Commit a5868736 authored by unknown's avatar unknown

Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.

Fixes:
- Bug #24712: SHOW TABLE STATUS for file-per-table showing incorrect time fields
- Bug #24386: Performance degradation caused by instrumentation in mutex_struct
- Bug #24190: many exportable definitions of field_in_record_is_null
- Bug #21468: InnoDB crash during recovery with corrupted data pages: XA bug?


storage/innobase/buf/buf0buf.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1045:
  buf_page_init_for_read(): Correct the indentation.
storage/innobase/buf/buf0flu.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1038:
  Port r983 from branches/zip: Enclose some more debug code in
  #ifdef UNIV_SYNC_DEBUG to allow the code to be built
  with UNIV_DEBUG but without UNIV_SYNC_DEBUG.
storage/innobase/dict/dict0dict.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
storage/innobase/ha/ha0ha.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/ha/hash0hash.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/handler/ha_innodb.cc:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1028:
  Merge a change from MySQL AB.
  
  ChangeSet
    2006/11/10 17:22:43+02:00 aelkin@dsl-hkibras-fe30f900-107.dhcp.inet.fi 
    Bug #24190  many exportable definitions of field_in_record_is_null
    
    mysql had several(2) exportable definitions of field_in_record_is_null function.
    
    Fixed with adding static.
  
  storage/innobase/handler/ha_innodb.cc
    2006/11/10 17:22:36+02:00 aelkin@dsl-hkibras-fe30f900-107.dhcp.inet.fi +1 -1
    made static
  
  
  Revision r1008:
  Minor cleanup.
  
  ha_innobase::rnd_pos(): Use correct format in DBUG_PRINT statements.
  
  buf_page_release(): Remove the local variable buf_fix_count.
  
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
  
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
  
  Revision r1134:
  Fix a potential bug in ha_innodb.cc:innobase_query_is_update() where the
  function can be called with "current_thd == NULL". Minor non-functional fix
  in log0recv.c
  
  
  Revision r1098:
  Fix bug #24712: SHOW TABLE STATUS for file-per-table showing incorrect
  time fields
  
  
  Revision r1109:
  ha_innodb.cc: Remove unused define MAX_ULONG_BIT.
storage/innobase/include/btr0sea.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1088:
  Replace the Latin abbreviation "cf." in comments.
storage/innobase/include/buf0buf.ic:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1008:
  Minor cleanup.
  
  ha_innobase::rnd_pos(): Use correct format in DBUG_PRINT statements.
  
  buf_page_release(): Remove the local variable buf_fix_count.
storage/innobase/include/data0type.ic:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1088:
  Replace the Latin abbreviation "cf." in comments.
storage/innobase/include/dict0dict.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
storage/innobase/include/ha0ha.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/include/hash0hash.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/include/sync0rw.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/include/sync0sync.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1038:
  Port r983 from branches/zip: Enclose some more debug code in
  #ifdef UNIV_SYNC_DEBUG to allow the code to be built
  with UNIV_DEBUG but without UNIV_SYNC_DEBUG.
  
  
  Revision r1037:
  Port r972 from branches/zip: Enclose some debug code in #ifdef UNIV_SYNC_DEBUG.
  The code was previously unused in non-debug builds.
  
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
  
  
  Revision r1084:
  Remove the unused constants SYNC_INFINITE_TIME and SYNC_TIME_EXCEEDED.
storage/innobase/include/sync0sync.ic:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/include/univ.i:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1006:
  Introduce #define UNIV_LIST_DEBUG for enabling the debug code in
  UT_LIST_REMOVE_CLEAR().
  
  
  Revision r1088:
  Replace the Latin abbreviation "cf." in comments.
  
  
  Revision r1083:
  univ.i: Document the debug flags (UNIV_DEBUG et al).
storage/innobase/include/ut0lst.h:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1006:
  Introduce #define UNIV_LIST_DEBUG for enabling the debug code in
  UT_LIST_REMOVE_CLEAR().
storage/innobase/lock/lock0lock.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1009:
  Remove duplicate printing of row lock counts in SHOW INNODB STATUS, and in
  the remaining print, add a comma so it doesn't get combined with the heap
  size. Both problems were introduced in r383.
storage/innobase/log/log0recv.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1134:
  Fix a potential bug in ha_innodb.cc:innobase_query_is_update() where the
  function can be called with "current_thd == NULL". Minor non-functional fix
  in log0recv.c
  
  
  Revision r1078:
  Merge r1067:1077 from branches/5.0:
  
  innobase_start_or_create_for_mysql(): Remove unnecessary delay now that
  we moved the setting sync_order_checks_on=TRUE to log0recv.c,
  to the start of the rollback phase in crash recovery.
  
  Fix assertion failure sync0sync.c line 1239
  (the latter ut_error in sync_thread_reset_level())
  in crash recovery when UNIV_SYNC_DEBUG is enabled.
  
  
  Revision r1080:
  Merge r1079 from branches/5.0:
  
  recv_recovery_from_checkpoint_finish(): Add 1 sec delay
  before switching on the sync order checks in crash recovery, so that
  file I/O threads have time to suspend themselves.
storage/innobase/row/row0ins.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
storage/innobase/row/row0sel.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1038:
  Port r983 from branches/zip: Enclose some more debug code in
  #ifdef UNIV_SYNC_DEBUG to allow the code to be built
  with UNIV_DEBUG but without UNIV_SYNC_DEBUG.
  
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
storage/innobase/row/row0upd.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1124:
  Unify dict_table_get_and_increment_handle_count() with dict_table_get() by
  adding a second parameter, adjust callers.
storage/innobase/srv/srv0start.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1038:
  Port r983 from branches/zip: Enclose some more debug code in
  #ifdef UNIV_SYNC_DEBUG to allow the code to be built
  with UNIV_DEBUG but without UNIV_SYNC_DEBUG.
  
  
  Revision r1078:
  Merge r1067:1077 from branches/5.0:
  
  innobase_start_or_create_for_mysql(): Remove unnecessary delay now that
  we moved the setting sync_order_checks_on=TRUE to log0recv.c,
  to the start of the rollback phase in crash recovery.
  
  Fix assertion failure sync0sync.c line 1239
  (the latter ut_error in sync_thread_reset_level())
  in crash recovery when UNIV_SYNC_DEBUG is enabled.
  
  
  Revision r1070:
  Remove another accidentally committed change to srv0start.c.
  The change was accidentally committed with the merge in r1068.
  This revision corresponds to r1051 with a properly merged
  r1067 of branches/5.0.
  
  
  Revision r1068:
  Merge r1067 from branches/5.0:
  
  trx_rollback_for_mysql(), trx_commit_for_mysql():
  Protect the creation of trx_dummy_sess with kernel_mutex.
  This error was introduced in r1046 and r1050.
  
  
  Revision r1069:
  Remove an accidentally committed change to srv0start.c.
  The change was accidentally committed with the merge in r1068.
storage/innobase/sync/sync0rw.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/sync/sync0sync.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1038:
  Port r983 from branches/zip: Enclose some more debug code in
  #ifdef UNIV_SYNC_DEBUG to allow the code to be built
  with UNIV_DEBUG but without UNIV_SYNC_DEBUG.
  
  
  Revision r1037:
  Port r972 from branches/zip: Enclose some debug code in #ifdef UNIV_SYNC_DEBUG.
  The code was previously unused in non-debug builds.
  
  
  Revision r1039:
  Port r1034 from branches/zip: Remove some instrumentation and reduce
  the output of SHOW MUTEX STATUS in non-debug builds.  (Bug #24386)
storage/innobase/trx/trx0roll.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1049:
  Merge r1048 from branches/5.0:
  
  trx_rollback_for_mysql(): Do not set trx->sess back to NULL.
  This bug was introduced in r1046.
  
  
  Revision r1047:
  Merge r1046 from branches/5.0:
  
  branches/5.0: trx_rollback_for_mysql(): Ensure that trx->sess is non-NULL
  when calling trx_general_rollback_for_mysql().  This removes a segmentation
  fault when rolling back a prepared transaction in XA recovery.  (Bug #21468)
  
  
  Revision r1068:
  Merge r1067 from branches/5.0:
  
  trx_rollback_for_mysql(), trx_commit_for_mysql():
  Protect the creation of trx_dummy_sess with kernel_mutex.
  This error was introduced in r1046 and r1050.
  
  
  Revision r1051:
  Merge r1050 from branches/5.0:
  
  trx_rollback_for_mysql(): Fix the comment introduced in r1046.
  
  trx_commit_for_mysql(): Use the dummy trx->sess also for committing a prepared
  transaction in XA recovery, just in case our code would need the session
  object also in that case (does not seem to need it right now).
storage/innobase/trx/trx0trx.c:
  Applied innodb-5.1-ss1039 and innodb-5.1-ss1134 snapshots.
  
  Revision r1009:
  Remove duplicate printing of row lock counts in SHOW INNODB STATUS, and in
  the remaining print, add a comma so it doesn't get combined with the heap
  size. Both problems were introduced in r383.
  
  Revision r1068:
  Merge r1067 from branches/5.0:
  
  trx_rollback_for_mysql(), trx_commit_for_mysql():
  Protect the creation of trx_dummy_sess with kernel_mutex.
  This error was introduced in r1046 and r1050.
  
  
  Revision r1051:
  Merge r1050 from branches/5.0:
  
  trx_rollback_for_mysql(): Fix the comment introduced in r1046.
  
  trx_commit_for_mysql(): Use the dummy trx->sess also for committing a prepared
  transaction in XA recovery, just in case our code would need the session
  object also in that case (does not seem to need it right now).
parent 4dabfa5d
......@@ -1755,22 +1755,22 @@ buf_page_init_for_read(
if (*err == DB_TABLESPACE_DELETED
|| NULL != buf_page_hash_get(space, offset)) {
/* The page belongs to a space which has been
deleted or is being deleted, or the page is
already in buf_pool, return */
/* The page belongs to a space which has been
deleted or is being deleted, or the page is
already in buf_pool, return */
mutex_exit(&block->mutex);
mutex_exit(&(buf_pool->mutex));
mutex_exit(&(buf_pool->mutex));
buf_block_free(block);
buf_block_free(block);
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
mtr_commit(&mtr);
}
mtr_commit(&mtr);
}
return(NULL);
}
return(NULL);
}
ut_ad(block);
......
......@@ -854,8 +854,10 @@ buf_flush_batch(
ut_ad((flush_type == BUF_FLUSH_LRU)
|| (flush_type == BUF_FLUSH_LIST));
#ifdef UNIV_SYNC_DEBUG
ut_ad((flush_type != BUF_FLUSH_LIST)
|| sync_thread_levels_empty_gen(TRUE));
#endif /* UNIV_SYNC_DEBUG */
mutex_enter(&(buf_pool->mutex));
if ((buf_pool->n_flush[flush_type] > 0)
......
......@@ -801,43 +801,20 @@ dict_init(void)
}
/**************************************************************************
Returns a table object. NOTE! This is a high-level function to be used
mainly from outside the 'dict' directory. Inside this directory
dict_table_get_low is usually the appropriate function. */
Returns a table object and optionally increment its MySQL open handle count.
NOTE! This is a high-level function to be used mainly from outside the
'dict' directory. Inside this directory dict_table_get_low is usually the
appropriate function. */
dict_table_t*
dict_table_get(
/*===========*/
/* out: table, NULL if
does not exist */
const char* table_name) /* in: table name */
{
dict_table_t* table;
mutex_enter(&(dict_sys->mutex));
table = dict_table_get_low(table_name);
mutex_exit(&(dict_sys->mutex));
if (table != NULL) {
if (!table->stat_initialized) {
dict_update_statistics(table);
}
}
return(table);
}
/**************************************************************************
Returns a table object and increments MySQL open handle count on the table. */
dict_table_t*
dict_table_get_and_increment_handle_count(
/*======================================*/
/* out: table, NULL if
does not exist */
const char* table_name) /* in: table name */
const char* table_name, /* in: table name */
ibool inc_mysql_count)
/* in: whether to increment the open
handle count on the table */
{
dict_table_t* table;
......@@ -845,15 +822,17 @@ dict_table_get_and_increment_handle_count(
table = dict_table_get_low(table_name);
if (table != NULL) {
if (inc_mysql_count && table) {
table->n_mysql_handles_opened++;
}
mutex_exit(&(dict_sys->mutex));
if (table != NULL) {
if (!table->stat_initialized && !table->ibd_file_missing) {
if (!table->stat_initialized) {
/* If table->ibd_file_missing == TRUE, this will
print an error message and return without doing
anything. */
dict_update_statistics(table);
}
}
......
......@@ -18,16 +18,18 @@ Creates a hash table with >= n array cells. The actual number of cells is
chosen to be a prime number slightly bigger than n. */
hash_table_t*
ha_create(
/*======*/
ha_create_func(
/*===========*/
/* out, own: created table */
ibool in_btr_search, /* in: TRUE if the hash table is used in
the btr_search module */
ulint n, /* in: number of array cells */
ulint n_mutexes, /* in: number of mutexes to protect the
hash table: must be a power of 2, or 0 */
ulint mutex_level) /* in: level of the mutexes in the latching
#ifdef UNIV_SYNC_DEBUG
ulint mutex_level, /* in: level of the mutexes in the latching
order: this is used in the debug version */
#endif /* UNIV_SYNC_DEBUG */
ulint n_mutexes) /* in: number of mutexes to protect the
hash table: must be a power of 2, or 0 */
{
hash_table_t* table;
ulint i;
......
......@@ -129,13 +129,15 @@ hash_table_free(
Creates a mutex array to protect a hash table. */
void
hash_create_mutexes(
/*================*/
hash_create_mutexes_func(
/*=====================*/
hash_table_t* table, /* in: hash table */
ulint n_mutexes, /* in: number of mutexes, must be a
power of 2 */
ulint sync_level) /* in: latching order level of the
#ifdef UNIV_SYNC_DEBUG
ulint sync_level, /* in: latching order level of the
mutexes: used in the debug version */
#endif /* UNIV_SYNC_DEBUG */
ulint n_mutexes) /* in: number of mutexes, must be a
power of 2 */
{
ulint i;
......
......@@ -39,9 +39,6 @@ have disables the InnoDB inlining in this file. */
#include <myisampack.h>
#include <mysys_err.h>
#include <my_sys.h>
#define MAX_ULONG_BIT ((ulong) 1 << (sizeof(ulong)*8-1))
#include "ha_innodb.h"
pthread_mutex_t innobase_share_mutex, /* to protect innobase_open_files */
......@@ -1261,18 +1258,6 @@ trx_is_interrupted(
return(trx && trx->mysql_thd && ((THD*) trx->mysql_thd)->killed);
}
/**************************************************************************
Obtain a pointer to the MySQL THD object, as in current_thd(). This
definition must match the one in sql/ha_innodb.cc! */
extern "C"
void*
innobase_current_thd(void)
/*======================*/
/* out: MySQL THD object */
{
return(current_thd);
}
/*********************************************************************
Call this when you have opened a new table handle in HANDLER, before you
call index_read_idx() etc. Actually, we can let the cursor stay open even
......@@ -2354,7 +2339,7 @@ ha_innobase::open(
/* Get pointer to a table object in InnoDB dictionary cache */
ib_table = dict_table_get_and_increment_handle_count(norm_name);
ib_table = dict_table_get(norm_name, TRUE);
if (NULL == ib_table) {
ut_print_timestamp(stderr);
......@@ -4451,7 +4436,7 @@ ha_innobase::rnd_pos(
}
if (error) {
DBUG_PRINT("error", ("Got error: %d", error));
DBUG_PRINT("error", ("Got error: %d", error));
DBUG_RETURN(error);
}
......@@ -4932,7 +4917,7 @@ ha_innobase::create(
log_buffer_flush_to_disk();
innobase_table = dict_table_get(norm_name);
innobase_table = dict_table_get(norm_name, FALSE);
DBUG_ASSERT(innobase_table != 0);
......@@ -5543,16 +5528,10 @@ ha_innobase::info(
prebuilt->trx->op_info = (char*)
"returning various info to MySQL";
if (ib_table->space != 0) {
my_snprintf(path, sizeof(path), "%s/%s%s",
mysql_data_home, ib_table->name, ".ibd");
unpack_filename(path,path);
} else {
my_snprintf(path, sizeof(path), "%s/%s%s",
my_snprintf(path, sizeof(path), "%s/%s%s",
mysql_data_home, ib_table->name, reg_ext);
unpack_filename(path,path);
}
unpack_filename(path,path);
/* Note that we do not know the access time of the table,
nor the CHECK TABLE time, nor the UPDATE or INSERT time. */
......@@ -6605,22 +6584,23 @@ innodb_mutex_show_status(
{
char buf1[IO_SIZE], buf2[IO_SIZE];
mutex_t* mutex;
#ifdef UNIV_DEBUG
ulint rw_lock_count= 0;
ulint rw_lock_count_spin_loop= 0;
ulint rw_lock_count_spin_rounds= 0;
ulint rw_lock_count_os_wait= 0;
ulint rw_lock_count_os_yield= 0;
ulonglong rw_lock_wait_time= 0;
#endif /* UNIV_DEBUG */
uint hton_name_len= strlen(innobase_hton_name), buf1len, buf2len;
DBUG_ENTER("innodb_mutex_show_status");
#ifdef MUTEX_PROTECT_TO_BE_ADDED_LATER
mutex_enter(&mutex_list_mutex);
#endif
mutex_enter_noninline(&mutex_list_mutex);
mutex = UT_LIST_GET_FIRST(mutex_list);
while (mutex != NULL) {
#ifdef UNIV_DEBUG
if (mutex->mutex_type != 1) {
if (mutex->count_using > 0) {
buf1len= my_snprintf(buf1, sizeof(buf1),
......@@ -6636,14 +6616,13 @@ innodb_mutex_show_status(
mutex->count_spin_rounds,
mutex->count_os_wait,
mutex->count_os_yield,
(ulong) (mutex->lspent_time/1000));
(ulong) (mutex->lspent_time/1000));
if (stat_print(thd, innobase_hton_name,
hton_name_len, buf1, buf1len,
buf2, buf2len)) {
#ifdef MUTEX_PROTECT_TO_BE_ADDED_LATER
mutex_exit(&mutex_list_mutex);
#endif
mutex_exit_noninline(
&mutex_list_mutex);
DBUG_RETURN(1);
}
}
......@@ -6656,26 +6635,39 @@ innodb_mutex_show_status(
rw_lock_count_os_yield += mutex->count_os_yield;
rw_lock_wait_time += mutex->lspent_time;
}
#else /* UNIV_DEBUG */
buf1len= my_snprintf(buf1, sizeof(buf1), "%s:%lu",
mutex->cfile_name, (ulong) mutex->cline);
buf2len= my_snprintf(buf2, sizeof(buf2), "os_waits=%lu",
mutex->count_os_wait);
if (stat_print(thd, innobase_hton_name,
hton_name_len, buf1, buf1len,
buf2, buf2len)) {
mutex_exit_noninline(&mutex_list_mutex);
DBUG_RETURN(1);
}
#endif /* UNIV_DEBUG */
mutex = UT_LIST_GET_NEXT(list, mutex);
}
mutex_exit_noninline(&mutex_list_mutex);
#ifdef UNIV_DEBUG
buf2len= my_snprintf(buf2, sizeof(buf2),
"count=%lu, spin_waits=%lu, spin_rounds=%lu, "
"os_waits=%lu, os_yields=%lu, os_wait_times=%lu",
rw_lock_count, rw_lock_count_spin_loop,
rw_lock_count_spin_rounds,
rw_lock_count_os_wait, rw_lock_count_os_yield,
(ulong) (rw_lock_wait_time/1000));
(ulong) (rw_lock_wait_time/1000));
if (stat_print(thd, innobase_hton_name, hton_name_len,
STRING_WITH_LEN("rw_lock_mutexes"), buf2, buf2len)) {
DBUG_RETURN(1);
}
#ifdef MUTEX_PROTECT_TO_BE_ADDED_LATER
mutex_exit(&mutex_list_mutex);
#endif
#endif /* UNIV_DEBUG */
DBUG_RETURN(FALSE);
}
......@@ -7348,7 +7340,6 @@ innobase_get_at_most_n_mbchars(
}
}
extern "C" {
/**********************************************************************
This function returns true if
......@@ -7358,33 +7349,34 @@ is either REPLACE or LOAD DATA INFILE REPLACE.
2) SQL-query in the current thread
is INSERT ON DUPLICATE KEY UPDATE.
NOTE that /mysql/innobase/row/row0ins.c must contain the
NOTE that storage/innobase/row/row0ins.c must contain the
prototype for this function ! */
extern "C"
ibool
innobase_query_is_update(void)
/*==========================*/
{
THD* thd;
THD* thd = current_thd;
thd = (THD *)innobase_current_thd();
if (!thd) {
/* InnoDB's internal threads may run InnoDB stored procedures
that call this function. Then current_thd is not defined
(it is probably NULL). */
if (thd->lex->sql_command == SQLCOM_REPLACE ||
thd->lex->sql_command == SQLCOM_REPLACE_SELECT ||
(thd->lex->sql_command == SQLCOM_LOAD &&
thd->lex->duplicates == DUP_REPLACE)) {
return(1);
return(FALSE);
}
if (thd->lex->sql_command == SQLCOM_INSERT &&
thd->lex->duplicates == DUP_UPDATE) {
return(1);
switch (thd->lex->sql_command) {
case SQLCOM_REPLACE:
case SQLCOM_REPLACE_SELECT:
return(TRUE);
case SQLCOM_LOAD:
return(thd->lex->duplicates == DUP_REPLACE);
case SQLCOM_INSERT:
return(thd->lex->duplicates == DUP_UPDATE);
default:
return(FALSE);
}
return(0);
}
}
/***********************************************************************
......
......@@ -160,7 +160,7 @@ struct btr_search_struct{
number of full fields */
ulint n_bytes; /* recommended prefix: number of bytes in
an incomplete field;
cf. BTR_PAGE_MAX_REC_SIZE */
see also BTR_PAGE_MAX_REC_SIZE */
ibool left_side; /* TRUE or FALSE, depending on whether
the leftmost record of several records with
the same prefix should be indexed in the
......
......@@ -625,8 +625,6 @@ buf_page_release(
RW_NO_LATCH */
mtr_t* mtr) /* in: mtr */
{
ulint buf_fix_count;
ut_ad(block);
ut_a(block->state == BUF_BLOCK_FILE_PAGE);
......@@ -643,8 +641,7 @@ buf_page_release(
#ifdef UNIV_SYNC_DEBUG
rw_lock_s_unlock(&(block->debug_latch));
#endif
buf_fix_count = block->buf_fix_count;
block->buf_fix_count = buf_fix_count - 1;
block->buf_fix_count--;
mutex_exit(&block->mutex);
......
......@@ -65,8 +65,8 @@ dtype_get_mblen(
innobase_get_cset_width(dtype_get_charset_coll(prtype),
mbminlen, mbmaxlen);
ut_ad(*mbminlen <= *mbmaxlen);
ut_ad(*mbminlen <= 2); /* cf. the bit-field in dtype_t */
ut_ad(*mbmaxlen < 1 << 3); /* cf. the bit-field in dtype_t */
ut_ad(*mbminlen <= 2); /* mbminlen in dtype_t is 0..3 */
ut_ad(*mbmaxlen < 1 << 3); /* mbmaxlen in dtype_t is 0..7 */
#else /* !UNIV_HOTBACKUP */
ut_a(mtype <= DATA_BINARY);
*mbminlen = *mbmaxlen = 1;
......
......@@ -326,26 +326,20 @@ dict_foreign_parse_drop_constraints(
const char*** constraints_to_drop); /* out: id's of the
constraints to drop */
/**************************************************************************
Returns a table object. NOTE! This is a high-level function to be used
mainly from outside the 'dict' directory. Inside this directory
dict_table_get_low is usually the appropriate function. */
Returns a table object and optionally increment its MySQL open handle count.
NOTE! This is a high-level function to be used mainly from outside the
'dict' directory. Inside this directory dict_table_get_low is usually the
appropriate function. */
dict_table_t*
dict_table_get(
/*===========*/
/* out: table, NULL if
does not exist */
const char* table_name); /* in: table name */
/**************************************************************************
Returns a table object and increments MySQL open handle count on the table.
*/
dict_table_t*
dict_table_get_and_increment_handle_count(
/*======================================*/
/* out: table, NULL if
does not exist */
const char* table_name); /* in: table name */
const char* table_name, /* in: table name */
ibool inc_mysql_count);
/* in: whether to increment the open
handle count on the table */
/**************************************************************************
Returns a table object based on table id. */
......
......@@ -41,16 +41,23 @@ Creates a hash table with >= n array cells. The actual number of cells is
chosen to be a prime number slightly bigger than n. */
hash_table_t*
ha_create(
/*======*/
ha_create_func(
/*===========*/
/* out, own: created table */
ibool in_btr_search, /* in: TRUE if the hash table is used in
the btr_search module */
ulint n, /* in: number of array cells */
ulint n_mutexes, /* in: number of mutexes to protect the
hash table: must be a power of 2 */
ulint mutex_level); /* in: level of the mutexes in the latching
#ifdef UNIV_SYNC_DEBUG
ulint mutex_level, /* in: level of the mutexes in the latching
order: this is used in the debug version */
#endif /* UNIV_SYNC_DEBUG */
ulint n_mutexes); /* in: number of mutexes to protect the
hash table: must be a power of 2 */
#ifdef UNIV_SYNC_DEBUG
# define ha_create(b,n_c,n_m,level) ha_create_func(b,n_c,level,n_m)
#else /* UNIV_SYNC_DEBUG */
# define ha_create(b,n_c,n_m,level) ha_create_func(b,n_c,n_m)
#endif /* UNIV_SYNC_DEBUG */
/*****************************************************************
Inserts an entry into a hash table. If an entry with the same fold number
is found, its node is updated to point to the new data, and no new node
......
......@@ -31,12 +31,20 @@ hash_create(
Creates a mutex array to protect a hash table. */
void
hash_create_mutexes(
/*================*/
hash_create_mutexes_func(
/*=====================*/
hash_table_t* table, /* in: hash table */
ulint n_mutexes, /* in: number of mutexes */
ulint sync_level); /* in: latching order level of the
#ifdef UNIV_SYNC_DEBUG
ulint sync_level, /* in: latching order level of the
mutexes: used in the debug version */
#endif /* UNIV_SYNC_DEBUG */
ulint n_mutexes); /* in: number of mutexes */
#ifdef UNIV_SYNC_DEBUG
# define hash_create_mutexes(t,n,level) hash_create_mutexes_func(t,level,n)
#else /* UNIV_SYNC_DEBUG */
# define hash_create_mutexes(t,n,level) hash_create_mutexes_func(t,n)
#endif /* UNIV_SYNC_DEBUG */
/*****************************************************************
Frees a hash table. */
......
......@@ -61,8 +61,18 @@ Creates, or rather, initializes an rw-lock object in a specified memory
location (which must be appropriately aligned). The rw-lock is initialized
to the non-locked state. Explicit freeing of the rw-lock with rw_lock_free
is necessary only if the memory block containing it is freed. */
#define rw_lock_create(L, level) \
rw_lock_create_func((L), (level), __FILE__, __LINE__, #L)
#ifdef UNIV_DEBUG
# ifdef UNIV_SYNC_DEBUG
# define rw_lock_create(L, level) \
rw_lock_create_func((L), (level), #L, __FILE__, __LINE__)
# else /* UNIV_SYNC_DEBUG */
# define rw_lock_create(L, level) \
rw_lock_create_func((L), #L, __FILE__, __LINE__)
# endif /* UNIV_SYNC_DEBUG */
#else /* UNIV_DEBUG */
# define rw_lock_create(L, level) \
rw_lock_create_func((L), __FILE__, __LINE__)
#endif /* UNIV_DEBUG */
/**********************************************************************
Creates, or rather, initializes an rw-lock object in a specified memory
......@@ -74,10 +84,14 @@ void
rw_lock_create_func(
/*================*/
rw_lock_t* lock, /* in: pointer to memory */
#ifdef UNIV_DEBUG
# ifdef UNIV_SYNC_DEBUG
ulint level, /* in: level */
# endif /* UNIV_SYNC_DEBUG */
const char* cmutex_name, /* in: mutex name */
#endif /* UNIV_DEBUG */
const char* cfile_name, /* in: file name where created */
ulint cline, /* in: file line where created */
const char* cmutex_name); /* in: mutex name */
ulint cline); /* in: file line where created */
/**********************************************************************
Calling this function is obligatory only if the memory buffer containing
the rw-lock is freed. Removes an rw-lock object from the global list. The
......
......@@ -39,8 +39,18 @@ location (which must be appropriately aligned). The mutex is initialized
in the reset state. Explicit freeing of the mutex with mutex_free is
necessary only if the memory block containing it is freed. */
#define mutex_create(M, level) \
mutex_create_func((M), (level), __FILE__, __LINE__, #M)
#ifdef UNIV_DEBUG
# ifdef UNIV_SYNC_DEBUG
# define mutex_create(M, level) \
mutex_create_func((M), #M, (level), __FILE__, __LINE__)
# else
# define mutex_create(M, level) \
mutex_create_func((M), #M, __FILE__, __LINE__)
# endif
#else
# define mutex_create(M, level) \
mutex_create_func((M), __FILE__, __LINE__)
#endif
/**********************************************************************
Creates, or rather, initializes a mutex object in a specified memory
......@@ -52,10 +62,14 @@ void
mutex_create_func(
/*==============*/
mutex_t* mutex, /* in: pointer to memory */
#ifdef UNIV_DEBUG
const char* cmutex_name, /* in: mutex name */
# ifdef UNIV_SYNC_DEBUG
ulint level, /* in: level */
# endif /* UNIV_SYNC_DEBUG */
#endif /* UNIV_DEBUG */
const char* cfile_name, /* in: file name where created */
ulint cline, /* in: file line where created */
const char* cmutex_name); /* in: mutex name */
ulint cline); /* in: file line where created */
/**********************************************************************
Calling this function is obligatory only if the memory buffer containing
the mutex is freed. Removes a mutex object from the mutex list. The mutex
......@@ -149,6 +163,7 @@ void
sync_print(
/*=======*/
FILE* file); /* in: file where to print */
#ifdef UNIV_DEBUG
/**********************************************************************
Checks that the mutex has been initialized. */
......@@ -156,6 +171,8 @@ ibool
mutex_validate(
/*===========*/
mutex_t* mutex);
#endif /* UNIV_DEBUG */
#ifdef UNIV_SYNC_DEBUG
/**********************************************************************
Adds a latch and its level in the thread level array. Allocates the memory
for the array if called first time for this OS thread. Makes the checks
......@@ -197,7 +214,6 @@ sync_thread_levels_empty_gen(
allowed to be owned by the thread,
also purge_is_running mutex is
allowed */
#ifdef UNIV_SYNC_DEBUG
/**********************************************************************
Checks that the current thread owns the mutex. Works only
in the debug version. */
......@@ -465,26 +481,29 @@ struct mutex_struct {
ulint line; /* Line where the mutex was locked */
os_thread_id_t thread_id; /* Debug version: The thread id of the
thread which locked the mutex. */
#endif /* UNIV_SYNC_DEBUG */
ulint level; /* Level in the global latching order */
#endif /* UNIV_SYNC_DEBUG */
const char* cfile_name;/* File name where mutex created */
ulint cline; /* Line where created */
ulint magic_n;
ulint cline; /* Line where created */
#ifdef UNIV_DEBUG
ulint magic_n;
# define MUTEX_MAGIC_N (ulint)979585
#endif /* UNIV_DEBUG */
#ifndef UNIV_HOTBACKUP
ulong count_using; /* count of times mutex used */
ulong count_spin_loop; /* count of spin loops */
ulong count_spin_rounds; /* count of spin rounds */
ulong count_os_wait; /* count of os_wait */
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 */
ulong count_os_wait; /* count of os_wait */
# ifdef UNIV_DEBUG
ulong count_using; /* count of times mutex used */
ulong count_spin_loop; /* count of spin loops */
ulong count_spin_rounds; /* count of spin rounds */
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 */
#endif /* !UNIV_HOTBACKUP */
};
#define MUTEX_MAGIC_N (ulint)979585
/* The global array of wait cells for implementation of the databases own
mutexes and read-write locks. Appears here for debugging purposes only! */
......@@ -496,20 +515,16 @@ to 20 microseconds. */
#define SYNC_SPIN_ROUNDS srv_n_spin_wait_rounds
#define SYNC_INFINITE_TIME ((ulint)(-1))
/* Means that a timeout elapsed when waiting */
#define SYNC_TIME_EXCEEDED (ulint)1
/* The number of system calls made in this module. Intended for performance
monitoring. */
extern ulint mutex_system_call_count;
extern ulint mutex_exit_count;
#ifdef UNIV_SYNC_DEBUG
/* Latching order checks start when this is set TRUE */
extern ibool sync_order_checks_on;
#endif /* UNIV_SYNC_DEBUG */
/* This variable is set to TRUE when sync_init is called */
extern ibool sync_initialized;
......
......@@ -254,9 +254,9 @@ mutex_enter_func(
/* Note that we do not peek at the value of lock_word before trying
the atomic test_and_set; we could peek, and possibly save time. */
#ifndef UNIV_HOTBACKUP
#if defined UNIV_DEBUG && !defined UNIV_HOTBACKUP
mutex->count_using++;
#endif /* UNIV_HOTBACKUP */
#endif /* UNIV_DEBUG && !UNIV_HOTBACKUP */
if (!mutex_test_and_set(mutex)) {
#ifdef UNIV_SYNC_DEBUG
......
......@@ -83,19 +83,27 @@ memory is read outside the allocated blocks. */
/* Make a non-inline debug version */
#if 0
#define UNIV_DEBUG
#define UNIV_MEM_DEBUG
#define UNIV_IBUF_DEBUG
#define UNIV_SYNC_DEBUG
#define UNIV_SEARCH_DEBUG
#define UNIV_SYNC_PERF_STAT
#define UNIV_SEARCH_PERF_STAT
#define UNIV_SRV_PRINT_LATCH_WAITS
#define UNIV_BTR_PRINT
#define UNIV_DEBUG /* Enable ut_ad() assertions */
#define UNIV_LIST_DEBUG /* debug UT_LIST_ macros */
#define UNIV_MEM_DEBUG /* detect memory leaks etc */
#define UNIV_IBUF_DEBUG /* debug the insert buffer;
this limits the database to IBUF_COUNT_N_SPACES and IBUF_COUNT_N_PAGES,
and the insert buffer must be empty when the database is started */
#define UNIV_SYNC_DEBUG /* debug mutex and latch
operations (very slow); also UNIV_DEBUG must be defined */
#define UNIV_SEARCH_DEBUG /* debug B-tree comparisons */
#define UNIV_SYNC_PERF_STAT /* operation counts for
rw-locks and mutexes */
#define UNIV_SEARCH_PERF_STAT /* statistics for the
adaptive hash index */
#define UNIV_SRV_PRINT_LATCH_WAITS /* enable diagnostic output
in sync0sync.c */
#define UNIV_BTR_PRINT /* enable functions for
printing B-trees */
#endif
#define UNIV_BTR_DEBUG
#define UNIV_LIGHT_MEM_DEBUG
#define UNIV_BTR_DEBUG /* check B-tree links */
#define UNIV_LIGHT_MEM_DEBUG /* light memory debugging */
#ifdef HAVE_purify
/* The following sets all new allocated memory to zero before use:
......
......@@ -124,7 +124,7 @@ name, NODE1 and NODE2 are pointers to nodes. */
}\
/* Invalidate the pointers in a list node. */
#ifdef UNIV_DEBUG
#ifdef UNIV_LIST_DEBUG
# define UT_LIST_REMOVE_CLEAR(NAME, N) \
((N)->NAME.prev = (N)->NAME.next = (void*) -1)
#else
......
......@@ -4391,10 +4391,6 @@ lock_print_info_all_transactions(
trx->read_view->up_limit_id));
}
fprintf(file,
"Trx has approximately %lu row locks\n",
(ulong) lock_number_of_rows_locked(trx));
if (trx->que_state == TRX_QUE_LOCK_WAIT) {
fprintf(file,
"------- TRX HAS BEEN WAITING %lu SEC"
......
......@@ -33,6 +33,7 @@ Created 9/20/1997 Heikki Tuuri
#include "btr0cur.h"
#include "dict0boot.h"
#include "fil0fil.h"
#include "sync0sync.h"
#ifdef UNIV_HOTBACKUP
/* This is set to FALSE if the backup was originally taken with the
......@@ -190,6 +191,7 @@ recv_sys_empty_hash(void)
recv_sys->addr_hash = hash_create(buf_pool_get_curr_size() / 256);
}
#ifndef UNIV_LOG_DEBUG
/************************************************************
Frees the recovery system. */
static
......@@ -209,6 +211,7 @@ recv_sys_free(void)
mutex_exit(&(recv_sys->mutex));
}
#endif /* UNIV_LOG_DEBUG */
/************************************************************
Truncates possible corrupted or extra records from a log group. */
......@@ -2855,6 +2858,15 @@ recv_recovery_from_checkpoint_finish(void)
#ifndef UNIV_LOG_DEBUG
recv_sys_free();
#endif
#ifdef UNIV_SYNC_DEBUG
/* Wait for a while so that created threads have time to suspend
themselves before we switch the latching order checks on */
os_thread_sleep(1000000);
/* Switch latching order checks on in sync0sync.c */
sync_order_checks_on = TRUE;
#endif
if (srv_force_recovery < SRV_FORCE_NO_TRX_UNDO) {
/* Rollback the uncommitted transactions which have no user
session */
......
......@@ -1524,7 +1524,8 @@ row_ins_check_foreign_constraints(
if (foreign->foreign_index == index) {
if (foreign->referenced_table == NULL) {
dict_table_get(foreign->referenced_table_name);
dict_table_get(foreign->referenced_table_name,
FALSE);
}
if (0 == trx->dict_operation_lock_mode) {
......
......@@ -1831,7 +1831,9 @@ row_sel(
mtr_commit(&mtr);
#ifdef UNIV_SYNC_DEBUG
ut_ad(sync_thread_levels_empty_gen(TRUE));
#endif /* UNIV_SYNC_DEBUG */
err = DB_SUCCESS;
goto func_exit;
......@@ -1850,7 +1852,9 @@ row_sel(
leaf_contains_updates = FALSE;
mtr_has_extra_clust_latch = FALSE;
#ifdef UNIV_SYNC_DEBUG
ut_ad(sync_thread_levels_empty_gen(TRUE));
#endif /* UNIV_SYNC_DEBUG */
goto table_loop;
......@@ -1866,7 +1870,9 @@ row_sel(
mtr_commit(&mtr);
#ifdef UNIV_SYNC_DEBUG
ut_ad(sync_thread_levels_empty_gen(TRUE));
#endif /* UNIV_SYNC_DEBUG */
func_exit:
if (UNIV_LIKELY_NULL(heap)) {
......@@ -4447,7 +4453,7 @@ row_search_check_if_query_cache_permitted(
dict_table_t* table;
ibool ret = FALSE;
table = dict_table_get(norm_name);
table = dict_table_get(norm_name, FALSE);
if (table == NULL) {
......
......@@ -202,7 +202,8 @@ row_upd_check_references_constraints(
foreign->n_fields))) {
if (foreign->foreign_table == NULL) {
dict_table_get(foreign->foreign_table_name);
dict_table_get(foreign->foreign_table_name,
FALSE);
}
if (foreign->foreign_table) {
......
......@@ -1601,17 +1601,6 @@ innobase_start_or_create_for_mysql(void)
srv_was_started = TRUE;
srv_is_being_started = FALSE;
#ifdef UNIV_DEBUG
/* Wait a while so that the created threads have time to suspend
themselves before we switch sync debugging on; otherwise a thread may
execute mutex_enter() before the checks are on, and mutex_exit() after
the checks are on, which will cause an assertion failure in sync
debug. */
os_thread_sleep(3000000);
#endif
sync_order_checks_on = TRUE;
if (trx_doublewrite == NULL) {
/* Create the doublewrite buffer to a new tablespace */
......
......@@ -89,11 +89,14 @@ void
rw_lock_create_func(
/*================*/
rw_lock_t* lock, /* in: pointer to memory */
ulint level __attribute__((unused)),
/* in: level */
#ifdef UNIV_DEBUG
# ifdef UNIV_SYNC_DEBUG
ulint level, /* in: level */
# endif /* UNIV_SYNC_DEBUG */
const char* cmutex_name, /* in: mutex name */
#endif /* UNIV_DEBUG */
const char* cfile_name, /* in: file name where created */
ulint cline, /* in: file line where created */
const char* cmutex_name) /* in: mutex name */
ulint cline) /* in: file line where created */
{
/* If this is the very first time a synchronization object is
created, then the following call initializes the sync system. */
......@@ -103,10 +106,10 @@ rw_lock_create_func(
lock->mutex.cfile_name = cfile_name;
lock->mutex.cline = cline;
#ifndef UNIV_HOTBACKUP
#if defined UNIV_DEBUG && !defined UNIV_HOTBACKUP
lock->mutex.cmutex_name = cmutex_name;
lock->mutex.mutex_type = 1;
#endif /* !UNIV_HOTBACKUP */
#endif /* UNIV_DEBUG && !UNIV_HOTBACKUP */
rw_lock_set_waiters(lock, 0);
rw_lock_set_writer(lock, RW_LOCK_NOT_LOCKED);
......
This diff is collapsed.
......@@ -131,8 +131,26 @@ trx_rollback_for_mysql(
trx->op_info = "rollback";
err = trx_general_rollback_for_mysql(trx, FALSE, NULL);
/* If we are doing the XA recovery of prepared transactions, then
the transaction object does not have an InnoDB session object, and we
set a dummy session that we use for all MySQL transactions. */
mutex_enter(&kernel_mutex);
if (trx->sess == NULL) {
/* Open a dummy session */
if (!trx_dummy_sess) {
trx_dummy_sess = sess_open();
}
trx->sess = trx_dummy_sess;
}
mutex_exit(&kernel_mutex);
err = trx_general_rollback_for_mysql(trx, FALSE, NULL);
trx->op_info = "";
return(err);
......
......@@ -1598,6 +1598,24 @@ trx_commit_for_mysql(
trx->op_info = "committing";
/* If we are doing the XA recovery of prepared transactions, then
the transaction object does not have an InnoDB session object, and we
set the dummy session that we use for all MySQL transactions. */
mutex_enter(&kernel_mutex);
if (trx->sess == NULL) {
/* Open a dummy session */
if (!trx_dummy_sess) {
trx_dummy_sess = sess_open();
}
trx->sess = trx_dummy_sess;
}
mutex_exit(&kernel_mutex);
trx_start_if_not_started(trx);
mutex_enter(&kernel_mutex);
......@@ -1761,11 +1779,10 @@ trx_print(
|| mem_heap_get_size(trx->lock_heap) > 400) {
newline = TRUE;
fprintf(f, "%lu lock struct(s), heap size %lu",
fprintf(f, "%lu lock struct(s), heap size %lu,"
" %lu row lock(s)",
(ulong) UT_LIST_GET_LEN(trx->trx_locks),
(ulong) mem_heap_get_size(trx->lock_heap));
fprintf(f, "%lu row lock(s)",
(ulong) mem_heap_get_size(trx->lock_heap),
(ulong) lock_number_of_rows_locked(trx));
}
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment