mirror of
https://github.com/MariaDB/server.git
synced 2025-07-27 21:55:03 +02:00

Added option 'aria-pagecache-segments', default 1. For values > 1, this split the aria-pagecache-buffer into the given number of segments, each independent from each other. Having multiple pagecaches improve performance when multiple connections runs queries concurrently using different tables. Each pagecache will use aria-pageache-buffer/segments amount of memory, however at least 128K. Each opened table has its index and data file use the segments in a a round-robin fashion. Internal changes: - All programs allocating the maria pagecache themselves should now call multi_init_pagecache() instead of init_pagecache(). - pagecache statistics is now stored in 'pagecache_stats' instead of maria_pagecache. One must call multi_update_pagecache_stats() to update the statistics. - Added into PAGECACHE_FILE a pointer to files pagecache. This was done to ensure that index and data file are using the same pagecache and simplified the checkpoint code. I kept pagecache in TABLE_SHARE to minimize the changes. - really_execute_checkpoint() was update to handle a dynamic number of pagecaches. - pagecache_collect_changed_blocks_with_lsn() was slight changed to allow it to be called for each pagecache. - undefined not used functions maria_assign_pagecache() and maria_change_pagecache() - ma_pagecaches.c is totally rewritten. It now contains all multi_pagecache functions. Errors found be QA that are fixed: MDEV-36872 UBSAN errors in ma_checkpoint.c MDEV-36874 Behavior upon too small aria_pagecache_buffer_size in case of multiple segments is not very user-friendly MDEV-36914 ma_checkpoint.c(285,9): conversion from '__int64' to 'uint' treated as an error MDEV-36912 sys_vars.sysvars_server_embedded and sys_vars.sysvars_server_notembedded fail on x86
168 lines
5.3 KiB
C
168 lines
5.3 KiB
C
/* Copyright (C) 2006 MySQL AB
|
|
|
|
This program is free software; you can redistribute it and/or modify
|
|
it under the terms of the GNU General Public License as published by
|
|
the Free Software Foundation; version 2 of the License.
|
|
|
|
This program is distributed in the hope that it will be useful,
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
GNU General Public License for more details.
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
along with this program; if not, write to the Free Software
|
|
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335 USA */
|
|
|
|
/*
|
|
Key cache assignments
|
|
*/
|
|
|
|
#include "maria_def.h"
|
|
|
|
/*
|
|
Assign pages of the index file for a table to a key cache
|
|
|
|
SYNOPSIS
|
|
maria_assign_to_pagecache()
|
|
info open table
|
|
key_map map of indexes to assign to the key cache
|
|
pagecache_ptr pointer to the key cache handle
|
|
assign_lock Mutex to lock during assignment
|
|
|
|
PREREQUESTS
|
|
One must have a READ lock or a WRITE lock on the table when calling
|
|
the function to ensure that there is no other writers to it.
|
|
|
|
The caller must also ensure that one doesn't call this function from
|
|
two different threads with the same table.
|
|
|
|
NOTES
|
|
At present pages for all indexes must be assigned to the same key cache.
|
|
In future only pages for indexes specified in the key_map parameter
|
|
of the table will be assigned to the specified key cache.
|
|
|
|
RETURN VALUE
|
|
0 If a success
|
|
# Error code
|
|
*/
|
|
|
|
#ifdef NOT_USED
|
|
int maria_assign_to_pagecache(MARIA_HA *info,
|
|
ulonglong key_map __attribute__((unused)),
|
|
PAGECACHE *pagecache)
|
|
{
|
|
int error= 0;
|
|
MARIA_SHARE* share= info->s;
|
|
DBUG_ENTER("maria_assign_to_pagecache");
|
|
DBUG_PRINT("enter",
|
|
("old_pagecache_handle:%p new_pagecache_handle:%p",
|
|
share->pagecache, pagecache));
|
|
|
|
/*
|
|
Skip operation if we didn't change key cache. This can happen if we
|
|
call this for all open instances of the same table
|
|
*/
|
|
if (share->pagecache == pagecache)
|
|
DBUG_RETURN(0);
|
|
|
|
/*
|
|
First flush all blocks for the table in the old key cache.
|
|
This is to ensure that the disk is consistent with the data pages
|
|
in memory (which may not be the case if the table uses delayed_key_write)
|
|
|
|
Note that some other read thread may still fill in the key cache with
|
|
new blocks during this call and after, but this doesn't matter as
|
|
all threads will start using the new key cache for their next call to
|
|
maria library and we know that there will not be any changed blocks
|
|
in the old key cache.
|
|
*/
|
|
|
|
if (flush_pagecache_blocks(share->pagecache, &share->kfile, FLUSH_RELEASE))
|
|
{
|
|
error= my_errno;
|
|
/* Mark that table must be checked */
|
|
_ma_set_fatal_error(info, error);
|
|
}
|
|
|
|
/*
|
|
Flush the new key cache for this file. This is needed to ensure
|
|
that there is no old blocks (with outdated data) left in the new key
|
|
cache from an earlier assign_to_keycache operation
|
|
|
|
(This can never fail as there is never any not written data in the
|
|
new key cache)
|
|
*/
|
|
(void) flush_pagecache_blocks(pagecache, &share->kfile, FLUSH_RELEASE);
|
|
|
|
/*
|
|
ensure that setting the key cache and changing the multi_pagecache
|
|
is done atomicly
|
|
*/
|
|
mysql_mutex_lock(&share->intern_lock);
|
|
/*
|
|
Tell all threads to use the new key cache
|
|
This should be seen at the lastes for the next call to an maria function.
|
|
*/
|
|
share->pagecache= pagecache;
|
|
|
|
/* store the key cache in the global hash structure for future opens */
|
|
if (multi_pagecache_set((uchar*) share->unique_file_name.str,
|
|
(uint)share->unique_file_name.length,
|
|
share->pagecache))
|
|
error= my_errno;
|
|
mysql_mutex_unlock(&share->intern_lock);
|
|
DBUG_RETURN(error);
|
|
}
|
|
#endif
|
|
|
|
|
|
/*
|
|
Change all MARIA entries that uses one key cache to another key cache
|
|
|
|
SYNOPSIS
|
|
maria_change_pagecache()
|
|
old_pagecache Old key cache
|
|
new_pagecache New key cache
|
|
|
|
NOTES
|
|
This is used when we delete one key cache.
|
|
|
|
To handle the case where some other threads tries to open an MARIA
|
|
table associated with the to-be-deleted key cache while this operation
|
|
is running, we have to call 'multi_pagecache_change()' from this
|
|
function while we have a lock on the MARIA table list structure.
|
|
|
|
This is safe as long as it's only MARIA that is using this specific
|
|
key cache.
|
|
*/
|
|
|
|
|
|
#ifdef NOT_USED
|
|
void maria_change_pagecache(PAGECACHE *old_pagecache,
|
|
PAGECACHE *new_pagecache)
|
|
{
|
|
LIST *pos;
|
|
DBUG_ENTER("maria_change_pagecache");
|
|
|
|
/*
|
|
Lock list to ensure that no one can close the table while we manipulate it
|
|
*/
|
|
mysql_mutex_lock(&THR_LOCK_maria);
|
|
for (pos=maria_open_list ; pos ; pos=pos->next)
|
|
{
|
|
MARIA_HA *info= (MARIA_HA*) pos->data;
|
|
MARIA_SHARE *share= info->s;
|
|
if (share->pagecache == old_pagecache)
|
|
maria_assign_to_pagecache(info, (ulonglong) ~0, new_pagecache);
|
|
}
|
|
|
|
/*
|
|
We have to do the following call while we have the lock on the
|
|
MARIA list structure to ensure that another thread is not trying to
|
|
open a new table that will be associted with the old key cache
|
|
*/
|
|
multi_pagecache_change(old_pagecache, new_pagecache);
|
|
mysql_mutex_unlock(&THR_LOCK_maria);
|
|
DBUG_VOID_RETURN;
|
|
}
|
|
#endif
|