up */ -UNIV_INTERN ibool recv_is_making_a_backup = FALSE; +UNIV_INTERN ibool recv_is_making_a_backup = FALSE; /** TRUE when recovering from a backed up redo log file */ UNIV_INTERN ibool recv_is_from_backup = FALSE; # define buf_pool_get_curr_size() (5 * 1024 * 1024) #endif /* !UNIV_HOTBACKUP */ /** The following counter is used to decide when to print info on log scan */ -static ulint recv_scan_print_counter = 0; +static ulint recv_scan_print_counter; /** The type of the previous parsed redo log record */ -static ulint recv_previous_parsed_rec_type = 999999; +static ulint recv_previous_parsed_rec_type; /** The offset of the previous parsed redo log record */ -static ulint recv_previous_parsed_rec_offset = 0; +static ulint recv_previous_parsed_rec_offset; /** The 'multi' flag of the previous parsed redo log record */ -static ulint recv_previous_parsed_rec_is_multi = 0; +static ulint recv_previous_parsed_rec_is_multi; /** Maximum page number encountered in the redo log */ -UNIV_INTERN ulint recv_max_parsed_page_no = 0; +UNIV_INTERN ulint recv_max_parsed_page_no; /** This many frames must be left free in the buffer pool when we scan the log and store the scanned log records in the buffer pool: we will use these free frames to read in pages when we start applying the log records to the database. */ -UNIV_INTERN ulint recv_n_pool_free_frames = 1024; +UNIV_INTERN ulint recv_n_pool_free_frames; /** The maximum lsn we see for a page during the recovery process. If this is bigger than the lsn we are able to scan up to, that is an indication that @@ -165,15 +170,119 @@ recv_sys_create(void) return; } - recv_sys = mem_alloc(sizeof(recv_sys_t)); + recv_sys = mem_alloc(sizeof(*recv_sys)); + memset(recv_sys, 0x0, sizeof(*recv_sys)); mutex_create(&recv_sys->mutex, SYNC_RECV); recv_sys->heap = NULL; recv_sys->addr_hash = NULL; + + recv_sys->stats_recv_start_time = time(NULL); + recv_sys->stats_oldest_modified_lsn = IB_ULONGLONG_MAX; } /********************************************************//** +Release recovery system mutexes. */ +UNIV_INTERN +void +recv_sys_close(void) +/*================*/ +{ + if (recv_sys != NULL) { + if (recv_sys->addr_hash != NULL) { + hash_table_free(recv_sys->addr_hash); + } + + if (recv_sys->heap != NULL) { + mem_heap_free(recv_sys->heap); + } + + if (recv_sys->buf != NULL) { + ut_free(recv_sys->buf); + } + + if (recv_sys->last_block_buf_start != NULL) { + mem_free(recv_sys->last_block_buf_start); + } + + mutex_free(&recv_sys->mutex); + + mem_free(recv_sys); + recv_sys = NULL; + } +} + +/********************************************************//** +Frees the recovery system memory. */ +UNIV_INTERN +void +recv_sys_mem_free(void) +/*===================*/ +{ + if (recv_sys != NULL) { + if (recv_sys->addr_hash != NULL) { + hash_table_free(recv_sys->addr_hash); + } + + if (recv_sys->heap != NULL) { + mem_heap_free(recv_sys->heap); + } + + if (recv_sys->buf != NULL) { + ut_free(recv_sys->buf); + } + + if (recv_sys->last_block_buf_start != NULL) { + mem_free(recv_sys->last_block_buf_start); + } + + mem_free(recv_sys); + recv_sys = NULL; + } +} + +/************************************************************ +Reset the state of the recovery system variables. */ +UNIV_INTERN +void +recv_sys_var_init(void) +/*===================*/ +{ + recv_lsn_checks_on = FALSE; + + recv_n_pool_free_frames = 1024; + + recv_recovery_on = FALSE; + +#ifdef UNIV_LOG_ARCHIVE + recv_recovery_from_backup_on = FALSE; +#endif /* UNIV_LOG_ARCHIVE */ + + recv_needed_recovery = FALSE; + + recv_lsn_checks_on = FALSE; + + recv_log_scan_is_startup_type = FALSE; + + recv_no_ibuf_operations = FALSE; + + recv_scan_print_counter = 0; + + recv_previous_parsed_rec_type = 999999; + + recv_previous_parsed_rec_offset = 0; + + recv_previous_parsed_rec_is_multi = 0; + + recv_max_parsed_page_no = 0; + + recv_n_pool_free_frames = 1024; + + recv_max_page_lsn = 0; +} + +/************************************************************ Inits the recovery system for a recovery operation. */ UNIV_INTERN void @@ -248,8 +357,8 @@ recv_sys_empty_hash(void) Frees the recovery system. */ static void -recv_sys_free(void) -/*===============*/ +recv_sys_debug_free(void) +/*=====================*/ { mutex_enter(&(recv_sys->mutex)); @@ -258,8 +367,10 @@ recv_sys_free(void) ut_free(recv_sys->buf); mem_free(recv_sys->last_block_buf_start); - recv_sys->addr_hash = NULL; + recv_sys->buf = NULL; recv_sys->heap = NULL; + recv_sys->addr_hash = NULL; + recv_sys->last_block_buf_start = NULL; mutex_exit(&(recv_sys->mutex)); } @@ -853,6 +964,11 @@ recv_parse_or_apply_log_rec_body( } switch (type) { +#ifdef UNIV_LOG_LSN_DEBUG + case MLOG_LSN: + /* The LSN is checked in recv_parse_log_rec(). */ + break; +#endif /* UNIV_LOG_LSN_DEBUG */ case MLOG_1BYTE: case MLOG_2BYTES: case MLOG_4BYTES: case MLOG_8BYTES: #ifdef UNIV_DEBUG if (page && page_type == FIL_PAGE_TYPE_ALLOCATED @@ -1223,6 +1339,11 @@ recv_add_to_hash_table( len = rec_end - body; + if (srv_recovery_stats) { + recv_sys->stats_log_recs++; + recv_sys->stats_log_len_sum += len; + } + recv = mem_heap_alloc(recv_sys->heap, sizeof(recv_t)); recv->type = type; recv->len = rec_end - body; @@ -1269,7 +1390,7 @@ recv_add_to_hash_table( sizeof(recv_data_t) + len); *prev_field = recv_data; - ut_memcpy(((byte*)recv_data) + sizeof(recv_data_t), body, len); + memcpy(recv_data + 1, body, len); prev_field = &(recv_data->next); @@ -1327,12 +1448,14 @@ recv_recover_page_func( buf_block_t* block) /*!< in/out: buffer block */ { page_t* page; + page_zip_des_t* page_zip; recv_addr_t* recv_addr; recv_t* recv; byte* buf; ib_uint64_t start_lsn; ib_uint64_t end_lsn; ib_uint64_t page_lsn; + ib_uint64_t page_lsn_orig; ib_uint64_t page_newest_lsn; ibool modification_to_page; #ifndef UNIV_HOTBACKUP @@ -1372,12 +1495,21 @@ recv_recover_page_func( recv_addr->state = RECV_BEING_PROCESSED; + if (srv_recovery_stats) { + if (just_read_in) { + recv_sys->stats_recover_pages_with_read++; + } else { + recv_sys->stats_recover_pages_without_read++; + } + } + mutex_exit(&(recv_sys->mutex)); mtr_start(&mtr); mtr_set_log_mode(&mtr, MTR_LOG_NONE); page = block->frame; + page_zip = buf_block_get_page_zip(block); #ifndef UNIV_HOTBACKUP if (just_read_in) { @@ -1400,6 +1532,7 @@ recv_recover_page_func( /* Read the newest modification lsn from the page */ page_lsn = mach_read_ull(page + FIL_PAGE_LSN); + page_lsn_orig = page_lsn; #ifndef UNIV_HOTBACKUP /* It may be that the page has been modified in the buffer @@ -1419,6 +1552,21 @@ recv_recover_page_func( modification_to_page = FALSE; start_lsn = end_lsn = 0; + if (srv_recovery_stats) { + mutex_enter(&(recv_sys->mutex)); + if (page_lsn_orig && recv_sys->stats_oldest_modified_lsn > page_lsn_orig) { + recv_sys->stats_oldest_modified_lsn = page_lsn_orig; + } + if (page_lsn_orig && recv_sys->stats_newest_modified_lsn < page_lsn_orig) { + recv_sys->stats_newest_modified_lsn = page_lsn_orig; + } + if (UT_LIST_GET_LAST(recv_addr->rec_list)->start_lsn + < page_lsn_orig) { + recv_sys->stats_pages_already_new++; + } + mutex_exit(&(recv_sys->mutex)); + } + recv = UT_LIST_GET_FIRST(recv_addr->rec_list); while (recv) { @@ -1438,13 +1586,19 @@ recv_recover_page_func( if (recv->type == MLOG_INIT_FILE_PAGE) { page_lsn = page_newest_lsn; - mach_write_ull(page + UNIV_PAGE_SIZE - - FIL_PAGE_END_LSN_OLD_CHKSUM, 0); - mach_write_ull(page + FIL_PAGE_LSN, 0); + memset(FIL_PAGE_LSN + page, 0, 8); + memset(UNIV_PAGE_SIZE - FIL_PAGE_END_LSN_OLD_CHKSUM + + page, 0, 8); + + if (page_zip) { + memset(FIL_PAGE_LSN + page_zip->data, 0, 8); + } } if (recv->start_lsn >= page_lsn) { + ib_uint64_t end_lsn; + if (!modification_to_page) { modification_to_page = TRUE; @@ -1466,11 +1620,24 @@ recv_recover_page_func( recv_parse_or_apply_log_rec_body(recv->type, buf, buf + recv->len, block, &mtr); - mach_write_ull(page + UNIV_PAGE_SIZE - - FIL_PAGE_END_LSN_OLD_CHKSUM, - recv->start_lsn + recv->len); - mach_write_ull(page + FIL_PAGE_LSN, - recv->start_lsn + recv->len); + + if (srv_recovery_stats) { + mutex_enter(&(recv_sys->mutex)); + recv_sys->stats_applied_log_recs++; + recv_sys->stats_applied_log_len_sum += recv->len; + mutex_exit(&(recv_sys->mutex)); + } + + end_lsn = recv->start_lsn + recv->len; + mach_write_ull(FIL_PAGE_LSN + page, end_lsn); + mach_write_ull(UNIV_PAGE_SIZE + - FIL_PAGE_END_LSN_OLD_CHKSUM + + page, end_lsn); + + if (page_zip) { + mach_write_ull(FIL_PAGE_LSN + + page_zip->data, end_lsn); + } } if (recv->len > RECV_DATA_BLOCK_SIZE) { @@ -1561,6 +1728,13 @@ recv_read_in_area( } } + if (srv_recovery_stats && n) { + mutex_enter(&(recv_sys->mutex)); + recv_sys->stats_read_requested_pages += n; + recv_sys->stats_read_in_area[n - 1]++; + mutex_exit(&(recv_sys->mutex)); + } + buf_read_recv_pages(FALSE, space, zip_size, page_nos, n); /* fprintf(stderr, "Recv pages at %lu n %lu\n", page_nos[0], n); @@ -1688,6 +1862,7 @@ loop: /* Flush all the file pages to disk and invalidate them in the buffer pool */ + ut_d(recv_no_log_write = TRUE); mutex_exit(&(recv_sys->mutex)); mutex_exit(&(log_sys->mutex)); @@ -1701,6 +1876,7 @@ loop: mutex_enter(&(log_sys->mutex)); mutex_enter(&(recv_sys->mutex)); + ut_d(recv_no_log_write = FALSE); recv_no_ibuf_operations = FALSE; } @@ -1712,6 +1888,10 @@ loop: if (has_printed) { fprintf(stderr, "InnoDB: Apply batch completed\n"); + + if (srv_recovery_stats) { + recv_sys->stats_recv_turns++; + } } mutex_exit(&(recv_sys->mutex)); @@ -1912,6 +2092,17 @@ recv_parse_log_rec( return(0); } +#ifdef UNIV_LOG_LSN_DEBUG + if (*type == MLOG_LSN) { + ib_uint64_t lsn = (ib_uint64_t) *space << 32 | *page_no; +# ifdef UNIV_LOG_DEBUG + ut_a(lsn == log_sys->old_lsn); +# else /* UNIV_LOG_DEBUG */ + ut_a(lsn == recv_sys->recovered_lsn); +# endif /* UNIV_LOG_DEBUG */ + } +#endif /* UNIV_LOG_LSN_DEBUG */ + /* Check that page_no is sensible */ if (UNIV_UNLIKELY(*page_no > 0x8FFFFFFFUL)) { @@ -2169,6 +2360,12 @@ loop: #endif /* In normal mysqld crash recovery we do not try to replay file operations */ +#ifdef UNIV_LOG_LSN_DEBUG + } else if (type == MLOG_LSN) { + /* Do not add these records to the hash table. + The page number and space id fields are misused + for something else. */ +#endif /* UNIV_LOG_LSN_DEBUG */ } else { recv_add_to_hash_table(type, space, page_no, body, ptr + len, old_lsn, @@ -2200,11 +2397,11 @@ loop: = recv_sys->recovered_offset + total_len; recv_previous_parsed_rec_is_multi = 1; - if ((!store_to_hash) && (type != MLOG_MULTI_REC_END)) { #ifdef UNIV_LOG_DEBUG + if ((!store_to_hash) && (type != MLOG_MULTI_REC_END)) { recv_check_incomplete_log_recs(ptr, len); -#endif /* UNIV_LOG_DEBUG */ } +#endif /* UNIV_LOG_DEBUG */ #ifdef UNIV_DEBUG if (log_debug_writes) { @@ -2268,7 +2465,11 @@ loop: break; } - if (store_to_hash) { + if (store_to_hash +#ifdef UNIV_LOG_LSN_DEBUG + && type != MLOG_LSN +#endif /* UNIV_LOG_LSN_DEBUG */ + ) { recv_add_to_hash_table(type, space, page_no, body, ptr + len, old_lsn, @@ -2417,8 +2618,7 @@ recv_scan_log_recs( scanned_lsn = start_lsn; more_data = FALSE; - while (log_block < buf + len && !finished) { - + do { no = log_block_get_hdr_no(log_block); /* fprintf(stderr, "Log block header no %lu\n", no); @@ -2548,10 +2748,11 @@ recv_scan_log_recs( /* Log data for this group ends here */ finished = TRUE; + break; } else { log_block += OS_FILE_LOG_BLOCK_SIZE; } - } + } while (log_block < buf + len && !finished); *group_scanned_lsn = scanned_lsn; @@ -3078,6 +3279,84 @@ recv_recovery_from_checkpoint_finish(void) } #endif /* UNIV_DEBUG */ + if (recv_needed_recovery && srv_recovery_stats) { + FILE* file = stderr; + ulint i; + + fprintf(stderr, + "InnoDB: Applying log records was done. Its statistics are followings.\n"); + + fprintf(stderr, + "============================================================\n" + "-------------------\n" + "RECOVERY STATISTICS\n" + "-------------------\n"); + fprintf(stderr, + "Recovery time: %g sec. (%lu turns)\n", + difftime(time(NULL), recv_sys->stats_recv_start_time), + recv_sys->stats_recv_turns); + + fprintf(stderr, + "\n" + "Data page IO statistics\n" + " Requested pages: %lu\n" + " Read pages: %lu\n" + " Written pages: %lu\n" + " (Dirty blocks): %lu\n", + recv_sys->stats_read_requested_pages, + recv_sys->stats_read_io_pages, + recv_sys->stats_write_io_pages, + UT_LIST_GET_LEN(buf_pool->flush_list)); + + fprintf(stderr, + " Grouping IO [times]:\n" + "\tnumber of pages,\n" + "\t\tread request neighbors (in %d pages chunk),\n" + "\t\t\tcombined read IO,\n" + "\t\t\t\tcombined write IO\n", + RECV_READ_AHEAD_AREA); + for (i = 0; i < ut_max(RECV_READ_AHEAD_AREA, + OS_AIO_MERGE_N_CONSECUTIVE); i++) { + fprintf(stderr, + "\t%3lu,\t%lu,\t%lu,\t%lu\n", i + 1, + (i < RECV_READ_AHEAD_AREA) ? + recv_sys->stats_read_in_area[i] : 0, + (i < OS_AIO_MERGE_N_CONSECUTIVE) ? + recv_sys->stats_read_io_consecutive[i] : 0, + (i < OS_AIO_MERGE_N_CONSECUTIVE) ? + recv_sys->stats_write_io_consecutive[i] : 0); + } + + fprintf(stderr, + "\n" + "Recovery process statistics\n" + " Checked pages by doublewrite buffer: %lu\n" + " Overwritten pages from doublewrite: %lu\n" + " Recovered pages by io_thread: %lu\n" + " Recovered pages by main thread: %lu\n" + " Parsed log records to apply: %lu\n" + " Sum of the length: %lu\n" + " Applied log records: %lu\n" + " Sum of the length: %lu\n" + " Pages which are already new enough: %lu (It may not be accurate, if turns > 1)\n" + " Oldest page's LSN: %llu\n" + " Newest page's LSN: %llu\n", + recv_sys->stats_doublewrite_check_pages, + recv_sys->stats_doublewrite_overwrite_pages, + recv_sys->stats_recover_pages_with_read, + recv_sys->stats_recover_pages_without_read, + recv_sys->stats_log_recs, + recv_sys->stats_log_len_sum, + recv_sys->stats_applied_log_recs, + recv_sys->stats_applied_log_len_sum, + recv_sys->stats_pages_already_new, + recv_sys->stats_oldest_modified_lsn, + recv_sys->stats_newest_modified_lsn); + + fprintf(stderr, + "============================================================\n"); + } + if (recv_needed_recovery) { trx_sys_print_mysql_master_log_pos(); trx_sys_print_mysql_binlog_offset(); @@ -3104,8 +3383,13 @@ recv_recovery_from_checkpoint_finish(void) recv_recovery_on = FALSE; #ifndef UNIV_LOG_DEBUG - recv_sys_free(); + recv_sys_debug_free(); #endif + /* Roll back any recovered data dictionary transactions, so + that the data dictionary tables will be free of any locks. + The data dictionary latch should guarantee that there is at + most one data dictionary transaction active at a time. */ + trx_rollback_or_clean_recovered(FALSE); /* Drop partially created indexes. */ row_merge_drop_temp_indexes(); diff --git a/storage/xtradb/mem/mem0dbg.c b/storage/xtradb/mem/mem0dbg.c index a20eb2ad7d2..01eda20ec45 100644 --- a/storage/xtradb/mem/mem0dbg.c +++ b/storage/xtradb/mem/mem0dbg.c @@ -170,6 +170,17 @@ mem_init( mem_comm_pool = mem_pool_create(size); } + +/******************************************************************//** +Closes the memory system. */ +UNIV_INTERN +void +mem_close(void) +/*===========*/ +{ + mem_pool_free(mem_comm_pool); + mem_comm_pool = NULL; +} #endif /* !UNIV_HOTBACKUP */ #ifdef UNIV_MEM_DEBUG diff --git a/storage/xtradb/mem/mem0mem.c b/storage/xtradb/mem/mem0mem.c index e0dc8716f13..ccb2fd8a7b4 100644 --- a/storage/xtradb/mem/mem0mem.c +++ b/storage/xtradb/mem/mem0mem.c @@ -475,16 +475,18 @@ mem_heap_block_free( len = block->len; block->magic_n = MEM_FREED_BLOCK_MAGIC_N; +#ifndef UNIV_HOTBACKUP + if (!srv_use_sys_malloc) { #ifdef UNIV_MEM_DEBUG - /* In the debug version we set the memory to a random combination - of hex 0xDE and 0xAD. */ + /* In the debug version we set the memory to a random + combination of hex 0xDE and 0xAD. */ - mem_erase_buf((byte*)block, len); + mem_erase_buf((byte*)block, len); #else /* UNIV_MEM_DEBUG */ - UNIV_MEM_ASSERT_AND_FREE(block, len); + UNIV_MEM_ASSERT_AND_FREE(block, len); #endif /* UNIV_MEM_DEBUG */ -#ifndef UNIV_HOTBACKUP + } if (type == MEM_HEAP_DYNAMIC || len < UNIV_PAGE_SIZE / 2) { ut_ad(!buf_block); @@ -495,6 +497,14 @@ mem_heap_block_free( buf_block_free(buf_block); } #else /* !UNIV_HOTBACKUP */ +#ifdef UNIV_MEM_DEBUG + /* In the debug version we set the memory to a random + combination of hex 0xDE and 0xAD. */ + + mem_erase_buf((byte*)block, len); +#else /* UNIV_MEM_DEBUG */ + UNIV_MEM_ASSERT_AND_FREE(block, len); +#endif /* UNIV_MEM_DEBUG */ ut_free(block); #endif /* !UNIV_HOTBACKUP */ } diff --git a/storage/xtradb/mem/mem0pool.c b/storage/xtradb/mem/mem0pool.c index c8fea97a6a3..c4f8af607e0 100644 --- a/storage/xtradb/mem/mem0pool.c +++ b/storage/xtradb/mem/mem0pool.c @@ -261,6 +261,18 @@ mem_pool_create( } /********************************************************************//** +Frees a memory pool. */ +UNIV_INTERN +void +mem_pool_free( +/*==========*/ + mem_pool_t* pool) /*!< in, own: memory pool */ +{ + ut_free(pool->buf); + ut_free(pool); +} + +/********************************************************************//** Fills the specified free list. @return TRUE if we were able to insert a block to the free list */ static diff --git a/storage/xtradb/mtr/mtr0mtr.c b/storage/xtradb/mtr/mtr0mtr.c index 183e575fadd..34e6d3ffc92 100644 --- a/storage/xtradb/mtr/mtr0mtr.c +++ b/storage/xtradb/mtr/mtr0mtr.c @@ -36,6 +36,7 @@ Created 11/26/1995 Heikki Tuuri #include "buf0flu.h" #ifndef UNIV_HOTBACKUP +# include "log0recv.h" /*****************************************************************//** Releases the item in the slot given. */ UNIV_INLINE @@ -148,7 +149,6 @@ mtr_log_reserve_and_write( dyn_array_t* mlog; dyn_block_t* block; ulint data_size; - ibool success; byte* first_data; ut_ad(mtr); @@ -167,8 +167,8 @@ mtr_log_reserve_and_write( if (mlog->heap == NULL) { mtr->end_lsn = log_reserve_and_write_fast( first_data, dyn_block_get_used(mlog), - &(mtr->start_lsn), &success); - if (success) { + &mtr->start_lsn); + if (mtr->end_lsn) { return; } @@ -215,6 +215,8 @@ mtr_commit( ut_d(mtr->state = MTR_COMMITTING); #ifndef UNIV_HOTBACKUP + /* This is a dirty read, for debugging. */ + ut_ad(!recv_no_log_write); write_log = mtr->modifications && mtr->n_log_recs; if (write_log) { diff --git a/storage/xtradb/os/os0file.c b/storage/xtradb/os/os0file.c index f961ea2adb2..d3b3edea29f 100644 --- a/storage/xtradb/os/os0file.c +++ b/storage/xtradb/os/os0file.c @@ -55,6 +55,9 @@ Created 10/21/1995 Heikki Tuuri #include "srv0start.h" #include "fil0fil.h" #include "buf0buf.h" +#include "trx0sys.h" +#include "trx0trx.h" +#include "log0recv.h" #ifndef UNIV_HOTBACKUP # include "os0sync.h" # include "os0thread.h" @@ -88,7 +91,9 @@ UNIV_INTERN ibool os_do_not_call_flush_at_each_write = FALSE; /* We do not call os_file_flush in every os_file_write. */ #endif /* UNIV_DO_FLUSH */ -#ifndef UNIV_HOTBACKUP +#ifdef UNIV_HOTBACKUP +# define os_aio_use_native_aio FALSE +#else /* UNIV_HOTBACKUP */ /* We use these mutexes to protect lseek + file i/o operation, if the OS does not provide an atomic pread or pwrite, or similar */ #define OS_FILE_N_SEEK_MUTEXES 16 @@ -235,7 +240,7 @@ static ulint os_aio_n_segments = ULINT_UNDEFINED; /** If the following is TRUE, read i/o handler threads try to wait until a batch of new read requests have been posted */ static volatile ibool os_aio_recommend_sleep_for_read_threads = FALSE; -#endif /* !UNIV_HOTBACKUP */ +#endif /* UNIV_HOTBACKUP */ UNIV_INTERN ulint os_n_file_reads = 0; UNIV_INTERN ulint os_bytes_read_since_printout = 0; @@ -352,6 +357,19 @@ os_file_get_last_error( " software or another instance\n" "InnoDB: of MySQL." " Please close it to get rid of this error.\n"); + } else if (err == ERROR_WORKING_SET_QUOTA + || err == ERROR_NO_SYSTEM_RESOURCES) { + fprintf(stderr, + "InnoDB: The error means that there are no" + " sufficient system resources or quota to" + " complete the operation.\n"); + } else if (err == ERROR_OPERATION_ABORTED) { + fprintf(stderr, + "InnoDB: The error means that the I/O" + " operation has been aborted\n" + "InnoDB: because of either a thread exit" + " or an application request.\n" + "InnoDB: Retry attempt is made.\n"); } else { fprintf(stderr, "InnoDB: Some operating system error numbers" @@ -373,6 +391,11 @@ os_file_get_last_error( } else if (err == ERROR_SHARING_VIOLATION || err == ERROR_LOCK_VIOLATION) { return(OS_FILE_SHARING_VIOLATION); + } else if (err == ERROR_WORKING_SET_QUOTA + || err == ERROR_NO_SYSTEM_RESOURCES) { + return(OS_FILE_INSUFFICIENT_RESOURCE); + } else if (err == ERROR_OPERATION_ABORTED) { + return(OS_FILE_OPERATION_ABORTED); } else { return(100 + err); } @@ -491,6 +514,14 @@ os_file_handle_error_cond_exit( os_thread_sleep(10000000); /* 10 sec */ return(TRUE); + } else if (err == OS_FILE_INSUFFICIENT_RESOURCE) { + + os_thread_sleep(100000); /* 100 ms */ + return(TRUE); + } else if (err == OS_FILE_OPERATION_ABORTED) { + + os_thread_sleep(100000); /* 100 ms */ + return(TRUE); } else { if (name) { fprintf(stderr, "InnoDB: File name %s\n", name); @@ -854,6 +885,23 @@ next_file: ret = stat(full_path, &statinfo); if (ret) { + + if (errno == ENOENT) { + /* readdir() returned a file that does not exist, + it must have been deleted in the meantime. Do what + would have happened if the file was deleted before + readdir() - ignore and go to the next entry. + If this is the last entry then info->name will still + contain the name of the deleted file when this + function returns, but this is not an issue since the + caller shouldn't be looking at info when end of + directory is returned. */ + + ut_free(full_path); + + goto next_file; + } + os_file_handle_error_no_exit(full_path, "stat"); ut_free(full_path); @@ -1282,6 +1330,7 @@ try_again: } #endif #ifdef UNIV_NON_BUFFERED_IO +# ifndef UNIV_HOTBACKUP if (type == OS_LOG_FILE && srv_flush_log_at_trx_commit == 2) { /* Do not use unbuffered i/o to log files because value 2 denotes that we do not flush the log at every @@ -1290,10 +1339,14 @@ try_again: == SRV_WIN_IO_UNBUFFERED) { attributes = attributes | FILE_FLAG_NO_BUFFERING; } -#endif +# else /* !UNIV_HOTBACKUP */ + attributes = attributes | FILE_FLAG_NO_BUFFERING; +# endif /* !UNIV_HOTBACKUP */ +#endif /* UNIV_NON_BUFFERED_IO */ } else if (purpose == OS_FILE_NORMAL) { attributes = 0; #ifdef UNIV_NON_BUFFERED_IO +# ifndef UNIV_HOTBACKUP if (type == OS_LOG_FILE && srv_flush_log_at_trx_commit == 2) { /* Do not use unbuffered i/o to log files because value 2 denotes that we do not flush the log at every @@ -1302,7 +1355,10 @@ try_again: == SRV_WIN_IO_UNBUFFERED) { attributes = attributes | FILE_FLAG_NO_BUFFERING; } -#endif +# else /* !UNIV_HOTBACKUP */ + attributes = attributes | FILE_FLAG_NO_BUFFERING; +# endif /* !UNIV_HOTBACKUP */ +#endif /* UNIV_NON_BUFFERED_IO */ } else { attributes = 0; ut_error; @@ -2046,20 +2102,30 @@ os_file_flush( /*******************************************************************//** Does a synchronous read operation in Posix. @return number of bytes read, -1 if error */ +#define os_file_pread(file, buf, n, offset, offset_high) \ + _os_file_pread(file, buf, n, offset, offset_high, NULL); + static ssize_t -os_file_pread( +_os_file_pread( /*==========*/ os_file_t file, /*!< in: handle to a file */ void* buf, /*!< in: buffer where to read */ ulint n, /*!< in: number of bytes to read */ ulint offset, /*!< in: least significant 32 bits of file offset from where to read */ - ulint offset_high) /*!< in: most significant 32 bits of + ulint offset_high, /*!< in: most significant 32 bits of offset */ + trx_t* trx) { off_t offs; +#if defined(HAVE_PREAD) && !defined(HAVE_BROKEN_PREAD) ssize_t n_bytes; +#endif /* HAVE_PREAD && !HAVE_BROKEN_PREAD */ + ulint sec; + ulint ms; + ib_uint64_t start_time; + ib_uint64_t finish_time; ut_a((offset & 0xFFFFFFFFUL) == offset); @@ -2080,6 +2146,15 @@ os_file_pread( os_n_file_reads++; + if (innobase_get_slow_log() && trx && trx->take_stats) + { + trx->io_reads++; + trx->io_read += n; + ut_usectime(&sec, &ms); + start_time = (ib_uint64_t)sec * 1000000 + ms; + } else { + start_time = 0; + } #if defined(HAVE_PREAD) && !defined(HAVE_BROKEN_PREAD) os_mutex_enter(os_file_count_mutex); os_file_n_pending_preads++; @@ -2093,21 +2168,32 @@ os_file_pread( os_n_pending_reads--; os_mutex_exit(os_file_count_mutex); + if (innobase_get_slow_log() && trx && trx->take_stats && start_time) + { + ut_usectime(&sec, &ms); + finish_time = (ib_uint64_t)sec * 1000000 + ms; + trx->io_reads_wait_timer += (ulint)(finish_time - start_time); + } + return(n_bytes); #else { off_t ret_offset; ssize_t ret; +#ifndef UNIV_HOTBACKUP ulint i; +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads++; os_mutex_exit(os_file_count_mutex); +#ifndef UNIV_HOTBACKUP /* Protect the seek / read operation with a mutex */ i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES; os_mutex_enter(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ ret_offset = lseek(file, offs, SEEK_SET); @@ -2117,12 +2203,21 @@ os_file_pread( ret = read(file, buf, (ssize_t)n); } +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads--; os_mutex_exit(os_file_count_mutex); + if (innobase_get_slow_log() && trx && trx->take_stats && start_time) + { + ut_usectime(&sec, &ms); + finish_time = (ib_uint64_t)sec * 1000000 + ms; + trx->io_reads_wait_timer += (ulint)(finish_time - start_time); + } + return(ret); } #endif @@ -2195,16 +2290,20 @@ os_file_pwrite( #else { off_t ret_offset; +# ifndef UNIV_HOTBACKUP ulint i; +# endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_writes++; os_mutex_exit(os_file_count_mutex); +# ifndef UNIV_HOTBACKUP /* Protect the seek / write operation with a mutex */ i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES; os_mutex_enter(os_file_seek_mutexes[i]); +# endif /* UNIV_HOTBACKUP */ ret_offset = lseek(file, offs, SEEK_SET); @@ -2230,7 +2329,9 @@ os_file_pwrite( # endif /* UNIV_DO_FLUSH */ func_exit: +# ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +# endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_writes--; @@ -2247,7 +2348,7 @@ Requests a synchronous positioned read operation. @return TRUE if request was successful, FALSE if fail */ UNIV_INTERN ibool -os_file_read( +_os_file_read( /*=========*/ os_file_t file, /*!< in: handle to a file */ void* buf, /*!< in: buffer where to read */ @@ -2255,7 +2356,8 @@ os_file_read( offset where to read */ ulint offset_high, /*!< in: most significant 32 bits of offset */ - ulint n) /*!< in: number of bytes to read */ + ulint n, /*!< in: number of bytes to read */ + trx_t* trx) { #ifdef __WIN__ BOOL ret; @@ -2264,7 +2366,9 @@ os_file_read( DWORD low; DWORD high; ibool retry; +#ifndef UNIV_HOTBACKUP ulint i; +#endif /* !UNIV_HOTBACKUP */ ut_a((offset & 0xFFFFFFFFUL) == offset); @@ -2283,16 +2387,20 @@ try_again: os_n_pending_reads++; os_mutex_exit(os_file_count_mutex); +#ifndef UNIV_HOTBACKUP /* Protect the seek / read operation with a mutex */ i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES; os_mutex_enter(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ ret2 = SetFilePointer(file, low, &high, FILE_BEGIN); if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) { +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads--; @@ -2303,7 +2411,9 @@ try_again: ret = ReadFile(file, buf, (DWORD) n, &len, NULL); +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads--; @@ -2312,14 +2422,14 @@ try_again: if (ret && len == n) { return(TRUE); } -#else +#else /* __WIN__ */ ibool retry; ssize_t ret; os_bytes_read_since_printout += n; try_again: - ret = os_file_pread(file, buf, n, offset, offset_high); + ret = _os_file_pread(file, buf, n, offset, offset_high, trx); if ((ulint)ret == n) { @@ -2331,7 +2441,7 @@ try_again: "InnoDB: Was only able to read %ld.\n", (ulong)n, (ulong)offset_high, (ulong)offset, (long)ret); -#endif +#endif /* __WIN__ */ #ifdef __WIN__ error_handling: #endif @@ -2380,7 +2490,9 @@ os_file_read_no_error_handling( DWORD low; DWORD high; ibool retry; +#ifndef UNIV_HOTBACKUP ulint i; +#endif /* !UNIV_HOTBACKUP */ ut_a((offset & 0xFFFFFFFFUL) == offset); @@ -2399,16 +2511,20 @@ try_again: os_n_pending_reads++; os_mutex_exit(os_file_count_mutex); +#ifndef UNIV_HOTBACKUP /* Protect the seek / read operation with a mutex */ i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES; os_mutex_enter(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ ret2 = SetFilePointer(file, low, &high, FILE_BEGIN); if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) { +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads--; @@ -2419,7 +2535,9 @@ try_again: ret = ReadFile(file, buf, (DWORD) n, &len, NULL); +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_reads--; @@ -2428,7 +2546,7 @@ try_again: if (ret && len == n) { return(TRUE); } -#else +#else /* __WIN__ */ ibool retry; ssize_t ret; @@ -2441,7 +2559,7 @@ try_again: return(TRUE); } -#endif +#endif /* __WIN__ */ #ifdef __WIN__ error_handling: #endif @@ -2500,9 +2618,11 @@ os_file_write( DWORD ret2; DWORD low; DWORD high; - ulint i; ulint n_retries = 0; ulint err; +#ifndef UNIV_HOTBACKUP + ulint i; +#endif /* !UNIV_HOTBACKUP */ ut_a((offset & 0xFFFFFFFF) == offset); @@ -2519,16 +2639,20 @@ retry: os_n_pending_writes++; os_mutex_exit(os_file_count_mutex); +#ifndef UNIV_HOTBACKUP /* Protect the seek / write operation with a mutex */ i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES; os_mutex_enter(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ ret2 = SetFilePointer(file, low, &high, FILE_BEGIN); if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) { +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_writes--; @@ -2562,7 +2686,9 @@ retry: } # endif /* UNIV_DO_FLUSH */ +#ifndef UNIV_HOTBACKUP os_mutex_exit(os_file_seek_mutexes[i]); +#endif /* !UNIV_HOTBACKUP */ os_mutex_enter(os_file_count_mutex); os_n_pending_writes--; @@ -2988,6 +3114,34 @@ os_aio_array_create( return(array); } +/************************************************************************//** +Frees an aio wait array. */ +static +void +os_aio_array_free( +/*==============*/ + os_aio_array_t* array) /*!< in, own: array to free */ +{ +#ifdef WIN_ASYNC_IO + ulint i; + + for (i = 0; i < array->n_slots; i++) { + os_aio_slot_t* slot = os_aio_array_get_nth_slot(array, i); + os_event_free(slot->event); + } +#endif /* WIN_ASYNC_IO */ + +#ifdef __WIN__ + ut_free(array->native_events); +#endif /* __WIN__ */ + os_mutex_free(array->mutex); + os_event_free(array->not_full); + os_event_free(array->is_empty); + + ut_free(array->slots); + ut_free(array); +} + /*********************************************************************** Initializes the asynchronous io system. Creates one array each for ibuf and log i/o. Also creates one array each for read and write where each @@ -3061,6 +3215,35 @@ os_aio_init( } +/*********************************************************************** +Frees the asynchronous io system. */ +UNIV_INTERN +void +os_aio_free(void) +/*=============*/ +{ + ulint i; + + os_aio_array_free(os_aio_ibuf_array); + os_aio_ibuf_array = NULL; + os_aio_array_free(os_aio_log_array); + os_aio_log_array = NULL; + os_aio_array_free(os_aio_read_array); + os_aio_read_array = NULL; + os_aio_array_free(os_aio_write_array); + os_aio_write_array = NULL; + os_aio_array_free(os_aio_sync_array); + os_aio_sync_array = NULL; + + for (i = 0; i < os_aio_n_segments; i++) { + os_event_free(os_aio_segment_wait_events[i]); + } + + ut_free(os_aio_segment_wait_events); + os_aio_segment_wait_events = 0; + os_aio_n_segments = 0; +} + #ifdef WIN_ASYNC_IO /************************************************************************//** Wakes up all async i/o threads in the array in Windows async i/o at @@ -3211,7 +3394,8 @@ os_aio_array_reserve_slot( offset */ ulint offset_high, /*!< in: most significant 32 bits of offset */ - ulint len) /*!< in: length of the block to read or write */ + ulint len, /*!< in: length of the block to read or write */ + trx_t* trx) { os_aio_slot_t* slot; #ifdef WIN_ASYNC_IO @@ -3432,9 +3616,21 @@ void os_aio_simulated_put_read_threads_to_sleep(void) /*============================================*/ { + +/* The idea of putting background IO threads to sleep is only for +Windows when using simulated AIO. Windows XP seems to schedule +background threads too eagerly to allow for coalescing during +readahead requests. */ +#ifdef __WIN__ os_aio_array_t* array; ulint g; + if (os_aio_use_native_aio) { + /* We do not use simulated aio: do nothing */ + + return; + } + os_aio_recommend_sleep_for_read_threads = TRUE; for (g = 0; g < os_aio_n_segments; g++) { @@ -3445,6 +3641,7 @@ os_aio_simulated_put_read_threads_to_sleep(void) os_event_reset(os_aio_segment_wait_events[g]); } } +#endif /* __WIN__ */ } /*******************************************************************//** @@ -3482,10 +3679,11 @@ os_aio( (can be used to identify a completed aio operation); ignored if mode is OS_AIO_SYNC */ - void* message2)/*!< in: message for the aio handler + void* message2,/*!< in: message for the aio handler (can be used to identify a completed aio operation); ignored if mode is OS_AIO_SYNC */ + trx_t* trx) { os_aio_array_t* array; os_aio_slot_t* slot; @@ -3524,8 +3722,8 @@ os_aio( wait in the Windows case. */ if (type == OS_FILE_READ) { - return(os_file_read(file, buf, offset, - offset_high, n)); + return(_os_file_read(file, buf, offset, + offset_high, n, trx)); } ut_a(type == OS_FILE_WRITE); @@ -3558,8 +3756,13 @@ try_again: ut_error; } + if (trx && type == OS_FILE_READ) + { + trx->io_reads++; + trx->io_read += n; + } slot = os_aio_array_reserve_slot(type, array, message1, message2, file, - name, buf, offset, offset_high, n); + name, buf, offset, offset_high, n, trx); if (type == OS_FILE_READ) { if (os_aio_use_native_aio) { #ifdef WIN_ASYNC_IO @@ -3679,6 +3882,7 @@ os_aio_windows_handle( ibool ret_val; BOOL ret; DWORD len; + BOOL retry = FALSE; if (segment == ULINT_UNDEFINED) { array = os_aio_sync_array; @@ -3732,14 +3936,52 @@ os_aio_windows_handle( ut_a(TRUE == os_file_flush(slot->file)); } #endif /* UNIV_DO_FLUSH */ + } else if (os_file_handle_error(slot->name, "Windows aio")) { + + retry = TRUE; } else { - os_file_handle_error(slot->name, "Windows aio"); ret_val = FALSE; } os_mutex_exit(array->mutex); + if (retry) { + /* retry failed read/write operation synchronously. + No need to hold array->mutex. */ + + switch (slot->type) { + case OS_FILE_WRITE: + ret = WriteFile(slot->file, slot->buf, + slot->len, &len, + &(slot->control)); + + break; + case OS_FILE_READ: + ret = ReadFile(slot->file, slot->buf, + slot->len, &len, + &(slot->control)); + + break; + default: + ut_error; + } + + if (!ret && GetLastError() == ERROR_IO_PENDING) { + /* aio was queued successfully! + We want a synchronous i/o operation on a + file where we also use async i/o: in Windows + we must use the same wait mechanism as for + async i/o */ + + ret = GetOverlappedResult(slot->file, + &(slot->control), + &len, TRUE); + } + + ret_val = ret && len == slot->len; + } + os_aio_array_free_slot(array, slot); return(ret_val); @@ -4018,6 +4260,18 @@ consecutive_loop: } } + if (srv_recovery_stats && recv_recovery_is_on() && n_consecutive) { + mutex_enter(&(recv_sys->mutex)); + if (slot->type == OS_FILE_READ) { + recv_sys->stats_read_io_pages += n_consecutive; + recv_sys->stats_read_io_consecutive[n_consecutive - 1]++; + } else if (slot->type == OS_FILE_WRITE) { + recv_sys->stats_write_io_pages += n_consecutive; + recv_sys->stats_write_io_consecutive[n_consecutive - 1]++; + } + mutex_exit(&(recv_sys->mutex)); + } + os_mutex_enter(array->mutex); /* Mark the i/os done in slots */ diff --git a/storage/xtradb/os/os0proc.c b/storage/xtradb/os/os0proc.c index a0ea9a1b258..48922886f23 100644 --- a/storage/xtradb/os/os0proc.c +++ b/storage/xtradb/os/os0proc.c @@ -97,6 +97,7 @@ os_mem_alloc_large( fprintf(stderr, "InnoDB: HugeTLB: Warning: Failed to" " attach shared memory segment, errno %d\n", errno); + ptr = NULL; } /* Remove the shared memory segment so that it will be diff --git a/storage/xtradb/os/os0sync.c b/storage/xtradb/os/os0sync.c index 4ec340b72b5..60467242e14 100644 --- a/storage/xtradb/os/os0sync.c +++ b/storage/xtradb/os/os0sync.c @@ -86,6 +86,9 @@ os_sync_init(void) UT_LIST_INIT(os_event_list); UT_LIST_INIT(os_mutex_list); + os_sync_mutex = NULL; + os_sync_mutex_inited = FALSE; + os_sync_mutex = os_mutex_create(NULL); os_sync_mutex_inited = TRUE; @@ -713,6 +716,7 @@ os_fast_mutex_free( os_mutex_enter(os_sync_mutex); } + ut_ad(os_fast_mutex_count > 0); os_fast_mutex_count--; if (UNIV_LIKELY(os_sync_mutex_inited)) { diff --git a/storage/xtradb/os/os0thread.c b/storage/xtradb/os/os0thread.c index 9a2d95cb166..34818ada804 100644 --- a/storage/xtradb/os/os0thread.c +++ b/storage/xtradb/os/os0thread.c @@ -233,6 +233,7 @@ os_thread_exit( #ifdef __WIN__ ExitThread((DWORD)exit_value); #else + pthread_detach(pthread_self()); pthread_exit(exit_value); #endif } diff --git a/storage/xtradb/page/page0cur.c b/storage/xtradb/page/page0cur.c index 65f3ba67439..f10f16a7dd9 100644 --- a/storage/xtradb/page/page0cur.c +++ b/storage/xtradb/page/page0cur.c @@ -1195,7 +1195,7 @@ page_cur_insert_rec_zip_reorg( } /* Out of space: restore the page */ - if (!page_zip_decompress(page_zip, page)) { + if (!page_zip_decompress(page_zip, page, FALSE)) { ut_error; /* Memory corrupted? */ } ut_ad(page_validate(page, index)); diff --git a/storage/xtradb/page/page0page.c b/storage/xtradb/page/page0page.c index f056ef77bdc..ab2ba60570e 100644 --- a/storage/xtradb/page/page0page.c +++ b/storage/xtradb/page/page0page.c @@ -45,7 +45,7 @@ Created 2/2/1994 Heikki Tuuri ============== The index page consists of a page header which contains the page's -id and other information. On top of it are the the index records +id and other information. On top of it are the index records in a heap linked into a one way linear list according to alphabetic order. Just below page end is an array of pointers which we call page directory, @@ -679,7 +679,7 @@ page_copy_rec_list_end( if (UNIV_UNLIKELY (!page_zip_decompress(new_page_zip, - new_page))) { + new_page, FALSE))) { ut_error; } ut_ad(page_validate(new_page, index)); @@ -792,7 +792,7 @@ page_copy_rec_list_start( if (UNIV_UNLIKELY (!page_zip_decompress(new_page_zip, - new_page))) { + new_page, FALSE))) { ut_error; } ut_ad(page_validate(new_page, index)); diff --git a/storage/xtradb/page/page0zip.c b/storage/xtradb/page/page0zip.c index 92ba0ec768a..aa5e39ff04a 100644 --- a/storage/xtradb/page/page0zip.c +++ b/storage/xtradb/page/page0zip.c @@ -47,8 +47,10 @@ Created June 2005 by Marko Makela # define buf_LRU_stat_inc_unzip() ((void) 0) #endif /* !UNIV_HOTBACKUP */ +#ifndef UNIV_HOTBACKUP /** Statistics on compression, indexed by page_zip_des_t::ssize - 1 */ UNIV_INTERN page_zip_stat_t page_zip_stat[PAGE_ZIP_NUM_SSIZE - 1]; +#endif /* !UNIV_HOTBACKUP */ /* Please refer to ../include/page0zip.ic for a description of the compressed page format. */ @@ -1144,7 +1146,9 @@ page_zip_compress( ulint* offsets = NULL; ulint n_blobs = 0; byte* storage;/* storage of uncompressed columns */ +#ifndef UNIV_HOTBACKUP ullint usec = ut_time_us(NULL); +#endif /* !UNIV_HOTBACKUP */ #ifdef PAGE_ZIP_COMPRESS_DBG FILE* logfile = NULL; #endif @@ -1208,7 +1212,9 @@ page_zip_compress( } } #endif /* PAGE_ZIP_COMPRESS_DBG */ +#ifndef UNIV_HOTBACKUP page_zip_stat[page_zip->ssize - 1].compressed++; +#endif /* !UNIV_HOTBACKUP */ if (UNIV_UNLIKELY(n_dense * PAGE_ZIP_DIR_SLOT_SIZE >= page_zip_get_size(page_zip))) { @@ -1345,8 +1351,10 @@ err_exit: fclose(logfile); } #endif /* PAGE_ZIP_COMPRESS_DBG */ +#ifndef UNIV_HOTBACKUP page_zip_stat[page_zip->ssize - 1].compressed_usec += ut_time_us(NULL) - usec; +#endif /* !UNIV_HOTBACKUP */ return(FALSE); } @@ -1404,12 +1412,14 @@ err_exit: fclose(logfile); } #endif /* PAGE_ZIP_COMPRESS_DBG */ +#ifndef UNIV_HOTBACKUP { page_zip_stat_t* zip_stat = &page_zip_stat[page_zip->ssize - 1]; zip_stat->compressed_ok++; zip_stat->compressed_usec += ut_time_us(NULL) - usec; } +#endif /* !UNIV_HOTBACKUP */ return(TRUE); } @@ -2811,7 +2821,11 @@ page_zip_decompress( /*================*/ page_zip_des_t* page_zip,/*!< in: data, ssize; out: m_start, m_end, m_nonempty, n_blobs */ - page_t* page) /*!< out: uncompressed page, may be trashed */ + page_t* page, /*!< out: uncompressed page, may be trashed */ + ibool all) /*!< in: TRUE=decompress the whole page; + FALSE=verify but do not copy some + page header fields that should not change + after page creation */ { z_stream d_stream; dict_index_t* index = NULL; @@ -2820,7 +2834,9 @@ page_zip_decompress( ulint trx_id_col = ULINT_UNDEFINED; mem_heap_t* heap; ulint* offsets; +#ifndef UNIV_HOTBACKUP ullint usec = ut_time_us(NULL); +#endif /* !UNIV_HOTBACKUP */ ut_ad(page_zip_simple_validate(page_zip)); UNIV_MEM_ASSERT_W(page, UNIV_PAGE_SIZE); @@ -2839,13 +2855,36 @@ page_zip_decompress( heap = mem_heap_create(n_dense * (3 * sizeof *recs) + UNIV_PAGE_SIZE); recs = mem_heap_alloc(heap, n_dense * (2 * sizeof *recs)); + if (all) { + /* Copy the page header. */ + memcpy(page, page_zip->data, PAGE_DATA); + } else { + /* Check that the bytes that we skip are identical. */ +#if defined UNIV_DEBUG || defined UNIV_ZIP_DEBUG + ut_a(!memcmp(FIL_PAGE_TYPE + page, + FIL_PAGE_TYPE + page_zip->data, + PAGE_HEADER - FIL_PAGE_TYPE)); + ut_a(!memcmp(PAGE_HEADER + PAGE_LEVEL + page, + PAGE_HEADER + PAGE_LEVEL + page_zip->data, + PAGE_DATA - (PAGE_HEADER + PAGE_LEVEL))); +#endif /* UNIV_DEBUG || UNIV_ZIP_DEBUG */ + + /* Copy the mutable parts of the page header. */ + memcpy(page, page_zip->data, FIL_PAGE_TYPE); + memcpy(PAGE_HEADER + page, PAGE_HEADER + page_zip->data, + PAGE_LEVEL - PAGE_N_DIR_SLOTS); + +#if defined UNIV_DEBUG || defined UNIV_ZIP_DEBUG + /* Check that the page headers match after copying. */ + ut_a(!memcmp(page, page_zip->data, PAGE_DATA)); +#endif /* UNIV_DEBUG || UNIV_ZIP_DEBUG */ + } + #ifdef UNIV_ZIP_DEBUG - /* Clear the page. */ - memset(page, 0x55, UNIV_PAGE_SIZE); + /* Clear the uncompressed page, except the header. */ + memset(PAGE_DATA + page, 0x55, UNIV_PAGE_SIZE - PAGE_DATA); #endif /* UNIV_ZIP_DEBUG */ - UNIV_MEM_INVALID(page, UNIV_PAGE_SIZE); - /* Copy the page header. */ - memcpy(page, page_zip->data, PAGE_DATA); + UNIV_MEM_INVALID(PAGE_DATA + page, UNIV_PAGE_SIZE - PAGE_DATA); /* Copy the page directory. */ if (UNIV_UNLIKELY(!page_zip_dir_decode(page_zip, page, recs, @@ -2976,12 +3015,14 @@ err_exit: page_zip_fields_free(index); mem_heap_free(heap); +#ifndef UNIV_HOTBACKUP { page_zip_stat_t* zip_stat = &page_zip_stat[page_zip->ssize - 1]; zip_stat->decompressed++; zip_stat->decompressed_usec += ut_time_us(NULL) - usec; } +#endif /* !UNIV_HOTBACKUP */ /* Update the stat counter for LRU policy. */ buf_LRU_stat_inc_unzip(); @@ -3084,7 +3125,7 @@ page_zip_validate_low( #endif /* UNIV_DEBUG_VALGRIND */ temp_page_zip = *page_zip; - valid = page_zip_decompress(&temp_page_zip, temp_page); + valid = page_zip_decompress(&temp_page_zip, temp_page, TRUE); if (!valid) { fputs("page_zip_validate(): failed to decompress\n", stderr); goto func_exit; @@ -4362,8 +4403,8 @@ IMPORTANT: if page_zip_reorganize() is invoked on a leaf page of a non-clustered index, the caller must update the insert buffer free bits in the same mini-transaction in such a way that the modification will be redo-logged. -@return TRUE on success, FALSE on failure; page and page_zip will be -left intact on failure. */ +@return TRUE on success, FALSE on failure; page_zip will be left +intact on failure, but page will be overwritten. */ UNIV_INTERN ibool page_zip_reorganize( @@ -4428,9 +4469,6 @@ page_zip_reorganize( if (UNIV_UNLIKELY(!page_zip_compress(page_zip, page, index, mtr))) { - /* Restore the old page and exit. */ - buf_frame_copy(page, temp_page); - #ifndef UNIV_HOTBACKUP buf_block_free(temp_block); #endif /* !UNIV_HOTBACKUP */ @@ -4591,7 +4629,8 @@ corrupt: memcpy(page_zip->data + page_zip_get_size(page_zip) - trailer_size, ptr + 8 + size, trailer_size); - if (UNIV_UNLIKELY(!page_zip_decompress(page_zip, page))) { + if (UNIV_UNLIKELY(!page_zip_decompress(page_zip, page, + TRUE))) { goto corrupt; } diff --git a/storage/xtradb/pars/lexyy.c b/storage/xtradb/pars/lexyy.c index 37d892e51e3..815395ea316 100644 --- a/storage/xtradb/pars/lexyy.c +++ b/storage/xtradb/pars/lexyy.c @@ -2778,3 +2778,16 @@ static void yyfree (void * ptr ) + +/********************************************************************** +Release any resources used by the lexer. */ +UNIV_INTERN +void +pars_lexer_close(void) +/*==================*/ +{ + yylex_destroy(); + free(stringbuf); + stringbuf = NULL; + stringbuf_len_alloc = stringbuf_len = 0; +} diff --git a/storage/xtradb/pars/pars0lex.l b/storage/xtradb/pars/pars0lex.l index 4abff65e98b..55ed17f82e1 100644 --- a/storage/xtradb/pars/pars0lex.l +++ b/storage/xtradb/pars/pars0lex.l @@ -661,3 +661,16 @@ In the state 'id', only two actions are possible (defined below). */ } %% + +/********************************************************************** +Release any resources used by the lexer. */ +UNIV_INTERN +void +pars_lexer_close(void) +/*==================*/ +{ + yylex_destroy(); + free(stringbuf); + stringbuf = NULL; + stringbuf_len_alloc = stringbuf_len = 0; +} diff --git a/storage/xtradb/plug.in b/storage/xtradb/plug.in index ed9555bb512..2d7bd51ec17 100644 --- a/storage/xtradb/plug.in +++ b/storage/xtradb/plug.in @@ -40,19 +40,11 @@ MYSQL_PLUGIN_ACTIONS(innobase, [ irix*|osf*|sysv5uw7*|openbsd*) CFLAGS="$CFLAGS -DUNIV_MUST_NOT_INLINE";; *solaris*|*SunOS*) - # Begin Solaris atomic function checks - AC_CHECK_FUNCS(atomic_cas_ulong atomic_cas_32 \ - atomic_cas_64 atomic_add_long, - AC_DEFINE( - [HAVE_SOLARIS_ATOMICS], - [1], - [Define to 1 if Solaris supports \ - atomic functions.])) - ### End Solaris atomic function checks - CFLAGS="$CFLAGS -DUNIV_SOLARIS";; esac + INNODB_DYNAMIC_CFLAGS="-DMYSQL_DYNAMIC_PLUGIN" + case "$target_cpu" in x86_64) # The AMD64 ABI forbids absolute addresses in shared libraries @@ -63,7 +55,60 @@ MYSQL_PLUGIN_ACTIONS(innobase, [ ;; esac AC_SUBST(INNODB_DYNAMIC_CFLAGS) + + AC_MSG_CHECKING(whether GCC atomic builtins are available) + # either define HAVE_IB_GCC_ATOMIC_BUILTINS or not + AC_TRY_RUN( + [ + int main() + { + long x; + long y; + long res; + char c; + + x = 10; + y = 123; + res = __sync_bool_compare_and_swap(&x, x, y); + if (!res || x != y) { + return(1); + } + + x = 10; + y = 123; + res = __sync_bool_compare_and_swap(&x, x + 1, y); + if (res || x != 10) { + return(1); + } + + x = 10; + y = 123; + res = __sync_add_and_fetch(&x, y); + if (res != 123 + 10 || x != 123 + 10) { + return(1); + } + + c = 10; + res = __sync_lock_test_and_set(&c, 123); + if (res != 10 || c != 123) { + return(1); + } + + return(0); + } + ], + [ + AC_DEFINE([HAVE_IB_GCC_ATOMIC_BUILTINS], [1], + [GCC atomic builtins are available]) + AC_MSG_RESULT(yes) + ], + [ + AC_MSG_RESULT(no) + ] + ) + AC_MSG_CHECKING(whether pthread_t can be used by GCC atomic builtins) + # either define HAVE_IB_ATOMIC_PTHREAD_T_GCC or not AC_TRY_RUN( [ #include <pthread.h> @@ -84,47 +129,73 @@ MYSQL_PLUGIN_ACTIONS(innobase, [ } ], [ - AC_DEFINE([HAVE_ATOMIC_PTHREAD_T], [1], + AC_DEFINE([HAVE_IB_ATOMIC_PTHREAD_T_GCC], [1], [pthread_t can be used by GCC atomic builtins]) AC_MSG_RESULT(yes) ], [ AC_MSG_RESULT(no) ] - ) + ) + + AC_MSG_CHECKING(whether Solaris libc atomic functions are available) + # either define HAVE_IB_SOLARIS_ATOMICS or not + AC_CHECK_FUNCS(atomic_add_long \ + atomic_cas_32 \ + atomic_cas_64 \ + atomic_cas_ulong, - # Try using solaris atomics on SunOS if GCC atomics are not available - AC_CHECK_DECLS( - [HAVE_ATOMIC_PTHREAD_T], + AC_DEFINE([HAVE_IB_SOLARIS_ATOMICS], [1], + [Define to 1 if Solaris libc atomic functions \ + are available]) + ) + + AC_MSG_CHECKING(whether pthread_t can be used by Solaris libc atomic functions) + # either define HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS or not + AC_TRY_RUN( [ - AC_MSG_NOTICE(no need to check pthread_t size) + #include <pthread.h> + #include <string.h> + + int main(int argc, char** argv) { + pthread_t x1; + pthread_t x2; + pthread_t x3; + + memset(&x1, 0x0, sizeof(x1)); + memset(&x2, 0x0, sizeof(x2)); + memset(&x3, 0x0, sizeof(x3)); + + if (sizeof(pthread_t) == 4) { + + atomic_cas_32(&x1, x2, x3); + + } else if (sizeof(pthread_t) == 8) { + + atomic_cas_64(&x1, x2, x3); + + } else { + + return(1); + } + + return(0); + } ], [ - AC_CHECK_DECLS( - [HAVE_SOLARIS_ATOMICS], - [ - AC_MSG_CHECKING(checking if pthread_t size is integral) - AC_TRY_RUN( - [ - #include <pthread.h> - int main() - { - pthread_t x = 0; - return(0); - } - ], - [ - AC_DEFINE([HAVE_ATOMIC_PTHREAD_T], [1], + AC_DEFINE([HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS], [1], [pthread_t can be used by solaris atomics]) - AC_MSG_RESULT(yes) - # size of pthread_t is needed for typed solaris atomics - AC_CHECK_SIZEOF([pthread_t], [], [#include <pthread.h>]) - ], - [ - AC_MSG_RESULT(no) - ]) - ]) - ]) + AC_MSG_RESULT(yes) + ], + [ + AC_MSG_RESULT(no) + ] + ) + + # this is needed to know which one of atomic_cas_32() or atomic_cas_64() + # to use in the source + AC_CHECK_SIZEOF([pthread_t], [], [#include <pthread.h>]) + # Check for x86 PAUSE instruction AC_MSG_CHECKING(for x86 PAUSE instruction) # We have to actually try running the test program, because of a bug @@ -141,7 +212,7 @@ MYSQL_PLUGIN_ACTIONS(innobase, [ } ], [ - AC_DEFINE([IB_HAVE_PAUSE_INSTRUCTION], [1], [Does x86 PAUSE instruction exist]) + AC_DEFINE([HAVE_IB_PAUSE_INSTRUCTION], [1], [Does x86 PAUSE instruction exist]) AC_MSG_RESULT(yes) ], [ diff --git a/storage/xtradb/que/que0que.c b/storage/xtradb/que/que0que.c index 54b1e7535fa..2fe046fa9b8 100644 --- a/storage/xtradb/que/que0que.c +++ b/storage/xtradb/que/que0que.c @@ -518,6 +518,7 @@ que_graph_free_recursive( upd_node_t* upd; tab_node_t* cre_tab; ind_node_t* cre_ind; + purge_node_t* purge; if (node == NULL) { @@ -579,6 +580,13 @@ que_graph_free_recursive( mem_heap_free(ins->entry_sys_heap); break; + case QUE_NODE_PURGE: + purge = node; + + mem_heap_free(purge->heap); + + break; + case QUE_NODE_UPDATE: upd = node; diff --git a/storage/xtradb/rem/rem0cmp.c b/storage/xtradb/rem/rem0cmp.c index 3924ed69b82..45230f1d7b1 100644 --- a/storage/xtradb/rem/rem0cmp.c +++ b/storage/xtradb/rem/rem0cmp.c @@ -36,7 +36,7 @@ Created 7/1/1994 Heikki Tuuri The records are put into alphabetical order in the following way: let F be the first field where two records disagree. -If there is a character in some position n where the the +If there is a character in some position n where the records disagree, the order is determined by comparison of the characters at position n, possibly after collating transformation. If there is no such character, @@ -76,7 +76,7 @@ cmp_debug_dtuple_rec_with_match( /*************************************************************//** This function is used to compare two data fields for which the data type is such that we must use MySQL code to compare them. The prototype here -must be a copy of the the one in ha_innobase.cc! +must be a copy of the one in ha_innobase.cc! @return 1, 0, -1, if a is greater, equal, less than b, respectively */ extern int @@ -399,7 +399,7 @@ next_byte: /*************************************************************//** This function is used to compare a data tuple to a physical record. Only dtuple->n_fields_cmp first fields are taken into account for -the the data tuple! If we denote by n = n_fields_cmp, then rec must +the data tuple! If we denote by n = n_fields_cmp, then rec must have either m >= n fields, or it must differ from dtuple in some of the m fields rec has. If rec has an externally stored field we do not compare it but return with value 0 if such a comparison should be diff --git a/storage/xtradb/row/row0ins.c b/storage/xtradb/row/row0ins.c index 930c9ec1fc7..fe51fce82c4 100644 --- a/storage/xtradb/row/row0ins.c +++ b/storage/xtradb/row/row0ins.c @@ -141,7 +141,7 @@ row_ins_alloc_sys_fields( dfield = dtuple_get_nth_field(row, dict_col_get_no(col)); - ptr = mem_heap_alloc(heap, DATA_ROW_ID_LEN); + ptr = mem_heap_zalloc(heap, DATA_ROW_ID_LEN); dfield_set_data(dfield, ptr, DATA_ROW_ID_LEN); @@ -152,7 +152,7 @@ row_ins_alloc_sys_fields( col = dict_table_get_sys_col(table, DATA_TRX_ID); dfield = dtuple_get_nth_field(row, dict_col_get_no(col)); - ptr = mem_heap_alloc(heap, DATA_TRX_ID_LEN); + ptr = mem_heap_zalloc(heap, DATA_TRX_ID_LEN); dfield_set_data(dfield, ptr, DATA_TRX_ID_LEN); @@ -163,7 +163,7 @@ row_ins_alloc_sys_fields( col = dict_table_get_sys_col(table, DATA_ROLL_PTR); dfield = dtuple_get_nth_field(row, dict_col_get_no(col)); - ptr = mem_heap_alloc(heap, DATA_ROLL_PTR_LEN); + ptr = mem_heap_zalloc(heap, DATA_ROLL_PTR_LEN); dfield_set_data(dfield, ptr, DATA_ROLL_PTR_LEN); } @@ -1191,7 +1191,7 @@ row_ins_check_foreign_constraint( /*=============================*/ ibool check_ref,/*!< in: TRUE if we want to check that the referenced table is ok, FALSE if we - want to to check the foreign key table */ + want to check the foreign key table */ dict_foreign_t* foreign,/*!< in: foreign constraint; NOTE that the tables mentioned in it must be in the dictionary cache if they exist at all */ diff --git a/storage/xtradb/row/row0merge.c b/storage/xtradb/row/row0merge.c index 05a45dc647c..25f041c0885 100644 --- a/storage/xtradb/row/row0merge.c +++ b/storage/xtradb/row/row0merge.c @@ -60,9 +60,19 @@ Completed by Sunny Bains and Marko Makela #ifdef UNIV_DEBUG /** Set these in order ot enable debug printout. */ /* @{ */ +/** Log the outcome of each row_merge_cmp() call, comparing records. */ static ibool row_merge_print_cmp; +/** Log each record read from temporary file. */ static ibool row_merge_print_read; +/** Log each record write to temporary file. */ static ibool row_merge_print_write; +/** Log each row_merge_blocks() call, merging two blocks of records to +a bigger one. */ +static ibool row_merge_print_block; +/** Log each block read from temporary file. */ +static ibool row_merge_print_block_read; +/** Log each block read from temporary file. */ +static ibool row_merge_print_block_write; /* @} */ #endif /* UNIV_DEBUG */ @@ -109,8 +119,9 @@ typedef struct row_merge_buf_struct row_merge_buf_t; /** Information about temporary files used in merge sort */ struct merge_file_struct { - int fd; /*!< file descriptor */ - ulint offset; /*!< file offset */ + int fd; /*!< file descriptor */ + ulint offset; /*!< file offset (end of file) */ + ib_uint64_t n_rec; /*!< number of records in the file */ }; /** Information about temporary files used in merge sort */ @@ -682,6 +693,13 @@ row_merge_read( ib_uint64_t ofs = ((ib_uint64_t) offset) * sizeof *buf; ibool success; +#ifdef UNIV_DEBUG + if (row_merge_print_block_read) { + fprintf(stderr, "row_merge_read fd=%d ofs=%lu\n", + fd, (ulong) offset); + } +#endif /* UNIV_DEBUG */ + success = os_file_read_no_error_handling(OS_FILE_FROM_FD(fd), buf, (ulint) (ofs & 0xFFFFFFFF), (ulint) (ofs >> 32), @@ -709,6 +727,13 @@ row_merge_write( ib_uint64_t ofs = ((ib_uint64_t) offset) * sizeof(row_merge_block_t); +#ifdef UNIV_DEBUG + if (row_merge_print_block_write) { + fprintf(stderr, "row_merge_write fd=%d ofs=%lu\n", + fd, (ulong) offset); + } +#endif /* UNIV_DEBUG */ + return(UNIV_LIKELY(os_file_write("(merge)", OS_FILE_FROM_FD(fd), buf, (ulint) (ofs & 0xFFFFFFFF), (ulint) (ofs >> 32), @@ -718,7 +743,7 @@ row_merge_write( /********************************************************************//** Read a merge record. @return pointer to next record, or NULL on I/O error or end of list */ -static +static __attribute__((nonnull)) const byte* row_merge_read_rec( /*===============*/ @@ -1070,7 +1095,7 @@ row_merge_cmp( Reads clustered index of the table and create temporary files containing the index entries for the indexes to be built. @return DB_SUCCESS or error */ -static +static __attribute__((nonnull)) ulint row_merge_read_clustered_index( /*===========================*/ @@ -1175,6 +1200,12 @@ row_merge_read_clustered_index( in order to release the latch on the old page. */ if (btr_pcur_is_after_last_on_page(&pcur)) { + if (UNIV_UNLIKELY(trx_is_interrupted(trx))) { + i = 0; + err = DB_INTERRUPTED; + goto err_exit; + } + btr_pcur_store_position(&pcur, &mtr); mtr_commit(&mtr); mtr_start(&mtr); @@ -1233,6 +1264,7 @@ row_merge_read_clustered_index( if (UNIV_LIKELY (row && row_merge_buf_add(buf, row, ext))) { + file->n_rec++; continue; } @@ -1274,14 +1306,19 @@ err_exit: UNIV_MEM_INVALID(block[0], sizeof block[0]); merge_buf[i] = row_merge_buf_empty(buf); - /* Try writing the record again, now that - the buffer has been written out and emptied. */ + if (UNIV_LIKELY(row != NULL)) { + /* Try writing the record again, now + that the buffer has been written out + and emptied. */ - if (UNIV_UNLIKELY - (row && !row_merge_buf_add(buf, row, ext))) { - /* An empty buffer should have enough - room for at least one record. */ - ut_error; + if (UNIV_UNLIKELY + (!row_merge_buf_add(buf, row, ext))) { + /* An empty buffer should have enough + room for at least one record. */ + ut_error; + } + + file->n_rec++; } } @@ -1320,7 +1357,7 @@ func_exit: b2 = row_merge_write_rec(&block[2], &buf[2], b2, \ of->fd, &of->offset, \ mrec##N, offsets##N); \ - if (UNIV_UNLIKELY(!b2)) { \ + if (UNIV_UNLIKELY(!b2 || ++of->n_rec > file->n_rec)) { \ goto corrupt; \ } \ b##N = row_merge_read_rec(&block[N], &buf[N], \ @@ -1336,14 +1373,14 @@ func_exit: } while (0) /*************************************************************//** -Merge two blocks of linked lists on disk and write a bigger block. +Merge two blocks of records on disk and write a bigger block. @return DB_SUCCESS or error code */ static ulint row_merge_blocks( /*=============*/ const dict_index_t* index, /*!< in: index being created */ - merge_file_t* file, /*!< in/out: file containing + const merge_file_t* file, /*!< in: file containing index entries */ row_merge_block_t* block, /*!< in/out: 3 buffers */ ulint* foffs0, /*!< in/out: offset of first @@ -1366,6 +1403,17 @@ row_merge_blocks( ulint* offsets0;/* offsets of mrec0 */ ulint* offsets1;/* offsets of mrec1 */ +#ifdef UNIV_DEBUG + if (row_merge_print_block) { + fprintf(stderr, + "row_merge_blocks fd=%d ofs=%lu + fd=%d ofs=%lu" + " = fd=%d ofs=%lu\n", + file->fd, (ulong) *foffs0, + file->fd, (ulong) *foffs1, + of->fd, (ulong) of->offset); + } +#endif /* UNIV_DEBUG */ + heap = row_merge_heap_create(index, &offsets0, &offsets1); /* Write a record and read the next record. Split the output @@ -1438,16 +1486,88 @@ done1: } /*************************************************************//** +Copy a block of index entries. +@return TRUE on success, FALSE on failure */ +static __attribute__((nonnull)) +ibool +row_merge_blocks_copy( +/*==================*/ + const dict_index_t* index, /*!< in: index being created */ + const merge_file_t* file, /*!< in: input file */ + row_merge_block_t* block, /*!< in/out: 3 buffers */ + ulint* foffs0, /*!< in/out: input file offset */ + merge_file_t* of) /*!< in/out: output file */ +{ + mem_heap_t* heap; /*!< memory heap for offsets0, offsets1 */ + + mrec_buf_t buf[3]; /*!< buffer for handling + split mrec in block[] */ + const byte* b0; /*!< pointer to block[0] */ + byte* b2; /*!< pointer to block[2] */ + const mrec_t* mrec0; /*!< merge rec, points to block[0] */ + ulint* offsets0;/* offsets of mrec0 */ + ulint* offsets1;/* dummy offsets */ + +#ifdef UNIV_DEBUG + if (row_merge_print_block) { + fprintf(stderr, + "row_merge_blocks_copy fd=%d ofs=%lu" + " = fd=%d ofs=%lu\n", + file->fd, (ulong) foffs0, + of->fd, (ulong) of->offset); + } +#endif /* UNIV_DEBUG */ + + heap = row_merge_heap_create(index, &offsets0, &offsets1); + + /* Write a record and read the next record. Split the output + file in two halves, which can be merged on the following pass. */ + + if (!row_merge_read(file->fd, *foffs0, &block[0])) { +corrupt: + mem_heap_free(heap); + return(FALSE); + } + + b0 = block[0]; + b2 = block[2]; + + b0 = row_merge_read_rec(&block[0], &buf[0], b0, index, file->fd, + foffs0, &mrec0, offsets0); + if (UNIV_UNLIKELY(!b0 && mrec0)) { + + goto corrupt; + } + + if (mrec0) { + /* append all mrec0 to output */ + for (;;) { + ROW_MERGE_WRITE_GET_NEXT(0, goto done0); + } + } +done0: + + /* The file offset points to the beginning of the last page + that has been read. Update it to point to the next block. */ + (*foffs0)++; + + mem_heap_free(heap); + return(row_merge_write_eof(&block[2], b2, of->fd, &of->offset) + != NULL); +} + +/*************************************************************//** Merge disk files. @return DB_SUCCESS or error code */ -static +static __attribute__((nonnull)) ulint row_merge( /*======*/ + trx_t* trx, /*!< in: transaction */ const dict_index_t* index, /*!< in: index being created */ merge_file_t* file, /*!< in/out: file containing index entries */ - ulint half, /*!< in: half the file */ + ulint* half, /*!< in/out: half the file */ row_merge_block_t* block, /*!< in/out: 3 buffers */ int* tmpfd, /*!< in/out: temporary file handle */ TABLE* table) /*!< in/out: MySQL table, for @@ -1458,43 +1578,87 @@ row_merge( ulint foffs1; /*!< second input offset */ ulint error; /*!< error code */ merge_file_t of; /*!< output file */ + const ulint ihalf = *half; + /*!< half the input file */ + ulint ohalf; /*!< half the output file */ UNIV_MEM_ASSERT_W(block[0], 3 * sizeof block[0]); - ut_ad(half > 0); + ut_ad(ihalf < file->offset); of.fd = *tmpfd; of.offset = 0; + of.n_rec = 0; /* Merge blocks to the output file. */ + ohalf = 0; foffs0 = 0; - foffs1 = half; + foffs1 = ihalf; + + for (; foffs0 < ihalf && foffs1 < file->offset; foffs0++, foffs1++) { + ulint ahalf; /*!< arithmetic half the input file */ + + if (UNIV_UNLIKELY(trx_is_interrupted(trx))) { + return(DB_INTERRUPTED); + } - for (; foffs0 < half && foffs1 < file->offset; foffs0++, foffs1++) { error = row_merge_blocks(index, file, block, &foffs0, &foffs1, &of, table); if (error != DB_SUCCESS) { return(error); } + + /* Record the offset of the output file when + approximately half the output has been generated. In + this way, the next invocation of row_merge() will + spend most of the time in this loop. The initial + estimate is ohalf==0. */ + ahalf = file->offset / 2; + ut_ad(ohalf <= of.offset); + + /* Improve the estimate until reaching half the input + file size, or we can not get any closer to it. All + comparands should be non-negative when !(ohalf < ahalf) + because ohalf <= of.offset. */ + if (ohalf < ahalf || of.offset - ahalf < ohalf - ahalf) { + ohalf = of.offset; + } } - /* Copy the last block, if there is one. */ - while (foffs0 < half) { - if (!row_merge_read(file->fd, foffs0++, block) - || !row_merge_write(of.fd, of.offset++, block)) { + /* Copy the last blocks, if there are any. */ + + while (foffs0 < ihalf) { + if (UNIV_UNLIKELY(trx_is_interrupted(trx))) { + return(DB_INTERRUPTED); + } + + if (!row_merge_blocks_copy(index, file, block, &foffs0, &of)) { return(DB_CORRUPTION); } } + + ut_ad(foffs0 == ihalf); + while (foffs1 < file->offset) { - if (!row_merge_read(file->fd, foffs1++, block) - || !row_merge_write(of.fd, of.offset++, block)) { + if (UNIV_UNLIKELY(trx_is_interrupted(trx))) { + return(DB_INTERRUPTED); + } + + if (!row_merge_blocks_copy(index, file, block, &foffs1, &of)) { return(DB_CORRUPTION); } } + ut_ad(foffs1 == file->offset); + + if (UNIV_UNLIKELY(of.n_rec != file->n_rec)) { + return(DB_CORRUPTION); + } + /* Swap file descriptors for the next pass. */ *tmpfd = file->fd; *file = of; + *half = ohalf; UNIV_MEM_INVALID(block[0], 3 * sizeof block[0]); @@ -1508,6 +1672,7 @@ static ulint row_merge_sort( /*===========*/ + trx_t* trx, /*!< in: transaction */ const dict_index_t* index, /*!< in: index being created */ merge_file_t* file, /*!< in/out: file containing index entries */ @@ -1517,20 +1682,26 @@ row_merge_sort( reporting erroneous key value if applicable */ { - ulint blksz; /*!< block size */ + ulint half = file->offset / 2; + + /* The file should always contain at least one byte (the end + of file marker). Thus, it must be at least one block. */ + ut_ad(file->offset > 0); - for (blksz = 1; blksz < file->offset; blksz *= 2) { - ulint half; + do { ulint error; - ut_ad(ut_is_2pow(blksz)); - half = ut_2pow_round((file->offset + (blksz - 1)) / 2, blksz); - error = row_merge(index, file, half, block, tmpfd, table); + error = row_merge(trx, index, file, &half, + block, tmpfd, table); if (error != DB_SUCCESS) { return(error); } - } + + /* half > 0 should hold except when the file consists + of one block. No need to merge further then. */ + ut_ad(half > 0 || file->offset == 1); + } while (half < file->offset && half > 0); return(DB_SUCCESS); } @@ -1797,7 +1968,15 @@ row_merge_drop_index( static const char str1[] = "PROCEDURE DROP_INDEX_PROC () IS\n" "BEGIN\n" + /* Rename the index, so that it will be dropped by + row_merge_drop_temp_indexes() at crash recovery + if the server crashes before this trx is committed. */ + "UPDATE SYS_INDEXES SET NAME=CONCAT('" + TEMP_INDEX_PREFIX_STR "', NAME) WHERE ID = :indexid;\n" + "COMMIT WORK;\n" + /* Drop the field definitions of the index. */ "DELETE FROM SYS_FIELDS WHERE INDEX_ID = :indexid;\n" + /* Drop the index definition and the B-tree. */ "DELETE FROM SYS_INDEXES WHERE ID = :indexid\n" " AND TABLE_ID = :tableid;\n" "END;\n"; @@ -1909,6 +2088,7 @@ row_merge_file_create( { merge_file->fd = innobase_mysql_tmpfile(); merge_file->offset = 0; + merge_file->n_rec = 0; } /*********************************************************************//** @@ -2129,7 +2309,7 @@ row_merge_rename_tables( if (err != DB_SUCCESS) { err_exit: trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; } @@ -2331,7 +2511,7 @@ row_merge_build_indexes( sorting and inserting. */ for (i = 0; i < n_indexes; i++) { - error = row_merge_sort(indexes[i], &merge_files[i], + error = row_merge_sort(trx, indexes[i], &merge_files[i], block, &tmpfd, table); if (error == DB_SUCCESS) { diff --git a/storage/xtradb/row/row0mysql.c b/storage/xtradb/row/row0mysql.c index 25946399fb6..8585b816911 100644 --- a/storage/xtradb/row/row0mysql.c +++ b/storage/xtradb/row/row0mysql.c @@ -510,7 +510,7 @@ handle_new_error: switch (err) { case DB_LOCK_WAIT_TIMEOUT: if (row_rollback_on_timeout) { - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); break; } /* fall through */ @@ -526,7 +526,7 @@ handle_new_error: /* Roll back the latest, possibly incomplete insertion or update */ - trx_general_rollback_for_mysql(trx, TRUE, savept); + trx_general_rollback_for_mysql(trx, savept); } /* MySQL will roll back the latest SQL statement */ break; @@ -548,7 +548,7 @@ handle_new_error: /* Roll back the whole transaction; this resolution was added to version 3.23.43 */ - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); break; case DB_MUST_GET_MORE_FILE_SPACE: @@ -869,18 +869,22 @@ row_update_statistics_if_needed( } /*********************************************************************//** -Unlocks AUTO_INC type locks that were possibly reserved by a trx. */ +Unlocks AUTO_INC type locks that were possibly reserved by a trx. This +function should be called at the the end of an SQL statement, by the +connection thread that owns the transaction (trx->mysql_thd). */ UNIV_INTERN void row_unlock_table_autoinc_for_mysql( /*===============================*/ trx_t* trx) /*!< in/out: transaction */ { - mutex_enter(&kernel_mutex); + if (lock_trx_holds_autoinc_locks(trx)) { + mutex_enter(&kernel_mutex); - lock_release_autoinc_locks(trx); + lock_release_autoinc_locks(trx); - mutex_exit(&kernel_mutex); + mutex_exit(&kernel_mutex); + } } /*********************************************************************//** @@ -1770,7 +1774,6 @@ row_create_table_for_mysql( const char* table_name; ulint table_name_len; ulint err; - ulint i; ut_ad(trx->mysql_thread_id == os_thread_get_curr_id()); #ifdef UNIV_SYNC_DEBUG @@ -1805,15 +1808,6 @@ err_exit: goto err_exit; } - /* Check that no reserved column names are used. */ - for (i = 0; i < dict_table_get_n_user_cols(table); i++) { - if (dict_col_name_is_reserved( - dict_table_get_col_name(table, i))) { - - goto err_exit; - } - } - trx_start_if_not_started(trx); /* The table name is prefixed with the database name and a '/'. @@ -1888,7 +1882,9 @@ err_exit: if (UNIV_UNLIKELY(err != DB_SUCCESS)) { trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); + /* TO DO: free table? The code below will dereference + table->name, though. */ } switch (err) { @@ -1907,31 +1903,6 @@ err_exit: break; case DB_DUPLICATE_KEY: - ut_print_timestamp(stderr); - fputs(" InnoDB: Error: table ", stderr); - ut_print_name(stderr, trx, TRUE, table->name); - fputs(" already exists in InnoDB internal\n" - "InnoDB: data dictionary. Have you deleted" - " the .frm file\n" - "InnoDB: and not used DROP TABLE?" - " Have you used DROP DATABASE\n" - "InnoDB: for InnoDB tables in" - " MySQL version <= 3.23.43?\n" - "InnoDB: See the Restrictions section" - " of the InnoDB manual.\n" - "InnoDB: You can drop the orphaned table" - " inside InnoDB by\n" - "InnoDB: creating an InnoDB table with" - " the same name in another\n" - "InnoDB: database and copying the .frm file" - " to the current database.\n" - "InnoDB: Then MySQL thinks the table exists," - " and DROP TABLE will\n" - "InnoDB: succeed.\n" - "InnoDB: You can look for further help from\n" - "InnoDB: " REFMAN "innodb-troubleshooting.html\n", - stderr); - /* We may also get err == DB_ERROR if the .ibd file for the table already exists */ @@ -2056,7 +2027,7 @@ error_handling: trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); row_drop_table_for_mysql(table_name, trx, FALSE); @@ -2077,7 +2048,7 @@ Scans a table create SQL string and adds to the data dictionary the foreign key constraints declared in the string. This function should be called after the indexes for a table have been created. Each foreign key constraint must be accompanied with indexes in -bot participating tables. The indexes are allowed to contain more +both participating tables. The indexes are allowed to contain more fields than mentioned in the constraint. Check also that foreign key constraints which reference this table are ok. @return error code or DB_SUCCESS */ @@ -2124,7 +2095,7 @@ row_table_add_foreign_constraints( trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); row_drop_table_for_mysql(name, trx, FALSE); @@ -2491,7 +2462,7 @@ row_discard_tablespace_for_mysql( if (err != DB_SUCCESS) { trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; } else { dict_table_change_id_in_cache(table, new_id); @@ -2500,7 +2471,7 @@ row_discard_tablespace_for_mysql( if (!success) { trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; err = DB_ERROR; @@ -2952,7 +2923,7 @@ next_rec: if (err != DB_SUCCESS) { trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; ut_print_timestamp(stderr); fputs(" InnoDB: Unable to assign a new identifier to table ", @@ -3593,7 +3564,7 @@ row_delete_constraint( if ((err == DB_SUCCESS) && !strchr(id, '/')) { /* Old format < 4.0.18 constraints have constraint ids - <number>_<number>. We only try deleting them if the + NUMBER_NUMBER. We only try deleting them if the constraint name does not contain a '/' character, otherwise deleting a new format constraint named 'foo/bar' from database 'baz' would remove constraint 'bar' from database @@ -3857,7 +3828,7 @@ end: "InnoDB: succeed.\n", stderr); } trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; } else { /* The following call will also rename the .ibd data file if @@ -3866,7 +3837,7 @@ end: if (!dict_table_rename_in_cache(table, new_name, !new_is_tmp)) { trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; goto funct_exit; } @@ -3906,7 +3877,7 @@ end: ut_a(dict_table_rename_in_cache(table, old_name, FALSE)); trx->error_state = DB_SUCCESS; - trx_general_rollback_for_mysql(trx, FALSE, NULL); + trx_general_rollback_for_mysql(trx, NULL); trx->error_state = DB_SUCCESS; } } @@ -4166,6 +4137,7 @@ row_check_table_for_mysql( } if (trx_is_interrupted(prebuilt->trx)) { + ret = DB_INTERRUPTED; break; } diff --git a/storage/xtradb/scripts/install_innodb_plugins.sql b/storage/xtradb/scripts/install_innodb_plugins.sql index b2e5e3f6de0..5a555a652f7 100644 --- a/storage/xtradb/scripts/install_innodb_plugins.sql +++ b/storage/xtradb/scripts/install_innodb_plugins.sql @@ -14,3 +14,4 @@ INSTALL PLUGIN INNODB_BUFFER_POOL_PAGES_INDEX SONAME 'ha_innodb.so'; INSTALL PLUGIN innodb_rseg SONAME 'ha_innodb.so'; INSTALL PLUGIN innodb_table_stats SONAME 'ha_innodb.so'; INSTALL PLUGIN innodb_index_stats SONAME 'ha_innodb.so'; +INSTALL PLUGIN xtradb_admin_command SONAME 'ha_innodb.so'; diff --git a/storage/xtradb/scripts/install_innodb_plugins_win.sql b/storage/xtradb/scripts/install_innodb_plugins_win.sql index 8c94b4e240d..7cda3335694 100644 --- a/storage/xtradb/scripts/install_innodb_plugins_win.sql +++ b/storage/xtradb/scripts/install_innodb_plugins_win.sql @@ -7,3 +7,11 @@ INSTALL PLUGIN innodb_cmp SONAME 'ha_innodb.dll'; INSTALL PLUGIN innodb_cmp_reset SONAME 'ha_innodb.dll';
INSTALL PLUGIN innodb_cmpmem SONAME 'ha_innodb.dll';
INSTALL PLUGIN innodb_cmpmem_reset SONAME 'ha_innodb.dll';
+INSTALL PLUGIN XTRADB_ENHANCEMENTS SONAME 'ha_innodb.dll';
+INSTALL PLUGIN INNODB_BUFFER_POOL_PAGES SONAME 'ha_innodb.dll';
+INSTALL PLUGIN INNODB_BUFFER_POOL_PAGES_BLOB SONAME 'ha_innodb.dll';
+INSTALL PLUGIN INNODB_BUFFER_POOL_PAGES_INDEX SONAME 'ha_innodb.dll';
+INSTALL PLUGIN innodb_rseg SONAME 'ha_innodb.dll';
+INSTALL PLUGIN innodb_table_stats SONAME 'ha_innodb.dll';
+INSTALL PLUGIN innodb_index_stats SONAME 'ha_innodb.dll';
+INSTALL PLUGIN xtradb_admin_command SONAME 'ha_innodb.dll';
diff --git a/storage/xtradb/srv/srv0srv.c b/storage/xtradb/srv/srv0srv.c index 76f3fc5097d..38176938e62 100644 --- a/storage/xtradb/srv/srv0srv.c +++ b/storage/xtradb/srv/srv0srv.c @@ -167,7 +167,7 @@ UNIV_INTERN ibool srv_extra_undoslots = FALSE; UNIV_INTERN ibool srv_fast_recovery = FALSE; UNIV_INTERN ibool srv_recovery_stats = FALSE; -UNIV_INTERN ibool srv_use_purge_thread = FALSE; +UNIV_INTERN ulint srv_use_purge_thread = 0; /* if TRUE, then we auto-extend the last data file */ UNIV_INTERN ibool srv_auto_extend_last_data_file = FALSE; @@ -307,12 +307,6 @@ UNIV_INTERN ulint srv_buf_pool_flushed = 0; reading of a disk page */ UNIV_INTERN ulint srv_buf_pool_reads = 0; -/** Number of sequential read-aheads */ -UNIV_INTERN ulint srv_read_ahead_seq = 0; - -/** Number of random read-aheads */ -UNIV_INTERN ulint srv_read_ahead_rnd = 0; - /* structure to pass status variables to MySQL */ UNIV_INTERN export_struc export_vars; @@ -403,6 +397,7 @@ UNIV_INTERN ulong srv_ibuf_active_contract = 0; /* 0:disable 1:enable */ UNIV_INTERN ulong srv_ibuf_accel_rate = 100; #define PCT_IBUF_IO(pct) ((ulint) (srv_io_capacity * srv_ibuf_accel_rate * ((double) pct / 10000.0))) +UNIV_INTERN ulint srv_checkpoint_age_target = 0; UNIV_INTERN ulong srv_flush_neighbor_pages = 1; /* 0:disable 1:enable */ UNIV_INTERN ulong srv_enable_unsafe_group_commit = 0; /* 0:disable 1:enable */ @@ -410,6 +405,7 @@ UNIV_INTERN ulong srv_read_ahead = 3; /* 1: random 2: linear 3: Both */ UNIV_INTERN ulong srv_adaptive_checkpoint = 0; /* 0: none 1: reflex 2: estimate */ UNIV_INTERN ulong srv_expand_import = 0; /* 0:disable 1:enable */ +UNIV_INTERN ulint srv_relax_table_creation = 0; /* 0:disable 1:enable */ UNIV_INTERN ulong srv_extra_rsegments = 0; /* extra rseg for users */ UNIV_INTERN ulong srv_dict_size_limit = 0; @@ -498,8 +494,6 @@ static ulint srv_main_background_loops = 0; static ulint srv_main_flush_loops = 0; /* Log writes involving flush. */ static ulint srv_log_writes_and_flush = 0; -/* Log writes not including flush. */ -static ulint srv_log_buffer_writes = 0; /* This is only ever touched by the master thread. It records the time when the last flush of log file has happened. The master @@ -648,7 +642,7 @@ future, but at the moment we plan to implement a more coarse solution, which could be called a global priority inheritance. If a thread has to wait for a long time, say 300 milliseconds, for a resource, we just guess that it may be waiting for a resource owned by a background -thread, and boost the the priority of all runnable background threads +thread, and boost the priority of all runnable background threads to the normal level. The background threads then themselves adjust their fixed priority back to background after releasing all resources they had (or, at some fixed points in their program code). @@ -748,9 +742,8 @@ srv_print_master_thread_info( srv_main_1_second_loops, srv_main_sleeps, srv_main_10_second_loops, srv_main_background_loops, srv_main_flush_loops); - fprintf(file, "srv_master_thread log flush and writes: %lu " - " log writes only: %lu\n", - srv_log_writes_and_flush, srv_log_buffer_writes); + fprintf(file, "srv_master_thread log flush and writes: %lu\n", + srv_log_writes_and_flush); } /*********************************************************************//** @@ -1048,13 +1041,26 @@ srv_init(void) } /*********************************************************************//** -Frees the OS fast mutex created in srv_init(). */ +Frees the data structures created in srv_init(). */ UNIV_INTERN void srv_free(void) /*==========*/ { os_fast_mutex_free(&srv_conc_mutex); + mem_free(srv_conc_slots); + srv_conc_slots = NULL; + + mem_free(srv_sys->threads); + mem_free(srv_sys); + srv_sys = NULL; + + mem_free(kernel_mutex_temp); + kernel_mutex_temp = NULL; + mem_free(srv_mysql_table); + srv_mysql_table = NULL; + + trx_i_s_cache_free(trx_i_s_cache); } /*********************************************************************//** @@ -1066,6 +1072,8 @@ srv_general_init(void) /*==================*/ { ut_mem_init(); + /* Reset the system variables in the recovery module. */ + recv_sys_var_init(); os_sync_init(); sync_init(); mem_init(srv_mem_pool_size); @@ -1159,6 +1167,10 @@ srv_conc_enter_innodb( ibool has_slept = FALSE; srv_conc_slot_t* slot = NULL; ulint i; + ib_uint64_t start_time = 0L; + ib_uint64_t finish_time = 0L; + ulint sec; + ulint ms; if (trx->mysql_thd != NULL && thd_is_replication_slave_thread(trx->mysql_thd)) { @@ -1235,6 +1247,7 @@ retry: switches. */ if (SRV_THREAD_SLEEP_DELAY > 0) { os_thread_sleep(SRV_THREAD_SLEEP_DELAY); + trx->innodb_que_wait_timer += SRV_THREAD_SLEEP_DELAY; } trx->op_info = ""; @@ -1290,12 +1303,25 @@ retry: /* Go to wait for the event; when a thread leaves InnoDB it will release this thread */ + if (innobase_get_slow_log() && trx->take_stats) { + ut_usectime(&sec, &ms); + start_time = (ib_uint64_t)sec * 1000000 + ms; + } else { + start_time = 0; + } + trx->op_info = "waiting in InnoDB queue"; os_event_wait(slot->event); trx->op_info = ""; + if (innobase_get_slow_log() && trx->take_stats && start_time) { + ut_usectime(&sec, &ms); + finish_time = (ib_uint64_t)sec * 1000000 + ms; + trx->innodb_que_wait_timer += (ulint)(finish_time - start_time); + } + os_fast_mutex_lock(&srv_conc_mutex); srv_conc_n_waiting_threads--; @@ -2085,14 +2111,16 @@ srv_export_innodb_status(void) export_vars.innodb_data_writes = os_n_file_writes; export_vars.innodb_data_written = srv_data_written; export_vars.innodb_dict_tables= (dict_sys ? UT_LIST_GET_LEN(dict_sys->table_LRU) : 0); - export_vars.innodb_buffer_pool_read_requests = buf_pool->n_page_gets; + export_vars.innodb_buffer_pool_read_requests = buf_pool->stat.n_page_gets; export_vars.innodb_buffer_pool_write_requests = srv_buf_pool_write_requests; export_vars.innodb_buffer_pool_wait_free = srv_buf_pool_wait_free; export_vars.innodb_buffer_pool_pages_flushed = srv_buf_pool_flushed; export_vars.innodb_buffer_pool_reads = srv_buf_pool_reads; - export_vars.innodb_buffer_pool_read_ahead_rnd = srv_read_ahead_rnd; - export_vars.innodb_buffer_pool_read_ahead_seq = srv_read_ahead_seq; + export_vars.innodb_buffer_pool_read_ahead + = buf_pool->stat.n_ra_pages_read; + export_vars.innodb_buffer_pool_read_ahead_evicted + = buf_pool->stat.n_ra_pages_evicted; export_vars.innodb_buffer_pool_pages_data = UT_LIST_GET_LEN(buf_pool->LRU); export_vars.innodb_buffer_pool_pages_dirty @@ -2123,9 +2151,9 @@ srv_export_innodb_status(void) export_vars.innodb_log_writes = srv_log_writes; export_vars.innodb_dblwr_pages_written = srv_dblwr_pages_written; export_vars.innodb_dblwr_writes = srv_dblwr_writes; - export_vars.innodb_pages_created = buf_pool->n_pages_created; - export_vars.innodb_pages_read = buf_pool->n_pages_read; - export_vars.innodb_pages_written = buf_pool->n_pages_written; + export_vars.innodb_pages_created = buf_pool->stat.n_pages_created; + export_vars.innodb_pages_read = buf_pool->stat.n_pages_read; + export_vars.innodb_pages_written = buf_pool->stat.n_pages_written; export_vars.innodb_row_lock_waits = srv_n_lock_wait_count; export_vars.innodb_row_lock_current_waits = srv_n_lock_wait_current_count; @@ -2492,12 +2520,6 @@ srv_sync_log_buffer_in_background(void) log_buffer_sync_in_background(TRUE); srv_last_log_flush_time = current_time; srv_log_writes_and_flush++; - } else { - /* Actually we don't need to write logs here. - We are just being extra safe here by forcing - the log buffer to log file. */ - log_buffer_sync_in_background(FALSE); - srv_log_buffer_writes++; } } @@ -2555,8 +2577,8 @@ loop: srv_main_thread_op_info = "reserving kernel mutex"; - n_ios_very_old = log_sys->n_log_ios + buf_pool->n_pages_read - + buf_pool->n_pages_written; + n_ios_very_old = log_sys->n_log_ios + buf_pool->stat.n_pages_read + + buf_pool->stat.n_pages_written; mutex_enter(&kernel_mutex); /* Store the user activity counter at the start of this loop */ @@ -2576,8 +2598,8 @@ loop: skip_sleep = FALSE; for (i = 0; i < 10; i++) { - n_ios_old = log_sys->n_log_ios + buf_pool->n_pages_read - + buf_pool->n_pages_written; + n_ios_old = log_sys->n_log_ios + buf_pool->stat.n_pages_read + + buf_pool->stat.n_pages_written; srv_main_thread_op_info = "sleeping"; srv_main_1_second_loops++; @@ -2629,8 +2651,8 @@ loop: n_pend_ios = buf_get_n_pending_ios() + log_sys->n_pending_writes; - n_ios = log_sys->n_log_ios + buf_pool->n_pages_read - + buf_pool->n_pages_written; + n_ios = log_sys->n_log_ios + buf_pool->stat.n_pages_read + + buf_pool->stat.n_pages_written; if (n_pend_ios < SRV_PEND_IO_THRESHOLD && (n_ios - n_ios_old < SRV_RECENT_IO_ACTIVITY)) { srv_main_thread_op_info = "doing insert buffer merge"; @@ -2646,6 +2668,8 @@ loop: /* Try to keep the number of modified pages in the buffer pool under the limit wished by the user */ + srv_main_thread_op_info = + "flushing buffer pool pages"; n_pages_flushed = buf_flush_batch(BUF_FLUSH_LIST, PCT_IO(100), IB_ULONGLONG_MAX); @@ -2668,6 +2692,8 @@ loop: ulint n_flush = buf_flush_get_desired_flush_rate(); if (n_flush) { + srv_main_thread_op_info = + "flushing buffer pool pages"; n_flush = ut_min(PCT_IO(100), n_flush); n_pages_flushed = buf_flush_batch( @@ -2839,8 +2865,8 @@ retry_flush_batch: are not required, and may be disabled. */ n_pend_ios = buf_get_n_pending_ios() + log_sys->n_pending_writes; - n_ios = log_sys->n_log_ios + buf_pool->n_pages_read - + buf_pool->n_pages_written; + n_ios = log_sys->n_log_ios + buf_pool->stat.n_pages_read + + buf_pool->stat.n_pages_written; srv_main_10_second_loops++; if (n_pend_ios < SRV_PEND_IO_THRESHOLD @@ -3134,6 +3160,7 @@ srv_purge_thread( ulint n_pages_purged_sum = 1; /* dummy */ ulint history_len; ulint sleep_ms= 10000; /* initial: 10 sec. */ + ibool can_be_last = FALSE; #ifdef UNIV_DEBUG_THREAD_CREATION fprintf(stderr, "Purge thread starts, id %lu\n", @@ -3146,8 +3173,21 @@ srv_purge_thread( mutex_exit(&kernel_mutex); loop: - if (srv_fast_shutdown && srv_shutdown_state > 0) { - goto exit_func; + if (srv_shutdown_state > 0) { + if (srv_fast_shutdown) { + /* someone other should wait the end of the workers */ + goto exit_func; + } + + mutex_enter(&kernel_mutex); + if (srv_n_threads_active[SRV_PURGE_WORKER]) { + can_be_last = FALSE; + } else { + can_be_last = TRUE; + } + mutex_exit(&kernel_mutex); + + sleep_ms = 10; } os_thread_sleep( sleep_ms * 1000 ); @@ -3168,6 +3208,15 @@ loop: n_pages_purged_sum += n_pages_purged; } while (n_pages_purged); + if (srv_shutdown_state > 0 && can_be_last) { + /* the last trx_purge() is executed without workers */ + goto exit_func; + } + + if (n_pages_purged_sum) { + srv_active_wake_master_thread(); + } + if (n_pages_purged_sum == 0) sleep_ms *= 10; if (sleep_ms > 10000) @@ -3176,9 +3225,62 @@ loop: goto loop; exit_func: - /* We count the number of threads in os_thread_exit(). A created - thread should always use that to exit and not use return() to exit. */ + trx_purge_worker_wake(); /* It may not make sense. for safety only */ + + /* wake master thread to flush the pages */ + srv_wake_master_thread(); + + mutex_enter(&kernel_mutex); + srv_n_threads_active[SRV_PURGE]--; + mutex_exit(&kernel_mutex); + os_thread_exit(NULL); + OS_THREAD_DUMMY_RETURN; +} + +/************************************************************************* +A thread which is devoted to purge, for take over the master thread's +purging */ +UNIV_INTERN +os_thread_ret_t +srv_purge_worker_thread( +/*====================*/ + void* arg) +{ + ulint worker_id; /* index for array */ + + worker_id = *((ulint*)arg); + +#ifdef UNIV_DEBUG_THREAD_CREATION + fprintf(stderr, "Purge worker thread starts, id %lu\n", + os_thread_pf(os_thread_get_curr_id())); +#endif + srv_table_reserve_slot(SRV_PURGE_WORKER); + mutex_enter(&kernel_mutex); + srv_n_threads_active[SRV_PURGE_WORKER]++; + mutex_exit(&kernel_mutex); + +loop: + /* purge worker threads only works when srv_shutdown_state==0 */ + /* for safety and exactness. */ + if (srv_shutdown_state > 0) { + goto exit_func; + } + + trx_purge_worker_wait(); + + if (srv_shutdown_state > 0) { + goto exit_func; + } + + trx_purge_worker(worker_id); + + goto loop; + +exit_func: + mutex_enter(&kernel_mutex); + srv_n_threads_active[SRV_PURGE_WORKER]--; + mutex_exit(&kernel_mutex); os_thread_exit(NULL); OS_THREAD_DUMMY_RETURN; diff --git a/storage/xtradb/srv/srv0start.c b/storage/xtradb/srv/srv0start.c index e8a9dabdc41..5b5599b893d 100644 --- a/storage/xtradb/srv/srv0start.c +++ b/storage/xtradb/srv/srv0start.c @@ -103,6 +103,8 @@ Created 2/16/1996 Heikki Tuuri # include "row0row.h" # include "row0mysql.h" # include "btr0pcur.h" +# include "thr0loc.h" +# include "os0sync.h" /* for INNODB_RW_LOCKS_USE_ATOMICS */ /** Log sequence number immediately after startup */ UNIV_INTERN ib_uint64_t srv_start_lsn; @@ -141,9 +143,9 @@ static mutex_t ios_mutex; static ulint ios; /** io_handler_thread parameters for thread identification */ -static ulint n[SRV_MAX_N_IO_THREADS + 5]; +static ulint n[SRV_MAX_N_IO_THREADS + 5 + 64]; /** io_handler_thread identifiers */ -static os_thread_id_t thread_ids[SRV_MAX_N_IO_THREADS + 5]; +static os_thread_id_t thread_ids[SRV_MAX_N_IO_THREADS + 5 + 64]; /** We use this mutex to test the return value of pthread_mutex_trylock on successful locking. HP-UX does NOT return 0, though Linux et al do. */ @@ -494,6 +496,8 @@ io_handler_thread( mutex_exit(&ios_mutex); } + thr_local_free(os_thread_get_curr_id()); + /* We count the number of threads in os_thread_exit(). A created thread should always use that to exit and not use return() to exit. The thread actually never comes here because it is exited in an @@ -530,32 +534,6 @@ srv_normalize_path_for_win( #endif } -/*********************************************************************//** -Adds a slash or a backslash to the end of a string if it is missing -and the string is not empty. -@return string which has the separator if the string is not empty */ -UNIV_INTERN -char* -srv_add_path_separator_if_needed( -/*=============================*/ - char* str) /*!< in: null-terminated character string */ -{ - char* out_str; - ulint len = ut_strlen(str); - - if (len == 0 || str[len - 1] == SRV_PATH_SEPARATOR) { - - return(str); - } - - out_str = ut_malloc(len + 2); - memcpy(out_str, str, len); - out_str[len] = SRV_PATH_SEPARATOR; - out_str[len + 1] = 0; - - return(out_str); -} - #ifndef UNIV_HOTBACKUP /*********************************************************************//** Calculates the low 32 bits when a file size which is given as a number @@ -604,19 +582,24 @@ open_or_create_log_file( ulint size; ulint size_high; char name[10000]; + ulint dirnamelen; UT_NOT_USED(create_new_db); *log_file_created = FALSE; srv_normalize_path_for_win(srv_log_group_home_dirs[k]); - srv_log_group_home_dirs[k] = srv_add_path_separator_if_needed( - srv_log_group_home_dirs[k]); - ut_a(strlen(srv_log_group_home_dirs[k]) - < (sizeof name) - 10 - sizeof "ib_logfile"); - sprintf(name, "%s%s%lu", srv_log_group_home_dirs[k], - "ib_logfile", (ulong) i); + dirnamelen = strlen(srv_log_group_home_dirs[k]); + ut_a(dirnamelen < (sizeof name) - 10 - sizeof "ib_logfile"); + memcpy(name, srv_log_group_home_dirs[k], dirnamelen); + + /* Add a path separator if needed. */ + if (dirnamelen && name[dirnamelen - 1] != SRV_PATH_SEPARATOR) { + name[dirnamelen++] = SRV_PATH_SEPARATOR; + } + + sprintf(name + dirnamelen, "%s%lu", "ib_logfile", (ulong) i); files[i] = os_file_create(name, OS_FILE_CREATE, OS_FILE_NORMAL, OS_LOG_FILE, &ret); @@ -779,14 +762,22 @@ open_or_create_data_files( *create_new_db = FALSE; srv_normalize_path_for_win(srv_data_home); - srv_data_home = srv_add_path_separator_if_needed(srv_data_home); for (i = 0; i < srv_n_data_files; i++) { + ulint dirnamelen; + srv_normalize_path_for_win(srv_data_file_names[i]); + dirnamelen = strlen(srv_data_home); - ut_a(strlen(srv_data_home) + strlen(srv_data_file_names[i]) + ut_a(dirnamelen + strlen(srv_data_file_names[i]) < (sizeof name) - 1); - sprintf(name, "%s%s", srv_data_home, srv_data_file_names[i]); + memcpy(name, srv_data_home, dirnamelen); + /* Add a path separator if needed. */ + if (dirnamelen && name[dirnamelen - 1] != SRV_PATH_SEPARATOR) { + name[dirnamelen++] = SRV_PATH_SEPARATOR; + } + + strcpy(name + dirnamelen, srv_data_file_names[i]); if (srv_data_file_is_raw_partition[i] == 0) { @@ -1008,7 +999,7 @@ skip_size_check: return(DB_SUCCESS); } -/****************************************************************//** +/******************************************************************** Starts InnoDB and creates a new database if database files are not found and the user wants. @return DB_SUCCESS or error code */ @@ -1096,6 +1087,10 @@ innobase_start_or_create_for_mysql(void) "InnoDB: !!!!!!!! UNIV_SEARCH_DEBUG switched on !!!!!!!!!\n"); #endif +#ifdef UNIV_LOG_LSN_DEBUG + fprintf(stderr, + "InnoDB: !!!!!!!! UNIV_LOG_LSN_DEBUG switched on !!!!!!!!!\n"); +#endif /* UNIV_LOG_LSN_DEBUG */ #ifdef UNIV_MEM_DEBUG fprintf(stderr, "InnoDB: !!!!!!!! UNIV_MEM_DEBUG switched on !!!!!!!!!\n"); @@ -1106,34 +1101,7 @@ innobase_start_or_create_for_mysql(void) "InnoDB: The InnoDB memory heap is disabled\n"); } -#ifdef HAVE_GCC_ATOMIC_BUILTINS -# ifdef INNODB_RW_LOCKS_USE_ATOMICS - fprintf(stderr, - "InnoDB: Mutexes and rw_locks use GCC atomic builtins.\n"); -# else /* INNODB_RW_LOCKS_USE_ATOMICS */ - fprintf(stderr, - "InnoDB: Mutexes use GCC atomic builtins, rw_locks do not.\n"); -# endif /* INNODB_RW_LOCKS_USE_ATOMICS */ -#elif defined(HAVE_SOLARIS_ATOMICS) -# ifdef INNODB_RW_LOCKS_USE_ATOMICS - fprintf(stderr, - "InnoDB: Mutexes and rw_locks use Solaris atomic functions.\n"); -# else - fprintf(stderr, - "InnoDB: Mutexes use Solaris atomic functions.\n"); -# endif /* INNODB_RW_LOCKS_USE_ATOMICS */ -#elif HAVE_WINDOWS_ATOMICS -# ifdef INNODB_RW_LOCKS_USE_ATOMICS - fprintf(stderr, - "InnoDB: Mutexes and rw_locks use Windows interlocked functions.\n"); -# else - fprintf(stderr, - "InnoDB: Mutexes use Windows interlocked functions.\n"); -# endif /* INNODB_RW_LOCKS_USE_ATOMICS */ -#else /* HAVE_GCC_ATOMIC_BUILTINS */ - fprintf(stderr, - "InnoDB: Neither mutexes nor rw_locks use GCC atomic builtins.\n"); -#endif /* HAVE_GCC_ATOMIC_BUILTINS */ + fprintf(stderr, "InnoDB: %s\n", IB_ATOMICS_STARTUP_MSG); /* Since InnoDB does not currently clean up all its internal data structures in MySQL Embedded Server Library server_end(), we @@ -1142,7 +1110,7 @@ innobase_start_or_create_for_mysql(void) if (srv_start_has_been_called) { fprintf(stderr, - "InnoDB: Error:startup called second time" + "InnoDB: Error: startup called second time" " during the process lifetime.\n" "InnoDB: In the MySQL Embedded Server Library" " you cannot call server_init()\n" @@ -1409,7 +1377,7 @@ innobase_start_or_create_for_mysql(void) sum_of_new_sizes += srv_data_file_sizes[i]; } - if (sum_of_new_sizes < 640) { + if (sum_of_new_sizes < 10485760 / UNIV_PAGE_SIZE) { fprintf(stderr, "InnoDB: Error: tablespace size must be" " at least 10 MB\n"); @@ -1739,8 +1707,17 @@ innobase_start_or_create_for_mysql(void) + (1 + SRV_MAX_N_IO_THREADS)); if (srv_use_purge_thread) { + ulint i; + os_thread_create(&srv_purge_thread, NULL, thread_ids + (4 + SRV_MAX_N_IO_THREADS)); + + for (i = 0; i < srv_use_purge_thread - 1; i++) { + n[5 + i + SRV_MAX_N_IO_THREADS] = i; /* using as index for arrays in purge_sys */ + os_thread_create(&srv_purge_worker_thread, + n + (5 + i + SRV_MAX_N_IO_THREADS), + thread_ids + (5 + i + SRV_MAX_N_IO_THREADS)); + } } #ifdef UNIV_DEBUG /* buf_debug_prints = TRUE; */ @@ -1853,7 +1830,7 @@ innobase_start_or_create_for_mysql(void) /* Actually, we did not change the undo log format between 4.0 and 4.1.1, and we would not need to run purge to completion. Note also that the purge algorithm in 4.1.1 - can process the the history list again even after a full + can process the history list again even after a full purge, because our algorithm does not cut the end of the history list in all cases so that it would become empty after a full purge. That mean that we may purge 4.0 type @@ -2005,8 +1982,10 @@ innobase_shutdown_for_mysql(void) /* All the threads have exited or are just exiting; NOTE that the threads may not have completed their exit yet. Should we use pthread_join() to make sure - they have exited? Now we just sleep 0.1 seconds and - hope that is enough! */ + they have exited? If we did, we would have to + remove the pthread_detach() from + os_thread_exit(). Now we just sleep 0.1 + seconds and hope that is enough! */ os_mutex_exit(os_sync_mutex); @@ -2045,37 +2024,41 @@ innobase_shutdown_for_mysql(void) srv_misc_tmpfile = 0; } + /* This must be disabled before closing the buffer pool + and closing the data dictionary. */ + btr_search_disable(); + + ibuf_close(); + log_shutdown(); + lock_sys_close(); + thr_local_close(); trx_sys_file_format_close(); + trx_sys_close(); mutex_free(&srv_monitor_file_mutex); mutex_free(&srv_dict_tmpfile_mutex); mutex_free(&srv_misc_tmpfile_mutex); + dict_close(); + btr_search_sys_free(); /* 3. Free all InnoDB's own mutexes and the os_fast_mutexes inside them */ + os_aio_free(); sync_close(); + srv_free(); + fil_close(); /* 4. Free the os_conc_mutex and all os_events and os_mutexes */ - srv_free(); os_sync_free(); - /* Check that all read views are closed except read view owned - by a purge. */ - - if (UT_LIST_GET_LEN(trx_sys->view_list) > 1) { - fprintf(stderr, - "InnoDB: Error: all read views were not closed" - " before shutdown:\n" - "InnoDB: %lu read views open \n", - UT_LIST_GET_LEN(trx_sys->view_list) - 1); - } - - /* 5. Free all allocated memory and the os_fast_mutex created in - ut0mem.c */ + /* 5. Free all allocated memory */ + pars_lexer_close(); + log_mem_free(); buf_pool_free(); ut_free_all_mem(); + mem_close(); if (os_thread_count != 0 || os_event_count != 0 @@ -2106,6 +2089,7 @@ innobase_shutdown_for_mysql(void) } srv_was_started = FALSE; + srv_start_has_been_called = FALSE; return((int) DB_SUCCESS); } diff --git a/storage/xtradb/sync/sync0arr.c b/storage/xtradb/sync/sync0arr.c index 0519e13dee0..cfa52cdcc88 100644 --- a/storage/xtradb/sync/sync0arr.c +++ b/storage/xtradb/sync/sync0arr.c @@ -227,24 +227,21 @@ sync_array_create( SYNC_ARRAY_MUTEX: determines the type of mutex protecting the data structure */ { + ulint sz; sync_array_t* arr; - sync_cell_t* cell_array; - sync_cell_t* cell; - ulint i; ut_a(n_cells > 0); /* Allocate memory for the data structures */ arr = ut_malloc(sizeof(sync_array_t)); + memset(arr, 0x0, sizeof(*arr)); - cell_array = ut_malloc(sizeof(sync_cell_t) * n_cells); + sz = sizeof(sync_cell_t) * n_cells; + arr->array = ut_malloc(sz); + memset(arr->array, 0x0, sz); arr->n_cells = n_cells; - arr->n_reserved = 0; - arr->array = cell_array; arr->protection = protection; - arr->sg_count = 0; - arr->res_count = 0; /* Then create the mutex to protect the wait array complex */ if (protection == SYNC_ARRAY_OS_MUTEX) { @@ -255,13 +252,6 @@ sync_array_create( ut_error; } - for (i = 0; i < n_cells; i++) { - cell = sync_array_get_nth_cell(arr, i); - cell->wait_object = NULL; - cell->waiting = FALSE; - cell->signal_count = 0; - } - return(arr); } @@ -492,12 +482,12 @@ sync_array_cell_print( mutex = cell->old_wait_mutex; fprintf(file, - "Mutex at %p created file %s line %lu, lock var %lu\n" + "Mutex at %p '%s', lock var %lu\n" #ifdef UNIV_SYNC_DEBUG "Last time reserved in file %s line %lu, " #endif /* UNIV_SYNC_DEBUG */ "waiters flag %lu\n", - (void*) mutex, mutex->cfile_name, (ulong) mutex->cline, + (void*) mutex, mutex->cmutex_name, (ulong) mutex->lock_word, #ifdef UNIV_SYNC_DEBUG mutex->file_name, (ulong) mutex->line, @@ -513,9 +503,8 @@ sync_array_cell_print( rwlock = cell->old_wait_rw_lock; fprintf(file, - " RW-latch at %p created in file %s line %lu\n", - (void*) rwlock, rwlock->cfile_name, - (ulong) rwlock->cline); + " RW-latch at %p '%s'\n", + (void*) rwlock, rwlock->lock_name); writer = rw_lock_get_writer(rwlock); if (writer != RW_LOCK_NOT_LOCKED) { fprintf(file, diff --git a/storage/xtradb/sync/sync0rw.c b/storage/xtradb/sync/sync0rw.c index 0ed114e330c..07eac403dfe 100644 --- a/storage/xtradb/sync/sync0rw.c +++ b/storage/xtradb/sync/sync0rw.c @@ -38,6 +38,7 @@ Created 9/11/1995 Heikki Tuuri #include "os0thread.h" #include "mem0mem.h" #include "srv0srv.h" +#include "os0sync.h" /* for INNODB_RW_LOCKS_USE_ATOMICS */ /* IMPLEMENTATION OF THE RW_LOCK @@ -230,8 +231,8 @@ rw_lock_create_func( # ifdef UNIV_SYNC_DEBUG ulint level, /*!< in: level */ # endif /* UNIV_SYNC_DEBUG */ - const char* cmutex_name, /*!< in: mutex name */ #endif /* UNIV_DEBUG */ + const char* cmutex_name, /*!< in: mutex name */ const char* cfile_name, /*!< in: file name where created */ ulint cline) /*!< in: file line where created */ { @@ -241,14 +242,15 @@ rw_lock_create_func( #ifndef INNODB_RW_LOCKS_USE_ATOMICS mutex_create(rw_lock_get_mutex(lock), SYNC_NO_ORDER_CHECK); - lock->mutex.cfile_name = cfile_name; - lock->mutex.cline = cline; + ut_d(lock->mutex.cfile_name = cfile_name); + ut_d(lock->mutex.cline = cline); - ut_d(lock->mutex.cmutex_name = cmutex_name); + lock->mutex.cmutex_name = cmutex_name; ut_d(lock->mutex.mutex_type = 1); #else /* INNODB_RW_LOCKS_USE_ATOMICS */ # ifdef UNIV_DEBUG - UT_NOT_USED(cmutex_name); + UT_NOT_USED(cfile_name); + UT_NOT_USED(cline); # endif #endif /* INNODB_RW_LOCKS_USE_ATOMICS */ @@ -268,8 +270,7 @@ rw_lock_create_func( lock->magic_n = RW_LOCK_MAGIC_N; - lock->cfile_name = cfile_name; - lock->cline = (unsigned int) cline; + lock->lock_name = cmutex_name; lock->count_os_wait = 0; lock->last_s_file_name = "not yet reserved"; @@ -304,8 +305,6 @@ rw_lock_free( ut_ad(rw_lock_validate(lock)); ut_a(lock->lock_word == X_LOCK_DECR); - lock->magic_n = 0; - #ifndef INNODB_RW_LOCKS_USE_ATOMICS mutex_free(rw_lock_get_mutex(lock)); #endif /* INNODB_RW_LOCKS_USE_ATOMICS */ @@ -325,6 +324,8 @@ rw_lock_free( UT_LIST_REMOVE(list, rw_lock_list, lock); mutex_exit(&rw_lock_list_mutex); + + lock->magic_n = 0; } #ifdef UNIV_DEBUG @@ -390,10 +391,10 @@ lock_loop: if (srv_print_latch_waits) { fprintf(stderr, "Thread %lu spin wait rw-s-lock at %p" - " cfile %s cline %lu rnds %lu\n", + " '%s' rnds %lu\n", (ulong) os_thread_pf(os_thread_get_curr_id()), (void*) lock, - lock->cfile_name, (ulong) lock->cline, (ulong) i); + lock->lock_name, (ulong) i); } /* We try once again to obtain the lock */ @@ -426,10 +427,9 @@ lock_loop: if (srv_print_latch_waits) { fprintf(stderr, "Thread %lu OS wait rw-s-lock at %p" - " cfile %s cline %lu\n", + " '%s'\n", os_thread_pf(os_thread_get_curr_id()), - (void*) lock, lock->cfile_name, - (ulong) lock->cline); + (void*) lock, lock->lock_name); } /* these stats may not be accurate */ @@ -648,9 +648,9 @@ lock_loop: if (srv_print_latch_waits) { fprintf(stderr, "Thread %lu spin wait rw-x-lock at %p" - " cfile %s cline %lu rnds %lu\n", + " '%s' rnds %lu\n", os_thread_pf(os_thread_get_curr_id()), (void*) lock, - lock->cfile_name, (ulong) lock->cline, (ulong) i); + lock->lock_name, (ulong) i); } sync_array_reserve_cell(sync_primary_wait_array, @@ -671,9 +671,9 @@ lock_loop: if (srv_print_latch_waits) { fprintf(stderr, "Thread %lu OS wait for rw-x-lock at %p" - " cfile %s cline %lu\n", + " '%s'\n", os_thread_pf(os_thread_get_curr_id()), (void*) lock, - lock->cfile_name, (ulong) lock->cline); + lock->lock_name); } /* these stats may not be accurate */ diff --git a/storage/xtradb/sync/sync0sync.c b/storage/xtradb/sync/sync0sync.c index 01aa49688b7..272492cb3e4 100644 --- a/storage/xtradb/sync/sync0sync.c +++ b/storage/xtradb/sync/sync0sync.c @@ -39,6 +39,7 @@ Created 9/5/1995 Heikki Tuuri #include "buf0buf.h" #include "srv0srv.h" #include "buf0types.h" +#include "os0sync.h" /* for HAVE_ATOMIC_BUILTINS */ /* REASONS FOR IMPLEMENTING THE SPIN LOCK MUTEX @@ -237,8 +238,8 @@ void mutex_create_func( /*==============*/ mutex_t* mutex, /*!< in: pointer to memory */ -#ifdef UNIV_DEBUG const char* cmutex_name, /*!< in: mutex name */ +#ifdef UNIV_DEBUG # ifdef UNIV_SYNC_DEBUG ulint level, /*!< in: level */ # endif /* UNIV_SYNC_DEBUG */ @@ -253,7 +254,7 @@ mutex_create_func( mutex->lock_word = 0; #endif mutex->event = os_event_create(NULL); - mutex_set_waiters(mutex, 0); + mutex->waiters = 0; #ifdef UNIV_DEBUG mutex->magic_n = MUTEX_MAGIC_N; #endif /* UNIV_DEBUG */ @@ -262,11 +263,13 @@ mutex_create_func( mutex->file_name = "not yet reserved"; mutex->level = level; #endif /* UNIV_SYNC_DEBUG */ +#ifdef UNIV_DEBUG mutex->cfile_name = cfile_name; mutex->cline = cline; +#endif /* UNIV_DEBUG */ mutex->count_os_wait = 0; -#ifdef UNIV_DEBUG mutex->cmutex_name= cmutex_name; +#ifdef UNIV_DEBUG mutex->count_using= 0; mutex->mutex_type= 0; mutex->lspent_time= 0; @@ -424,10 +427,18 @@ mutex_set_waiters( the value is stored to memory */ ut_ad(mutex); +#ifdef INNODB_RW_LOCKS_USE_ATOMICS + if (n) { + os_compare_and_swap_ulint(&mutex->waiters, 0, 1); + } else { + os_compare_and_swap_ulint(&mutex->waiters, 1, 0); + } +#else ptr = &(mutex->waiters); *ptr = n; /* Here we assume that the write of a single word in memory is atomic */ +#endif } /******************************************************************//** @@ -498,9 +509,9 @@ spin_loop: #ifdef UNIV_SRV_PRINT_LATCH_WAITS fprintf(stderr, "Thread %lu spin wait mutex at %p" - " cfile %s cline %lu rnds %lu\n", + " '%s' rnds %lu\n", (ulong) os_thread_pf(os_thread_get_curr_id()), (void*) mutex, - mutex->cfile_name, (ulong) mutex->cline, (ulong) i); + mutex->cmutex_name, (ulong) i); #endif mutex_spin_round_count += i; @@ -575,9 +586,9 @@ spin_loop: #ifdef UNIV_SRV_PRINT_LATCH_WAITS fprintf(stderr, - "Thread %lu OS wait mutex at %p cfile %s cline %lu rnds %lu\n", + "Thread %lu OS wait mutex at %p '%s' rnds %lu\n", (ulong) os_thread_pf(os_thread_get_curr_id()), (void*) mutex, - mutex->cfile_name, (ulong) mutex->cline, (ulong) i); + mutex->cmutex_name, (ulong) i); #endif mutex_os_wait_count++; @@ -849,7 +860,8 @@ sync_thread_levels_g( /*=================*/ sync_level_t* arr, /*!< in: pointer to level array for an OS thread */ - ulint limit) /*!< in: level limit */ + ulint limit, /*!< in: level limit */ + ulint warn) /*!< in: TRUE=display a diagnostic message */ { sync_level_t* slot; rw_lock_t* lock; @@ -863,6 +875,11 @@ sync_thread_levels_g( if (slot->latch != NULL) { if (slot->level <= limit) { + if (!warn) { + + return(FALSE); + } + lock = slot->latch; mutex = slot->latch; @@ -873,9 +890,8 @@ sync_thread_levels_g( if (mutex->magic_n == MUTEX_MAGIC_N) { fprintf(stderr, - "Mutex created at %s %lu\n", - mutex->cfile_name, - (ulong) mutex->cline); + "Mutex '%s'\n", + mutex->cmutex_name); if (mutex_get_lock_word(mutex) != 0) { const char* file_name; @@ -1106,7 +1122,7 @@ sync_thread_add_level( case SYNC_DICT_HEADER: case SYNC_TRX_I_S_RWLOCK: case SYNC_TRX_I_S_LAST_READ: - if (!sync_thread_levels_g(array, level)) { + if (!sync_thread_levels_g(array, level, TRUE)) { fprintf(stderr, "InnoDB: sync_thread_levels_g(array, %lu)" " does not hold!\n", level); @@ -1117,36 +1133,44 @@ sync_thread_add_level( /* Either the thread must own the buffer pool mutex (buf_pool_mutex), or it is allowed to latch only ONE buffer block (block->mutex or buf_pool_zip_mutex). */ - if (!sync_thread_levels_g(array, level)) { - ut_a(sync_thread_levels_g(array, level - 1)); + if (!sync_thread_levels_g(array, level, FALSE)) { + ut_a(sync_thread_levels_g(array, level - 1, TRUE)); ut_a(sync_thread_levels_contain(array, SYNC_BUF_LRU_LIST)); } break; case SYNC_REC_LOCK: - ut_a((sync_thread_levels_contain(array, SYNC_KERNEL) - && sync_thread_levels_g(array, SYNC_REC_LOCK - 1)) - || sync_thread_levels_g(array, SYNC_REC_LOCK)); + if (sync_thread_levels_contain(array, SYNC_KERNEL)) { + ut_a(sync_thread_levels_g(array, SYNC_REC_LOCK - 1, + TRUE)); + } else { + ut_a(sync_thread_levels_g(array, SYNC_REC_LOCK, TRUE)); + } break; case SYNC_IBUF_BITMAP: /* Either the thread must own the master mutex to all the bitmap pages, or it is allowed to latch only ONE bitmap page. */ - ut_a((sync_thread_levels_contain(array, SYNC_IBUF_BITMAP_MUTEX) - && sync_thread_levels_g(array, SYNC_IBUF_BITMAP - 1)) - || sync_thread_levels_g(array, SYNC_IBUF_BITMAP)); + if (sync_thread_levels_contain(array, + SYNC_IBUF_BITMAP_MUTEX)) { + ut_a(sync_thread_levels_g(array, SYNC_IBUF_BITMAP - 1, + TRUE)); + } else { + ut_a(sync_thread_levels_g(array, SYNC_IBUF_BITMAP, + TRUE)); + } break; case SYNC_FSP_PAGE: ut_a(sync_thread_levels_contain(array, SYNC_FSP)); break; case SYNC_FSP: ut_a(sync_thread_levels_contain(array, SYNC_FSP) - || sync_thread_levels_g(array, SYNC_FSP)); + || sync_thread_levels_g(array, SYNC_FSP, TRUE)); break; case SYNC_TRX_UNDO_PAGE: ut_a(sync_thread_levels_contain(array, SYNC_TRX_UNDO) || sync_thread_levels_contain(array, SYNC_RSEG) || sync_thread_levels_contain(array, SYNC_PURGE_SYS) - || sync_thread_levels_g(array, SYNC_TRX_UNDO_PAGE)); + || sync_thread_levels_g(array, SYNC_TRX_UNDO_PAGE, TRUE)); break; case SYNC_RSEG_HEADER: ut_a(sync_thread_levels_contain(array, SYNC_RSEG)); @@ -1158,37 +1182,41 @@ sync_thread_add_level( case SYNC_TREE_NODE: ut_a(sync_thread_levels_contain(array, SYNC_INDEX_TREE) || sync_thread_levels_contain(array, SYNC_DICT_OPERATION) - || sync_thread_levels_g(array, SYNC_TREE_NODE - 1)); + || sync_thread_levels_g(array, SYNC_TREE_NODE - 1, TRUE)); break; case SYNC_TREE_NODE_NEW: ut_a(sync_thread_levels_contain(array, SYNC_FSP_PAGE) || sync_thread_levels_contain(array, SYNC_IBUF_MUTEX)); break; case SYNC_INDEX_TREE: - ut_a((sync_thread_levels_contain(array, SYNC_IBUF_MUTEX) - && sync_thread_levels_contain(array, SYNC_FSP) - && sync_thread_levels_g(array, SYNC_FSP_PAGE - 1)) - || sync_thread_levels_g(array, SYNC_TREE_NODE - 1)); + if (sync_thread_levels_contain(array, SYNC_IBUF_MUTEX) + && sync_thread_levels_contain(array, SYNC_FSP)) { + ut_a(sync_thread_levels_g(array, SYNC_FSP_PAGE - 1, + TRUE)); + } else { + ut_a(sync_thread_levels_g(array, SYNC_TREE_NODE - 1, + TRUE)); + } break; case SYNC_IBUF_MUTEX: - ut_a(sync_thread_levels_g(array, SYNC_FSP_PAGE - 1)); + ut_a(sync_thread_levels_g(array, SYNC_FSP_PAGE - 1, TRUE)); break; case SYNC_IBUF_PESS_INSERT_MUTEX: - ut_a(sync_thread_levels_g(array, SYNC_FSP - 1) - && !sync_thread_levels_contain(array, SYNC_IBUF_MUTEX)); + ut_a(sync_thread_levels_g(array, SYNC_FSP - 1, TRUE)); + ut_a(!sync_thread_levels_contain(array, SYNC_IBUF_MUTEX)); break; case SYNC_IBUF_HEADER: - ut_a(sync_thread_levels_g(array, SYNC_FSP - 1) - && !sync_thread_levels_contain(array, SYNC_IBUF_MUTEX) - && !sync_thread_levels_contain( - array, SYNC_IBUF_PESS_INSERT_MUTEX)); + ut_a(sync_thread_levels_g(array, SYNC_FSP - 1, TRUE)); + ut_a(!sync_thread_levels_contain(array, SYNC_IBUF_MUTEX)); + ut_a(!sync_thread_levels_contain(array, + SYNC_IBUF_PESS_INSERT_MUTEX)); break; case SYNC_DICT: #ifdef UNIV_DEBUG ut_a(buf_debug_prints - || sync_thread_levels_g(array, SYNC_DICT)); + || sync_thread_levels_g(array, SYNC_DICT, TRUE)); #else /* UNIV_DEBUG */ - ut_a(sync_thread_levels_g(array, SYNC_DICT)); + ut_a(sync_thread_levels_g(array, SYNC_DICT, TRUE)); #endif /* UNIV_DEBUG */ break; default: @@ -1364,7 +1392,12 @@ sync_close(void) mutex_free(&mutex_list_mutex); #ifdef UNIV_SYNC_DEBUG mutex_free(&sync_thread_mutex); + + /* Switch latching order checks on in sync0sync.c */ + sync_order_checks_on = FALSE; #endif /* UNIV_SYNC_DEBUG */ + + sync_initialized = FALSE; } /*******************************************************************//** diff --git a/storage/xtradb/thr/thr0loc.c b/storage/xtradb/thr/thr0loc.c index ecc7d720dd5..5b9e83be920 100644 --- a/storage/xtradb/thr/thr0loc.c +++ b/storage/xtradb/thr/thr0loc.c @@ -63,7 +63,7 @@ struct thr_local_struct{ os_thread_t handle; /*!< operating system handle to the thread */ ulint slot_no;/*!< the index of the slot in the thread table for this thread */ - ibool in_ibuf;/*!< TRUE if the the thread is doing an ibuf + ibool in_ibuf;/*!< TRUE if the thread is doing an ibuf operation */ hash_node_t hash; /*!< hash chain node */ ulint magic_n;/*!< magic number (THR_LOCAL_MAGIC_N) */ @@ -250,6 +250,37 @@ thr_local_init(void) mutex_create(&thr_local_mutex, SYNC_THR_LOCAL); } +/******************************************************************** +Close the thread local storage module. */ +UNIV_INTERN +void +thr_local_close(void) +/*=================*/ +{ + ulint i; + + ut_a(thr_local_hash != NULL); + + /* Free the hash elements. We don't remove them from the table + because we are going to destroy the table anyway. */ + for (i = 0; i < hash_get_n_cells(thr_local_hash); i++) { + thr_local_t* local; + + local = HASH_GET_FIRST(thr_local_hash, i); + + while (local) { + thr_local_t* prev_local = local; + + local = HASH_GET_NEXT(hash, prev_local); + ut_a(prev_local->magic_n == THR_LOCAL_MAGIC_N); + mem_free(prev_local); + } + } + + hash_table_free(thr_local_hash); + thr_local_hash = NULL; +} + /************************************************************************* Return local hash table informations. */ diff --git a/storage/xtradb/trx/trx0i_s.c b/storage/xtradb/trx/trx0i_s.c index 90b73ad9e07..e7646c99227 100644 --- a/storage/xtradb/trx/trx0i_s.c +++ b/storage/xtradb/trx/trx0i_s.c @@ -60,7 +60,7 @@ Created July 17, 2007 Vasil Dimov /** @brief The maximum number of chunks to allocate for a table cache. The rows of a table cache are stored in a set of chunks. When a new -row is added a new chunk is allocated if necessary. Assuming that the +row is added a new chunk is allocated if necessary. Assuming that the first one is 1024 rows (TABLE_CACHE_INITIAL_ROWSNUM) and each subsequent is N/2 where N is the number of rows we have allocated till now, then 39th chunk would accommodate 1677416425 rows and all chunks @@ -238,6 +238,27 @@ table_cache_init( } /*******************************************************************//** +Frees a table cache. */ +static +void +table_cache_free( +/*=============*/ + i_s_table_cache_t* table_cache) /*!< in/out: table cache */ +{ + ulint i; + + for (i = 0; i < MEM_CHUNKS_IN_TABLE_CACHE; i++) { + + /* the memory is actually allocated in + table_cache_create_empty_row() */ + if (table_cache->chunks[i].base) { + mem_free(table_cache->chunks[i].base); + table_cache->chunks[i].base = NULL; + } + } +} + +/*******************************************************************//** Returns an empty row from a table cache. The row is allocated if no more empty rows are available. The number of used rows is incremented. If the memory limit is hit then NULL is returned and nothing is @@ -1252,6 +1273,22 @@ trx_i_s_cache_init( } /*******************************************************************//** +Free the INFORMATION SCHEMA trx related cache. */ +UNIV_INTERN +void +trx_i_s_cache_free( +/*===============*/ + trx_i_s_cache_t* cache) /*!< in, own: cache to free */ +{ + hash_table_free(cache->locks_hash); + ha_storage_free(cache->storage); + table_cache_free(&cache->innodb_trx); + table_cache_free(&cache->innodb_locks); + table_cache_free(&cache->innodb_lock_waits); + memset(cache, 0, sizeof *cache); +} + +/*******************************************************************//** Issue a shared/read lock on the tables cache. */ UNIV_INTERN void diff --git a/storage/xtradb/trx/trx0purge.c b/storage/xtradb/trx/trx0purge.c index cd79fd1c315..41e16b35e85 100644 --- a/storage/xtradb/trx/trx0purge.c +++ b/storage/xtradb/trx/trx0purge.c @@ -184,8 +184,9 @@ this query graph. @return own: the query graph */ static que_t* -trx_purge_graph_build(void) +trx_purge_graph_build( /*=======================*/ + trx_t* trx) { mem_heap_t* heap; que_fork_t* fork; @@ -194,7 +195,7 @@ trx_purge_graph_build(void) heap = mem_heap_create(512); fork = que_fork_create(NULL, NULL, QUE_FORK_PURGE, heap); - fork->trx = purge_sys->trx; + fork->trx = trx; thr = que_thr_create(fork, heap); @@ -243,10 +244,73 @@ trx_purge_sys_create(void) ut_a(trx_start_low(purge_sys->trx, ULINT_UNDEFINED)); - purge_sys->query = trx_purge_graph_build(); + purge_sys->query = trx_purge_graph_build(purge_sys->trx); purge_sys->view = read_view_oldest_copy_or_open_new(ut_dulint_zero, purge_sys->heap); + + purge_sys->n_worker = 0; + if (srv_use_purge_thread > 1) { + /* Use worker threads */ + ulint i; + + purge_sys->n_worker = srv_use_purge_thread - 1; + + purge_sys->sess_arr = mem_alloc(sizeof(sess_t*) * purge_sys->n_worker); + purge_sys->trx_arr = mem_alloc(sizeof(trx_t*) * purge_sys->n_worker); + purge_sys->query_arr = mem_alloc(sizeof(que_t*) * purge_sys->n_worker); + + purge_sys->worker_event = os_event_create(NULL); + os_event_reset(purge_sys->worker_event); + + for (i = 0; i < purge_sys->n_worker; i++) { + purge_sys->sess_arr[i] = sess_open(); + + purge_sys->trx_arr[i] = purge_sys->sess_arr[i]->trx; + purge_sys->trx_arr[i]->is_purge = 1; + ut_a(trx_start_low(purge_sys->trx_arr[i], ULINT_UNDEFINED)); + + purge_sys->query_arr[i] = trx_purge_graph_build(purge_sys->trx_arr[i]); + } + } +} + +/************************************************************************ +Frees the global purge system control structure. */ +UNIV_INTERN +void +trx_purge_sys_close(void) +/*======================*/ +{ + ut_ad(!mutex_own(&kernel_mutex)); + + que_graph_free(purge_sys->query); + + ut_a(purge_sys->sess->trx->is_purge); + purge_sys->sess->trx->conc_state = TRX_NOT_STARTED; + sess_close(purge_sys->sess); + purge_sys->sess = NULL; + + if (purge_sys->view != NULL) { + /* Because acquiring the kernel mutex is a pre-condition + of read_view_close(). We don't really need it here. */ + mutex_enter(&kernel_mutex); + + read_view_close(purge_sys->view); + purge_sys->view = NULL; + + mutex_exit(&kernel_mutex); + } + + trx_undo_arr_free(purge_sys->arr); + + rw_lock_free(&purge_sys->latch); + mutex_free(&purge_sys->mutex); + + mem_heap_free(purge_sys->heap); + mem_free(purge_sys); + + purge_sys = NULL; } /*================ UNDO LOG HISTORY LIST =============================*/ @@ -1110,7 +1174,7 @@ trx_purge(void) /* Handle at most 20 undo log pages in one purge batch */ - purge_sys->handle_limit = purge_sys->n_pages_handled + 20; + purge_sys->handle_limit = purge_sys->n_pages_handled + 20 * (srv_use_purge_thread + 1); old_pages_handled = purge_sys->n_pages_handled; @@ -1129,6 +1193,9 @@ trx_purge(void) mutex_exit(&kernel_mutex); + if (purge_sys->n_worker) + os_event_set(purge_sys->worker_event); + /* srv_que_task_enqueue(thr2); */ if (srv_print_thread_releases) { @@ -1138,6 +1205,9 @@ trx_purge(void) que_run_threads(thr); + if (purge_sys->n_worker) + os_event_reset(purge_sys->worker_event); + if (srv_print_thread_releases) { fprintf(stderr, @@ -1148,6 +1218,52 @@ trx_purge(void) return(purge_sys->n_pages_handled - old_pages_handled); } +/********************************************************************** +This function runs a purge worker batch */ +UNIV_INTERN +void +trx_purge_worker( +/*=============*/ + ulint worker_id) +{ + que_thr_t* thr; + + mutex_enter(&kernel_mutex); + + thr = que_fork_start_command(purge_sys->query_arr[worker_id]); + + ut_ad(thr); + + mutex_exit(&kernel_mutex); + + que_run_threads(thr); + + if (purge_sys->state == TRX_STOP_PURGE) { /* optimistic */ + os_event_reset(purge_sys->worker_event); + } +} + +/********************************************************************** +This function waits the event for worker batch */ +UNIV_INTERN +void +trx_purge_worker_wait(void) +/*=======================*/ +{ + os_event_wait(purge_sys->worker_event); +} + +/********************************************************************** +This function wakes the waiting worker batch */ +UNIV_INTERN +void +trx_purge_worker_wake(void) +/*=======================*/ +{ + if (purge_sys->n_worker) + os_event_set(purge_sys->worker_event); +} + /******************************************************************//** Prints information of the purge system to stderr. */ UNIV_INTERN diff --git a/storage/xtradb/trx/trx0rec.c b/storage/xtradb/trx/trx0rec.c index 36911c9df85..5097cf18dcd 100644 --- a/storage/xtradb/trx/trx0rec.c +++ b/storage/xtradb/trx/trx0rec.c @@ -1333,7 +1333,7 @@ trx_undo_get_undo_rec_low( ulint rseg_id; ulint page_no; ulint offset; - page_t* undo_page; + const page_t* undo_page; trx_rseg_t* rseg; ibool is_insert; mtr_t mtr; @@ -1572,7 +1572,7 @@ trx_undo_prev_version_build( /* We have to set the appropriate extern storage bits in the old version of the record: the extern bits in rec for those - fields that update does NOT update, as well as the the bits for + fields that update does NOT update, as well as the bits for those fields that update updates to become externally stored fields. Store the info: */ diff --git a/storage/xtradb/trx/trx0roll.c b/storage/xtradb/trx/trx0roll.c index 51d17192d5b..c925478cdf4 100644 --- a/storage/xtradb/trx/trx0roll.c +++ b/storage/xtradb/trx/trx0roll.c @@ -66,9 +66,9 @@ int trx_general_rollback_for_mysql( /*===========================*/ trx_t* trx, /*!< in: transaction handle */ - ibool partial,/*!< in: TRUE if partial rollback requested */ trx_savept_t* savept) /*!< in: pointer to savepoint undo number, if - partial rollback requested */ + partial rollback requested, or NULL for + complete rollback */ { mem_heap_t* heap; que_thr_t* thr; @@ -85,9 +85,8 @@ trx_general_rollback_for_mysql( roll_node = roll_node_create(heap); - roll_node->partial = partial; - - if (partial) { + if (savept) { + roll_node->partial = TRUE; roll_node->savept = *savept; } @@ -145,7 +144,7 @@ trx_rollback_for_mysql( the transaction object does not have an InnoDB session object, and we set a dummy session that we use for all MySQL transactions. */ - err = trx_general_rollback_for_mysql(trx, FALSE, NULL); + err = trx_general_rollback_for_mysql(trx, NULL); trx->op_info = ""; @@ -170,8 +169,7 @@ trx_rollback_last_sql_stat_for_mysql( trx->op_info = "rollback of SQL statement"; - err = trx_general_rollback_for_mysql(trx, TRUE, - &(trx->last_sql_stat_start)); + err = trx_general_rollback_for_mysql(trx, &trx->last_sql_stat_start); /* The following call should not be needed, but we play safe: */ trx_mark_sql_stat_end(trx); @@ -282,7 +280,7 @@ trx_rollback_to_savepoint_for_mysql( trx->op_info = "rollback to a savepoint"; - err = trx_general_rollback_for_mysql(trx, TRUE, &(savep->savept)); + err = trx_general_rollback_for_mysql(trx, &savep->savept); /* Store the current undo_no of the transaction so that we know where to roll back if we have to roll back the next SQL statement: */ @@ -534,28 +532,26 @@ trx_rollback_active( Rollback or clean up any incomplete transactions which were encountered in crash recovery. If the transaction already was committed, then we clean up a possible insert undo log. If the -transaction was not yet committed, then we roll it back. -Note: this is done in a background thread. -@return a dummy parameter */ +transaction was not yet committed, then we roll it back. */ UNIV_INTERN -os_thread_ret_t -trx_rollback_or_clean_all_recovered( -/*================================*/ - void* arg __attribute__((unused))) - /*!< in: a dummy parameter required by - os_thread_create */ +void +trx_rollback_or_clean_recovered( +/*============================*/ + ibool all) /*!< in: FALSE=roll back dictionary transactions; + TRUE=roll back all non-PREPARED transactions */ { trx_t* trx; mutex_enter(&kernel_mutex); - if (UT_LIST_GET_FIRST(trx_sys->trx_list)) { + if (!UT_LIST_GET_FIRST(trx_sys->trx_list)) { + goto leave_function; + } + if (all) { fprintf(stderr, "InnoDB: Starting in background the rollback" " of uncommitted transactions\n"); - } else { - goto leave_function; } mutex_exit(&kernel_mutex); @@ -584,18 +580,42 @@ loop: goto loop; case TRX_ACTIVE: - mutex_exit(&kernel_mutex); - trx_rollback_active(trx); - goto loop; + if (all || trx_get_dict_operation(trx) + != TRX_DICT_OP_NONE) { + mutex_exit(&kernel_mutex); + trx_rollback_active(trx); + goto loop; + } } } - ut_print_timestamp(stderr); - fprintf(stderr, - " InnoDB: Rollback of non-prepared transactions completed\n"); + if (all) { + ut_print_timestamp(stderr); + fprintf(stderr, + " InnoDB: Rollback of non-prepared" + " transactions completed\n"); + } leave_function: mutex_exit(&kernel_mutex); +} + +/*******************************************************************//** +Rollback or clean up any incomplete transactions which were +encountered in crash recovery. If the transaction already was +committed, then we clean up a possible insert undo log. If the +transaction was not yet committed, then we roll it back. +Note: this is done in a background thread. +@return a dummy parameter */ +UNIV_INTERN +os_thread_ret_t +trx_rollback_or_clean_all_recovered( +/*================================*/ + void* arg __attribute__((unused))) + /*!< in: a dummy parameter required by + os_thread_create */ +{ + trx_rollback_or_clean_recovered(TRUE); /* We count the number of threads in os_thread_exit(). A created thread should always use that to exit and not use return() to exit. */ diff --git a/storage/xtradb/trx/trx0rseg.c b/storage/xtradb/trx/trx0rseg.c index 580762e8716..8d754788e2a 100644 --- a/storage/xtradb/trx/trx0rseg.c +++ b/storage/xtradb/trx/trx0rseg.c @@ -132,6 +132,49 @@ trx_rseg_header_create( } /***********************************************************************//** +Free's an instance of the rollback segment in memory. */ +UNIV_INTERN +void +trx_rseg_mem_free( +/*==============*/ + trx_rseg_t* rseg) /* in, own: instance to free */ +{ + trx_undo_t* undo; + + mutex_free(&rseg->mutex); + + /* There can't be any active transactions. */ + ut_a(UT_LIST_GET_LEN(rseg->update_undo_list) == 0); + ut_a(UT_LIST_GET_LEN(rseg->insert_undo_list) == 0); + + undo = UT_LIST_GET_FIRST(rseg->update_undo_cached); + + while (undo != NULL) { + trx_undo_t* prev_undo = undo; + + undo = UT_LIST_GET_NEXT(undo_list, undo); + UT_LIST_REMOVE(undo_list, rseg->update_undo_cached, prev_undo); + + trx_undo_mem_free(prev_undo); + } + + undo = UT_LIST_GET_FIRST(rseg->insert_undo_cached); + + while (undo != NULL) { + trx_undo_t* prev_undo = undo; + + undo = UT_LIST_GET_NEXT(undo_list, undo); + UT_LIST_REMOVE(undo_list, rseg->insert_undo_cached, prev_undo); + + trx_undo_mem_free(prev_undo); + } + + trx_sys_set_nth_rseg(trx_sys, rseg->id, NULL); + + mem_free(rseg); +} + +/*************************************************************************** Creates and initializes a rollback segment object. The values for the fields are read from the header. The object is inserted to the rseg list of the trx system object and a pointer is inserted in the rseg diff --git a/storage/xtradb/trx/trx0sys.c b/storage/xtradb/trx/trx0sys.c index 81c614c8323..8ea34a8c81c 100644 --- a/storage/xtradb/trx/trx0sys.c +++ b/storage/xtradb/trx/trx0sys.c @@ -39,7 +39,9 @@ Created 3/26/1996 Heikki Tuuri #include "srv0srv.h" #include "trx0purge.h" #include "log0log.h" +#include "log0recv.h" #include "os0file.h" +#include "read0read.h" /** The file format tag structure with id and name. */ struct file_format_struct { @@ -552,6 +554,12 @@ trx_sys_doublewrite_init_or_restore_pages( zip_size ? zip_size : UNIV_PAGE_SIZE, read_buf, NULL); + if (srv_recovery_stats && recv_recovery_is_on()) { + mutex_enter(&(recv_sys->mutex)); + recv_sys->stats_doublewrite_check_pages++; + mutex_exit(&(recv_sys->mutex)); + } + /* Check if the page is corrupt */ if (UNIV_UNLIKELY @@ -599,6 +607,13 @@ trx_sys_doublewrite_init_or_restore_pages( zip_size, page_no, 0, zip_size ? zip_size : UNIV_PAGE_SIZE, page, NULL); + + if (srv_recovery_stats && recv_recovery_is_on()) { + mutex_enter(&(recv_sys->mutex)); + recv_sys->stats_doublewrite_overwrite_pages++; + mutex_exit(&(recv_sys->mutex)); + } + fprintf(stderr, "InnoDB: Recovered the page from" " the doublewrite buffer.\n"); @@ -1592,3 +1607,80 @@ trx_sys_file_format_id_to_name( } #endif /* !UNIV_HOTBACKUP */ + +/********************************************************************* +Shutdown/Close the transaction system. */ +UNIV_INTERN +void +trx_sys_close(void) +/*===============*/ +{ + trx_rseg_t* rseg; + read_view_t* view; + + ut_ad(trx_sys != NULL); + + /* Check that all read views are closed except read view owned + by a purge. */ + + if (UT_LIST_GET_LEN(trx_sys->view_list) > 1) { + fprintf(stderr, + "InnoDB: Error: all read views were not closed" + " before shutdown:\n" + "InnoDB: %lu read views open \n", + UT_LIST_GET_LEN(trx_sys->view_list) - 1); + } + + sess_close(trx_dummy_sess); + trx_dummy_sess = NULL; + + trx_purge_sys_close(); + + mutex_enter(&kernel_mutex); + + /* Free the double write data structures. */ + ut_a(trx_doublewrite != NULL); + ut_free(trx_doublewrite->write_buf_unaligned); + trx_doublewrite->write_buf_unaligned = NULL; + + mem_free(trx_doublewrite->buf_block_arr); + trx_doublewrite->buf_block_arr = NULL; + + mutex_free(&trx_doublewrite->mutex); + mem_free(trx_doublewrite); + trx_doublewrite = NULL; + + /* There can't be any active transactions. */ + rseg = UT_LIST_GET_FIRST(trx_sys->rseg_list); + + while (rseg != NULL) { + trx_rseg_t* prev_rseg = rseg; + + rseg = UT_LIST_GET_NEXT(rseg_list, prev_rseg); + UT_LIST_REMOVE(rseg_list, trx_sys->rseg_list, prev_rseg); + + trx_rseg_mem_free(prev_rseg); + } + + view = UT_LIST_GET_FIRST(trx_sys->view_list); + + while (view != NULL) { + read_view_t* prev_view = view; + + view = UT_LIST_GET_NEXT(view_list, prev_view); + + /* Views are allocated from the trx_sys->global_read_view_heap. + So, we simply remove the element here. */ + UT_LIST_REMOVE(view_list, trx_sys->view_list, prev_view); + } + + ut_a(UT_LIST_GET_LEN(trx_sys->trx_list) == 0); + ut_a(UT_LIST_GET_LEN(trx_sys->rseg_list) == 0); + ut_a(UT_LIST_GET_LEN(trx_sys->view_list) == 0); + ut_a(UT_LIST_GET_LEN(trx_sys->mysql_trx_list) == 0); + + mem_free(trx_sys); + + trx_sys = NULL; + mutex_exit(&kernel_mutex); +} diff --git a/storage/xtradb/trx/trx0trx.c b/storage/xtradb/trx/trx0trx.c index 12253c131a8..69bc5c6b18b 100644 --- a/storage/xtradb/trx/trx0trx.c +++ b/storage/xtradb/trx/trx0trx.c @@ -178,6 +178,15 @@ trx_create( trx->global_read_view = NULL; trx->read_view = NULL; + trx->io_reads = 0; + trx->io_read = 0; + trx->io_reads_wait_timer = 0; + trx->lock_que_wait_timer = 0; + trx->innodb_que_wait_timer = 0; + trx->distinct_page_access = 0; + trx->distinct_page_access_hash = NULL; + trx->take_stats = FALSE; + /* Set X/Open XA transaction identification to NULL */ memset(&trx->xid, 0, sizeof(trx->xid)); trx->xid.formatID = -1; @@ -215,6 +224,11 @@ trx_allocate_for_mysql(void) trx->mysql_process_no = os_proc_get_number(); + if (innobase_get_slow_log() && trx->take_stats) { + trx->distinct_page_access_hash = mem_alloc(DPAH_SIZE); + memset(trx->distinct_page_access_hash, 0, DPAH_SIZE); + } + return(trx); } @@ -346,6 +360,12 @@ trx_free_for_mysql( /*===============*/ trx_t* trx) /*!< in, own: trx object */ { + if (trx->distinct_page_access_hash) + { + mem_free(trx->distinct_page_access_hash); + trx->distinct_page_access_hash= NULL; + } + mutex_enter(&kernel_mutex); UT_LIST_REMOVE(mysql_trx_list, trx_sys->mysql_trx_list, trx); @@ -367,6 +387,12 @@ trx_free_for_background( /*====================*/ trx_t* trx) /*!< in, own: trx object */ { + if (trx->distinct_page_access_hash) + { + mem_free(trx->distinct_page_access_hash); + trx->distinct_page_access_hash= NULL; + } + mutex_enter(&kernel_mutex); trx_free(trx); @@ -820,7 +846,7 @@ trx_commit_off_kernel( in exactly the same order as commit lsn's, if the transactions have different rollback segments. To get exactly the same order we should hold the kernel mutex up to this point, - adding to to the contention of the kernel mutex. However, if + adding to the contention of the kernel mutex. However, if a transaction T2 is able to see modifications made by a transaction T1, T2 will always get a bigger transaction number and a bigger commit lsn than T1. */ @@ -967,7 +993,7 @@ trx_commit_off_kernel( /****************************************************************//** Cleans up a transaction at database startup. The cleanup is needed if the transaction already got to the middle of a commit when the database -crashed, andf we cannot roll it back. */ +crashed, and we cannot roll it back. */ UNIV_INTERN void trx_cleanup_at_db_startup( @@ -1072,6 +1098,9 @@ trx_end_lock_wait( trx_t* trx) /*!< in: transaction */ { que_thr_t* thr; + ulint sec; + ulint ms; + ib_uint64_t now; ut_ad(mutex_own(&kernel_mutex)); ut_ad(trx->que_state == TRX_QUE_LOCK_WAIT); @@ -1086,6 +1115,11 @@ trx_end_lock_wait( thr = UT_LIST_GET_FIRST(trx->wait_thrs); } + if (innobase_get_slow_log() && trx->take_stats) { + ut_usectime(&sec, &ms); + now = (ib_uint64_t)sec * 1000000 + ms; + trx->lock_que_wait_timer += (ulint)(now - trx->lock_que_wait_ustarted); + } trx->que_state = TRX_QUE_RUNNING; } @@ -1099,6 +1133,9 @@ trx_lock_wait_to_suspended( trx_t* trx) /*!< in: transaction in the TRX_QUE_LOCK_WAIT state */ { que_thr_t* thr; + ulint sec; + ulint ms; + ib_uint64_t now; ut_ad(mutex_own(&kernel_mutex)); ut_ad(trx->que_state == TRX_QUE_LOCK_WAIT); @@ -1113,6 +1150,11 @@ trx_lock_wait_to_suspended( thr = UT_LIST_GET_FIRST(trx->wait_thrs); } + if (innobase_get_slow_log() && trx->take_stats) { + ut_usectime(&sec, &ms); + now = (ib_uint64_t)sec * 1000000 + ms; + trx->lock_que_wait_timer += (ulint)(now - trx->lock_que_wait_ustarted); + } trx->que_state = TRX_QUE_RUNNING; } diff --git a/storage/xtradb/trx/trx0undo.c b/storage/xtradb/trx/trx0undo.c index 99672d4eb81..063a2f2b8a6 100644 --- a/storage/xtradb/trx/trx0undo.c +++ b/storage/xtradb/trx/trx0undo.c @@ -1560,7 +1560,7 @@ trx_undo_mem_init_for_reuse( /********************************************************************//** Frees an undo log memory copy. */ -static +UNIV_INTERN void trx_undo_mem_free( /*==============*/ diff --git a/storage/xtradb/usr/usr0sess.c b/storage/xtradb/usr/usr0sess.c index 990991a2c06..8087dcb4170 100644 --- a/storage/xtradb/usr/usr0sess.c +++ b/storage/xtradb/usr/usr0sess.c @@ -32,14 +32,6 @@ Created 6/25/1996 Heikki Tuuri #include "trx0trx.h" /*********************************************************************//** -Closes a session, freeing the memory occupied by it. */ -static -void -sess_close( -/*=======*/ - sess_t* sess); /*!< in, own: session object */ - -/*********************************************************************//** Opens a session. @return own: session object */ UNIV_INTERN @@ -64,35 +56,16 @@ sess_open(void) /*********************************************************************//** Closes a session, freeing the memory occupied by it. */ -static +UNIV_INTERN void sess_close( /*=======*/ sess_t* sess) /*!< in, own: session object */ { - ut_ad(mutex_own(&kernel_mutex)); - ut_ad(sess->trx == NULL); - - mem_free(sess); -} - -/*********************************************************************//** -Closes a session, freeing the memory occupied by it, if it is in a state -where it should be closed. -@return TRUE if closed */ -UNIV_INTERN -ibool -sess_try_close( -/*===========*/ - sess_t* sess) /*!< in, own: session object */ -{ - ut_ad(mutex_own(&kernel_mutex)); + ut_ad(!mutex_own(&kernel_mutex)); - if (UT_LIST_GET_LEN(sess->graphs) == 0) { - sess_close(sess); + ut_a(UT_LIST_GET_LEN(sess->graphs) == 0); - return(TRUE); - } - - return(FALSE); + trx_free_for_background(sess->trx); + mem_free(sess); } diff --git a/storage/xtradb/ut/ut0auxconf_atomic_pthread_t_solaris.c b/storage/xtradb/ut/ut0auxconf_atomic_pthread_t_solaris.c index a18a537d1d4..310603c7503 100644 --- a/storage/xtradb/ut/ut0auxconf_atomic_pthread_t_solaris.c +++ b/storage/xtradb/ut/ut0auxconf_atomic_pthread_t_solaris.c @@ -17,18 +17,38 @@ Place, Suite 330, Boston, MA 02111-1307 USA *****************************************************************************/ /***************************************************************************** -If this program compiles, then pthread_t objects can be used as arguments -to Solaris libc atomic functions. +If this program compiles and returns 0, then pthread_t objects can be used as +arguments to Solaris libc atomic functions. Created April 18, 2009 Vasil Dimov *****************************************************************************/ #include <pthread.h> +#include <string.h> int main(int argc, char** argv) { - pthread_t x = 0; + pthread_t x1; + pthread_t x2; + pthread_t x3; + + memset(&x1, 0x0, sizeof(x1)); + memset(&x2, 0x0, sizeof(x2)); + memset(&x3, 0x0, sizeof(x3)); + + if (sizeof(pthread_t) == 4) { + + atomic_cas_32(&x1, x2, x3); + + } else if (sizeof(pthread_t) == 8) { + + atomic_cas_64(&x1, x2, x3); + + } else { + + return(1); + } return(0); } diff --git a/storage/xtradb/ut/ut0auxconf_have_gcc_atomics.c b/storage/xtradb/ut/ut0auxconf_have_gcc_atomics.c new file mode 100644 index 00000000000..da5c13d7d79 --- /dev/null +++ b/storage/xtradb/ut/ut0auxconf_have_gcc_atomics.c @@ -0,0 +1,61 @@ +/***************************************************************************** + +Copyright (c) 2009, Innobase Oy. All Rights Reserved. + +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., 59 Temple +Place, Suite 330, Boston, MA 02111-1307 USA + +*****************************************************************************/ + +/***************************************************************************** +If this program compiles and returns 0, then GCC atomic funcions are available. + +Created September 12, 2009 Vasil Dimov +*****************************************************************************/ + +int +main(int argc, char** argv) +{ + long x; + long y; + long res; + char c; + + x = 10; + y = 123; + res = __sync_bool_compare_and_swap(&x, x, y); + if (!res || x != y) { + return(1); + } + + x = 10; + y = 123; + res = __sync_bool_compare_and_swap(&x, x + 1, y); + if (res || x != 10) { + return(1); + } + + x = 10; + y = 123; + res = __sync_add_and_fetch(&x, y); + if (res != 123 + 10 || x != 123 + 10) { + return(1); + } + + c = 10; + res = __sync_lock_test_and_set(&c, 123); + if (res != 10 || c != 123) { + return(1); + } + + return(0); +} diff --git a/storage/xtradb/ut/ut0mem.c b/storage/xtradb/ut/ut0mem.c index edb63c95700..35a325b9ccd 100644 --- a/storage/xtradb/ut/ut0mem.c +++ b/storage/xtradb/ut/ut0mem.c @@ -433,6 +433,8 @@ ut_free_all_mem(void) " total allocated memory is %lu\n", (ulong) ut_total_allocated_memory); } + + ut_mem_block_list_inited = FALSE; } #endif /* !UNIV_HOTBACKUP */ diff --git a/storage/xtradb/ut/ut0ut.c b/storage/xtradb/ut/ut0ut.c index e4cc226fbad..498873e290a 100644 --- a/storage/xtradb/ut/ut0ut.c +++ b/storage/xtradb/ut/ut0ut.c @@ -132,6 +132,7 @@ ut_time(void) return(time(NULL)); } +#ifndef UNIV_HOTBACKUP /**********************************************************//** Returns system time. Upon successful completion, the value 0 is returned; otherwise the @@ -200,6 +201,24 @@ ut_time_us( } /**********************************************************//** +Returns the number of milliseconds since some epoch. The +value may wrap around. It should only be used for heuristic +purposes. +@return ms since epoch */ +UNIV_INTERN +ulint +ut_time_ms(void) +/*============*/ +{ + struct timeval tv; + + ut_gettimeofday(&tv, NULL); + + return((ulint) tv.tv_sec * 1000 + tv.tv_usec / 1000); +} +#endif /* !UNIV_HOTBACKUP */ + +/**********************************************************//** Returns the difference of two times in seconds. @return time2 - time1 expressed in seconds */ UNIV_INTERN diff --git a/storage/xtradb/win-plugin/README b/storage/xtradb/win-plugin/README deleted file mode 100644 index 00f4e996a3f..00000000000 --- a/storage/xtradb/win-plugin/README +++ /dev/null @@ -1,22 +0,0 @@ -This directory contains patches that need to be applied to the MySQL -source tree in order to build the dynamic plugin on Windows -- -HA_INNODB.DLL. Please note the followings when adding the patches: - -* The patch must be applied from the mysql top-level source directory. - patch -p0 < win-plugin.diff -* The patch filenames end in ".diff". -* All patches here are expected to apply cleanly to the latest MySQL 5.1 - tree when storage/innobase is replaced with this InnoDB branch. - -When applying the patch, the following files will be modified: - - * CMakeLists.txt - * sql/CMakeLists.txt - * win/configure.js - -Also, two new files will be added: - - * sql/mysqld.def - * sql/mysqld_x64.def - -You can get "patch" utility for Windows from http://unxutils.sourceforge.net/ diff --git a/storage/xtradb/win-plugin/win-plugin.diff b/storage/xtradb/win-plugin/win-plugin.diff deleted file mode 100644 index 4b3354ac4de..00000000000 --- a/storage/xtradb/win-plugin/win-plugin.diff +++ /dev/null @@ -1,279 +0,0 @@ -diff -Nur CMakeLists.txt.orig CMakeLists.txt ---- CMakeLists.txt.orig 2008-10-03 12:25:41 -05:00 -+++ CMakeLists.txt 2008-09-26 17:32:51 -05:00 -@@ -254,9 +254,9 @@ - IF(WITH_FEDERATED_STORAGE_ENGINE) - ADD_SUBDIRECTORY(storage/federated) - ENDIF(WITH_FEDERATED_STORAGE_ENGINE) --IF(WITH_INNOBASE_STORAGE_ENGINE) -+IF(WITH_INNOBASE_STORAGE_ENGINE OR INNODB_DYNAMIC_PLUGIN) - ADD_SUBDIRECTORY(storage/innobase) --ENDIF(WITH_INNOBASE_STORAGE_ENGINE) -+ENDIF(WITH_INNOBASE_STORAGE_ENGINE OR INNODB_DYNAMIC_PLUGIN) - ADD_SUBDIRECTORY(sql) - ADD_SUBDIRECTORY(server-tools/instance-manager) - ADD_SUBDIRECTORY(libmysql) - -diff -Nur sql/CMakeLists.txt.orig sql/CMakeLists.txt ---- sql/CMakeLists.txt.orig 2008-10-03 12:25:41 -05:00 -+++ sql/CMakeLists.txt 2008-09-24 03:58:19 -05:00 -@@ -98,6 +98,15 @@ - LINK_FLAGS "/PDB:${CMAKE_CFG_INTDIR}/mysqld${MYSQLD_EXE_SUFFIX}.pdb") - ENDIF(cmake_version EQUAL 20406) - -+# Checks for 64-bit version -+IF(CMAKE_SIZEOF_VOID_P MATCHES 8) -+SET_TARGET_PROPERTIES(mysqld PROPERTIES -+ LINK_FLAGS "/def:\"${PROJECT_SOURCE_DIR}/sql/mysqld_x64.def\"") -+ELSE(CMAKE_SIZEOF_VOID_P MATCHES 8) -+SET_TARGET_PROPERTIES(mysqld PROPERTIES -+ LINK_FLAGS "/def:\"${PROJECT_SOURCE_DIR}/sql/mysqld.def\"") -+ENDIF(CMAKE_SIZEOF_VOID_P MATCHES 8) -+ - IF(EMBED_MANIFESTS) - MYSQL_EMBED_MANIFEST("mysqld" "asInvoker") - ENDIF(EMBED_MANIFESTS) - -diff -Nur sql/mysqld.def.orig sql/mysqld.def ---- sql/mysqld.def.orig 1969-12-31 18:00:00 -06:00 -+++ sql/mysqld.def 2009-04-09 02:20:32 -05:00 -@@ -0,0 +1,111 @@ -+EXPORTS -+ ?use_hidden_primary_key@handler@@UAEXXZ -+ ?get_dynamic_partition_info@handler@@UAEXPAUPARTITION_INFO@@I@Z -+ ?read_first_row@handler@@UAEHPAEI@Z -+ ?read_range_next@handler@@UAEHXZ -+ ?read_range_first@handler@@UAEHPBUst_key_range@@0_N1@Z -+ ?read_multi_range_first@handler@@UAEHPAPAUst_key_multi_range@@PAU2@I_NPAUst_handler_buffer@@@Z -+ ?read_multi_range_next@handler@@UAEHPAPAUst_key_multi_range@@@Z -+ ?index_read_idx_map@handler@@UAEHPAEIPBEKW4ha_rkey_function@@@Z -+ ?print_error@handler@@UAEXHH@Z -+ ?clone@handler@@UAEPAV1@PAUst_mem_root@@@Z -+ ?get_auto_increment@handler@@UAEX_K00PA_K1@Z -+ ?index_next_same@handler@@UAEHPAEPBEI@Z -+ ?get_error_message@handler@@UAE_NHPAVString@@@Z -+ ?ha_thd@handler@@IBEPAVTHD@@XZ -+ ?update_auto_increment@handler@@QAEHXZ -+ ?ha_statistic_increment@handler@@IBEXPQsystem_status_var@@K@Z -+ ?trans_register_ha@@YAXPAVTHD@@_NPAUhandlerton@@@Z -+ ?cmp@Field_blob@@QAEHPBEI0I@Z -+ ?set_time@Field_timestamp@@QAEXXZ -+ ?sql_print_error@@YAXPBDZZ -+ ?sql_print_warning@@YAXPBDZZ -+ ?check_global_access@@YA_NPAVTHD@@K@Z -+ ?schema_table_store_record@@YA_NPAVTHD@@PAUst_table@@@Z -+ ?get_quote_char_for_identifier@@YAHPAVTHD@@PBDI@Z -+ ?copy@String@@QAE_NXZ -+ ?copy@String@@QAE_NABV1@@Z -+ ?copy@String@@QAE_NPBDIPAUcharset_info_st@@@Z -+ ?copy_and_convert@@YAIPADIPAUcharset_info_st@@PBDI1PAI@Z -+ ?filename_to_tablename@@YAIPBDPADI@Z -+ ?strconvert@@YAIPAUcharset_info_st@@PBD0PADIPAI@Z -+ ?calculate_key_len@@YAIPAUst_table@@IPBEK@Z -+ ?sql_alloc@@YAPAXI@Z -+ ?localtime_to_TIME@@YAXPAUst_mysql_time@@PAUtm@@@Z -+ ?push_warning@@YAPAVMYSQL_ERROR@@PAVTHD@@W4enum_warning_level@1@IPBD@Z -+ ?push_warning_printf@@YAXPAVTHD@@W4enum_warning_level@MYSQL_ERROR@@IPBDZZ -+ ?drop_table@handler@@EAEXPBD@Z -+ ?column_bitmaps_signal@handler@@UAEXXZ -+ ?delete_table@handler@@MAEHPBD@Z -+ ?rename_table@handler@@MAEHPBD0@Z -+ ?key_map_empty@@3V?$Bitmap@$0EA@@@B -+ ?THR_THD@@3PAVTHD@@A -+ ?end_of_list@@3Ulist_node@@A -+ ?mysql_tmpdir_list@@3Ust_my_tmpdir@@A -+ mysql_query_cache_invalidate4 -+ thd_query -+ thd_sql_command -+ thd_get_thread_id -+ thd_get_xid -+ thd_slave_thread -+ thd_non_transactional_update -+ thd_mark_transaction_to_rollback -+ thd_security_context -+ thd_charset -+ thd_test_options -+ thd_ha_data -+ thd_killed -+ thd_tx_isolation -+ thd_tablespace_op -+ thd_sql_command -+ thd_memdup -+ thd_make_lex_string -+ thd_in_lock_tables -+ thd_binlog_format -+ _my_hash_init -+ my_hash_free -+ my_tmpdir -+ check_if_legal_filename -+ my_filename -+ my_sync_dir_by_file -+ alloc_root -+ thr_lock_data_init -+ thr_lock_init -+ thr_lock_delete -+ my_multi_malloc -+ get_charset -+ unpack_filename -+ my_hash_insert -+ my_hash_search -+ my_hash_delete -+ mysql_bin_log_file_pos -+ mysql_bin_log_file_name -+ mysqld_embedded -+ my_thread_name -+ my_malloc -+ my_no_flags_free -+ _sanity -+ _mymalloc -+ _myfree -+ _my_strdup -+ _my_thread_var -+ my_error -+ pthread_cond_init -+ pthread_cond_signal -+ pthread_cond_wait -+ pthread_cond_destroy -+ localtime_r -+ my_strdup -+ deflate -+ deflateEnd -+ deflateReset -+ deflateInit2_ -+ inflateEnd -+ inflateInit_ -+ inflate -+ compressBound -+ inflateInit2_ -+ adler32 -+ longlong2str -+ strend -+ my_snprintf - -diff -Nur sql/mysqld_x64.def.orig sql/mysqld_x64.def ---- sql/mysqld_x64.def.orig 1969-12-31 18:00:00 -06:00 -+++ sql/mysqld_x64.def 2009-04-09 02:22:04 -05:00 -@@ -0,0 +1,111 @@ -+EXPORTS -+ ?use_hidden_primary_key@handler@@UEAAXXZ -+ ?get_dynamic_partition_info@handler@@UEAAXPEAUPARTITION_INFO@@I@Z -+ ?read_first_row@handler@@UEAAHPEAEI@Z -+ ?read_range_next@handler@@UEAAHXZ -+ ?read_range_first@handler@@UEAAHPEBUst_key_range@@0_N1@Z -+ ?read_multi_range_first@handler@@UEAAHPEAPEAUst_key_multi_range@@PEAU2@I_NPEAUst_handler_buffer@@@Z -+ ?read_multi_range_next@handler@@UEAAHPEAPEAUst_key_multi_range@@@Z -+ ?index_read_idx_map@handler@@UEAAHPEAEIPEBEKW4ha_rkey_function@@@Z -+ ?print_error@handler@@UEAAXHH@Z -+ ?clone@handler@@UEAAPEAV1@PEAUst_mem_root@@@Z -+ ?get_auto_increment@handler@@UEAAX_K00PEA_K1@Z -+ ?index_next_same@handler@@UEAAHPEAEPEBEI@Z -+ ?get_error_message@handler@@UEAA_NHPEAVString@@@Z -+ ?ha_thd@handler@@IEBAPEAVTHD@@XZ -+ ?update_auto_increment@handler@@QEAAHXZ -+ ?ha_statistic_increment@handler@@IEBAXPEQsystem_status_var@@K@Z -+ ?trans_register_ha@@YAXPEAVTHD@@_NPEAUhandlerton@@@Z -+ ?cmp@Field_blob@@QEAAHPEBEI0I@Z -+ ?set_time@Field_timestamp@@QEAAXXZ -+ ?sql_print_error@@YAXPEBDZZ -+ ?sql_print_warning@@YAXPEBDZZ -+ ?check_global_access@@YA_NPEAVTHD@@K@Z -+ ?schema_table_store_record@@YA_NPEAVTHD@@PEAUst_table@@@Z -+ ?get_quote_char_for_identifier@@YAHPEAVTHD@@PEBDI@Z -+ ?copy@String@@QEAA_NXZ -+ ?copy@String@@QEAA_NAEBV1@@Z -+ ?copy@String@@QEAA_NPEBDIPEAUcharset_info_st@@@Z -+ ?copy_and_convert@@YAIPEADIPEAUcharset_info_st@@PEBDI1PEAI@Z -+ ?filename_to_tablename@@YAIPEBDPEADI@Z -+ ?strconvert@@YAIPEAUcharset_info_st@@PEBD0PEADIPEAI@Z -+ ?calculate_key_len@@YAIPEAUst_table@@IPEBEK@Z -+ ?sql_alloc@@YAPEAX_K@Z -+ ?localtime_to_TIME@@YAXPEAUst_mysql_time@@PEAUtm@@@Z -+ ?push_warning@@YAPEAVMYSQL_ERROR@@PEAVTHD@@W4enum_warning_level@1@IPEBD@Z -+ ?push_warning_printf@@YAXPEAVTHD@@W4enum_warning_level@MYSQL_ERROR@@IPEBDZZ -+ ?drop_table@handler@@EEAAXPEBD@Z -+ ?column_bitmaps_signal@handler@@UEAAXXZ -+ ?delete_table@handler@@MEAAHPEBD@Z -+ ?rename_table@handler@@MEAAHPEBD0@Z -+ ?key_map_empty@@3V?$Bitmap@$0EA@@@B -+ ?THR_THD@@3PEAVTHD@@EA -+ ?end_of_list@@3Ulist_node@@A -+ ?mysql_tmpdir_list@@3Ust_my_tmpdir@@A -+ mysql_query_cache_invalidate4 -+ thd_query -+ thd_sql_command -+ thd_get_thread_id -+ thd_get_xid -+ thd_slave_thread -+ thd_non_transactional_update -+ thd_mark_transaction_to_rollback -+ thd_security_context -+ thd_charset -+ thd_test_options -+ thd_ha_data -+ thd_killed -+ thd_tx_isolation -+ thd_tablespace_op -+ thd_sql_command -+ thd_memdup -+ thd_make_lex_string -+ thd_in_lock_tables -+ thd_binlog_format -+ _my_hash_init -+ my_hash_free -+ my_tmpdir -+ check_if_legal_filename -+ my_filename -+ my_sync_dir_by_file -+ alloc_root -+ thr_lock_data_init -+ thr_lock_init -+ thr_lock_delete -+ my_multi_malloc -+ get_charset -+ unpack_filename -+ my_hash_insert -+ my_hash_search -+ my_hash_delete -+ mysql_bin_log_file_pos -+ mysql_bin_log_file_name -+ mysqld_embedded -+ my_thread_name -+ my_malloc -+ my_no_flags_free -+ _sanity -+ _mymalloc -+ _myfree -+ _my_strdup -+ _my_thread_var -+ my_error -+ pthread_cond_init -+ pthread_cond_signal -+ pthread_cond_wait -+ pthread_cond_destroy -+ localtime_r -+ my_strdup -+ deflate -+ deflateEnd -+ deflateReset -+ deflateInit2_ -+ inflateEnd -+ inflateInit_ -+ inflate -+ compressBound -+ inflateInit2_ -+ adler32 -+ longlong2str -+ strend -+ my_snprintf - -diff -Nur win/configure.js.orig win/configure.js ---- win/configure.js.orig 2008-09-26 21:18:37 -05:00 -+++ win/configure.js 2008-10-01 11:21:27 -05:00 -@@ -50,6 +50,7 @@ - case "EMBED_MANIFESTS": - case "EXTRA_DEBUG": - case "WITH_EMBEDDED_SERVER": -+ case "INNODB_DYNAMIC_PLUGIN": - configfile.WriteLine("SET (" + args.Item(i) + " TRUE)"); - break; - case "MYSQL_SERVER_SUFFIX": |