mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-20 14:02:32 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/buf/buf0buf.c

3410 lines
89 KiB
C
Raw Normal View History

Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Innobase relational database engine; Copyright (C) 2001 Innobase Oy
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License 2
as published by the Free Software Foundation in June 1991.
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
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.
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
You should have received a copy of the GNU General Public License 2
along with this program (in file COPYING); if not, write to the Free
Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
/******************************************************
The database buffer buf_pool
(c) 1995 Innobase Oy
Created 11/5/1995 Heikki Tuuri
*******************************************************/
#include "buf0buf.h"
#ifdef UNIV_NONINL
#include "buf0buf.ic"
#endif
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
#include "buf0buddy.h"
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#include "mem0mem.h"
#include "btr0btr.h"
#include "fil0fil.h"
#include "lock0lock.h"
#include "btr0sea.h"
#include "ibuf0ibuf.h"
#include "dict0dict.h"
#include "log0recv.h"
#include "log0log.h"
#include "trx0undo.h"
#include "srv0srv.h"
branches/zip: Initial import of the zip-like compression of B-tree pages.
2005-10-27 13:48:10 +02:00
#include "page0zip.h"
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/*
IMPLEMENTATION OF THE BUFFER POOL
=================================
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Performance improvement:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
------------------------
Thread scheduling in NT may be so slow that the OS wait mechanism should
not be used even in waiting for disk reads to complete.
Rather, we should put waiting query threads to the queue of
waiting jobs, and let the OS thread do something useful while the i/o
is processed. In this way we could remove most OS thread switches in
an i/o-intensive benchmark like TPC-C.
A possibility is to put a user space thread library between the database
and NT. User space thread libraries might be very fast.
SQL Server 7.0 can be configured to use 'fibers' which are lightweight
threads in NT. These should be studied.
Buffer frames and blocks
------------------------
Following the terminology of Gray and Reuter, we call the memory
blocks where file pages are loaded buffer frames. For each buffer
frame there is a control block, or shortly, a block, in the buffer
control array. The control info which does not need to be stored
in the file along with the file page, resides in the control block.
Buffer pool struct
------------------
The buffer buf_pool contains a single mutex which protects all the
control data structures of the buf_pool. The content of a buffer frame is
protected by a separate read-write lock in its control block, though.
These locks can be locked and unlocked without owning the buf_pool mutex.
The OS events in the buf_pool struct can be waited for without owning the
buf_pool mutex.
The buf_pool mutex is a hot-spot in main memory, causing a lot of
memory bus traffic on multiprocessor systems when processors
alternately access the mutex. On our Pentium, the mutex is accessed
maybe every 10 microseconds. We gave up the solution to have mutexes
for each control block, for instance, because it seemed to be
complicated.
A solution to reduce mutex contention of the buf_pool mutex is to
create a separate mutex for the page hash table. On Pentium,
accessing the hash table takes 2 microseconds, about half
of the total buf_pool mutex hold time.
Control blocks
--------------
The control block contains, for instance, the bufferfix count
which is incremented when a thread wants a file page to be fixed
in a buffer frame. The bufferfix operation does not lock the
contents of the frame, however. For this purpose, the control
block contains a read-write lock.
The buffer frames have to be aligned so that the start memory
address of a frame is divisible by the universal page size, which
is a power of two.
We intend to make the buffer buf_pool size on-line reconfigurable,
that is, the buf_pool size can be changed without closing the database.
Then the database administarator may adjust it to be bigger
at night, for example. The control block array must
contain enough control blocks for the maximum buffer buf_pool size
which is used in the particular database.
If the buf_pool size is cut, we exploit the virtual memory mechanism of
the OS, and just refrain from using frames at high addresses. Then the OS
can swap them to disk.
The control blocks containing file pages are put to a hash table
according to the file address of the page.
We could speed up the access to an individual page by using
"pointer swizzling": we could replace the page references on
non-leaf index pages by direct pointers to the page, if it exists
in the buf_pool. We could make a separate hash table where we could
chain all the page references in non-leaf pages residing in the buf_pool,
using the page reference as the hash key,
and at the time of reading of a page update the pointers accordingly.
Drawbacks of this solution are added complexity and,
possibly, extra space required on non-leaf pages for memory pointers.
A simpler solution is just to speed up the hash table mechanism
in the database, using tables whose size is a power of 2.
Lists of blocks
---------------
There are several lists of control blocks. The free list contains
blocks which are currently not used.
The LRU-list contains all the blocks holding a file page
except those for which the bufferfix count is non-zero.
The pages are in the LRU list roughly in the order of the last
access to the page, so that the oldest pages are at the end of the
list. We also keep a pointer to near the end of the LRU list,
which we can use when we want to artificially age a page in the
buf_pool. This is used if we know that some page is not needed
again for some time: we insert the block right after the pointer,
causing it to be replaced sooner than would noramlly be the case.
Currently this aging mechanism is used for read-ahead mechanism
of pages, and it can also be used when there is a scan of a full
table which cannot fit in the memory. Putting the pages near the
of the LRU list, we make sure that most of the buf_pool stays in the
main memory, undisturbed.
The chain of modified blocks contains the blocks
holding file pages that have been modified in the memory
but not written to disk yet. The block with the oldest modification
which has not yet been written to disk is at the end of the chain.
Loading a file page
-------------------
First, a victim block for replacement has to be found in the
buf_pool. It is taken from the free list or searched for from the
end of the LRU-list. An exclusive lock is reserved for the frame,
the io_fix field is set in the block fixing the block in buf_pool,
and the io-operation for loading the page is queued. The io-handler thread
releases the X-lock on the frame and resets the io_fix field
when the io operation completes.
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
A thread may request the above operation using the function
buf_page_get(). It may then continue to request a lock on the frame.
Import 5.0 code.
2005-10-27 09:29:40 +02:00
The lock is granted when the io-handler releases the x-lock.
Read-ahead
----------
The read-ahead mechanism is intended to be intelligent and
isolated from the semantically higher levels of the database
index management. From the higher level we only need the
information if a file page has a natural successor or
predecessor page. On the leaf level of a B-tree index,
these are the next and previous pages in the natural
order of the pages.
Let us first explain the read-ahead mechanism when the leafs
of a B-tree are scanned in an ascending or descending order.
When a read page is the first time referenced in the buf_pool,
the buffer manager checks if it is at the border of a so-called
linear read-ahead area. The tablespace is divided into these
areas of size 64 blocks, for example. So if the page is at the
border of such an area, the read-ahead mechanism checks if
all the other blocks in the area have been accessed in an
ascending or descending order. If this is the case, the system
looks at the natural successor or predecessor of the page,
checks if that is at the border of another area, and in this case
issues read-requests for all the pages in that area. Maybe
we could relax the condition that all the pages in the area
have to be accessed: if data is deleted from a table, there may
appear holes of unused pages in the area.
A different read-ahead mechanism is used when there appears
to be a random access pattern to a file.
If a new page is referenced in the buf_pool, and several pages
of its random access area (for instance, 32 consecutive pages
in a tablespace) have recently been referenced, we may predict
that the whole area may be needed in the near future, and issue
the read requests for the whole area.
*/
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
/* Value in microseconds */
static const int WAIT_FOR_READ = 20000;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool_t* buf_pool = NULL; /* The buffer buf_pool of the database */
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
static ulint buf_dbg_counter = 0; /* This is used to insert validation
Import 5.0 code.
2005-10-27 09:29:40 +02:00
operations in excution in the
debug version */
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
#ifdef UNIV_DEBUG
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ibool buf_debug_prints = FALSE; /* If this is set TRUE,
the program prints info whenever
read-ahead or flush occurs */
#endif /* UNIV_DEBUG */
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* A chunk of buffers. The buffer pool is allocated in chunks. */
struct buf_chunk_struct{
ulint mem_size; /* allocated size of the chunk */
ulint size; /* size of frames[] and blocks[] */
void* mem; /* pointer to the memory area which
was allocated for the frames */
buf_block_t* blocks; /* array of buffer control blocks */
};
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Calculates a page checksum which is stored to the page when it is written
to a file. Note that we must be careful to calculate the same value on
32-bit and 64-bit architectures. */
ulint
buf_calc_page_new_checksum(
/*=======================*/
branches/zip: Minor cleanup. buf_calc_page_new_checksum(), buf_calc_page_old_checksum(), buf_page_is_corrupted(), buf_page_print(): Add const qualifier to byte* parameter. page_parse_create_zip(): Remove this declaration of non-existing function. page_rec_check(), page_rec_print(), row_vers_impl_x_locked_off_kernel(): Add const qualifier to rec_t* parameter. row_vers_old_has_index_entry(): Add const qualifier to dtuple_t* parameter. row_sel_sec_rec_is_for_blob(): Add const qualifiers. Replace some occurrences of mtr_memo_contains_page() with mtr_memo_contains().
2006-10-23 21:14:36 +02:00
/* out: checksum */
const byte* page) /* in: buffer page */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
ulint checksum;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* Since the field FIL_PAGE_FILE_FLUSH_LSN, and in versions <= 4.1.x
..._ARCH_LOG_NO, are written outside the buffer pool to the first
pages of data files, we have to skip them in the page checksum
calculation.
Import 5.0 code.
2005-10-27 09:29:40 +02:00
We must also skip the field FIL_PAGE_SPACE_OR_CHKSUM where the
checksum is stored, and also the last 8 bytes of page because
there we store the old formula checksum. */
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
checksum = ut_fold_binary(page + FIL_PAGE_OFFSET,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
FIL_PAGE_FILE_FLUSH_LSN - FIL_PAGE_OFFSET)
+ ut_fold_binary(page + FIL_PAGE_DATA,
UNIV_PAGE_SIZE - FIL_PAGE_DATA
- FIL_PAGE_END_LSN_OLD_CHKSUM);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
checksum = checksum & 0xFFFFFFFFUL;
return(checksum);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/************************************************************************
In versions < 4.0.14 and < 4.1.1 there was a bug that the checksum only
looked at the first few bytes of the page. This calculates that old
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
checksum.
Import 5.0 code.
2005-10-27 09:29:40 +02:00
NOTE: we must first store the new formula checksum to
FIL_PAGE_SPACE_OR_CHKSUM before calculating and storing this old checksum
because this takes that field as an input! */
ulint
buf_calc_page_old_checksum(
/*=======================*/
branches/zip: Minor cleanup. buf_calc_page_new_checksum(), buf_calc_page_old_checksum(), buf_page_is_corrupted(), buf_page_print(): Add const qualifier to byte* parameter. page_parse_create_zip(): Remove this declaration of non-existing function. page_rec_check(), page_rec_print(), row_vers_impl_x_locked_off_kernel(): Add const qualifier to rec_t* parameter. row_vers_old_has_index_entry(): Add const qualifier to dtuple_t* parameter. row_sel_sec_rec_is_for_blob(): Add const qualifiers. Replace some occurrences of mtr_memo_contains_page() with mtr_memo_contains().
2006-10-23 21:14:36 +02:00
/* out: checksum */
const byte* page) /* in: buffer page */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
ulint checksum;
checksum = ut_fold_binary(page, FIL_PAGE_FILE_FLUSH_LSN);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
checksum = checksum & 0xFFFFFFFFUL;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
return(checksum);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/************************************************************************
Checks if a page is corrupt. */
ibool
buf_page_is_corrupted(
/*==================*/
branches/zip: Minor cleanup. buf_calc_page_new_checksum(), buf_calc_page_old_checksum(), buf_page_is_corrupted(), buf_page_print(): Add const qualifier to byte* parameter. page_parse_create_zip(): Remove this declaration of non-existing function. page_rec_check(), page_rec_print(), row_vers_impl_x_locked_off_kernel(): Add const qualifier to rec_t* parameter. row_vers_old_has_index_entry(): Add const qualifier to dtuple_t* parameter. row_sel_sec_rec_is_for_blob(): Add const qualifiers. Replace some occurrences of mtr_memo_contains_page() with mtr_memo_contains().
2006-10-23 21:14:36 +02:00
/* out: TRUE if corrupted */
const byte* read_buf, /* in: a database page */
ulint zip_size) /* in: size of compressed page;
0 for uncompressed pages */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
ulint checksum_field;
ulint old_checksum_field;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifndef UNIV_HOTBACKUP
branches/zip: Replace ib_ulonglong with ib_uint64_t, since we want exactly 64 bits. "long long" might be longer than 64 bits on some systems.
2006-11-29 15:52:16 +01:00
ib_uint64_t current_lsn;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
if (UNIV_LIKELY(!zip_size)
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
&& memcmp(read_buf + FIL_PAGE_LSN + 4,
read_buf + UNIV_PAGE_SIZE
- FIL_PAGE_END_LSN_OLD_CHKSUM + 4, 4)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Stored log sequence numbers at the start and the end
of page do not match */
return(TRUE);
}
#ifndef UNIV_HOTBACKUP
if (recv_lsn_checks_on && log_peek_lsn(&current_lsn)) {
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
if (current_lsn < mach_read_ull(read_buf + FIL_PAGE_LSN)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Error: page %lu log sequence number"
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
" %llu\n"
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: is in the future! Current system "
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
"log sequence number %llu.\n"
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: Your database may be corrupt or "
"you may have copied the InnoDB\n"
"InnoDB: tablespace but not the InnoDB "
"log files. See\n"
"InnoDB: http://dev.mysql.com/doc/refman/"
"5.1/en/forcing-recovery.html\n"
"InnoDB: for more information.\n",
(ulong) mach_read_from_4(read_buf
+ FIL_PAGE_OFFSET),
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
mach_read_ull(read_buf + FIL_PAGE_LSN),
current_lsn);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
#endif
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* If we use checksums validation, make additional check before
returning TRUE to ensure that the checksum is not equal to
BUF_NO_CHECKSUM_MAGIC which might be stored by InnoDB with checksums
disabled. Otherwise, skip checksum calculation and return FALSE */
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
if (UNIV_LIKELY(srv_use_checksums)) {
checksum_field = mach_read_from_4(read_buf
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ FIL_PAGE_SPACE_OR_CHKSUM);
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
if (UNIV_UNLIKELY(zip_size)) {
return(checksum_field != BUF_NO_CHECKSUM_MAGIC
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
&& checksum_field
!= page_zip_calc_checksum(read_buf, zip_size));
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
old_checksum_field = mach_read_from_4(
read_buf + UNIV_PAGE_SIZE
- FIL_PAGE_END_LSN_OLD_CHKSUM);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* There are 2 valid formulas for old_checksum_field:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
1. Very old versions of InnoDB only stored 8 byte lsn to the
start and the end of the page.
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
2. Newer InnoDB versions store the old formula checksum
there. */
if (old_checksum_field != mach_read_from_4(read_buf
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ FIL_PAGE_LSN)
&& old_checksum_field != BUF_NO_CHECKSUM_MAGIC
&& old_checksum_field
!= buf_calc_page_old_checksum(read_buf)) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
return(TRUE);
}
/* InnoDB versions < 4.0.14 and < 4.1.1 stored the space id
(always equal to 0), to FIL_PAGE_SPACE_SPACE_OR_CHKSUM */
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
if (checksum_field != 0
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
&& checksum_field != BUF_NO_CHECKSUM_MAGIC
&& checksum_field
!= buf_calc_page_new_checksum(read_buf)) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
return(TRUE);
}
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(FALSE);
}
/************************************************************************
Prints a page to stderr. */
void
buf_page_print(
/*===========*/
branches/zip: Minor cleanup. buf_calc_page_new_checksum(), buf_calc_page_old_checksum(), buf_page_is_corrupted(), buf_page_print(): Add const qualifier to byte* parameter. page_parse_create_zip(): Remove this declaration of non-existing function. page_rec_check(), page_rec_print(), row_vers_impl_x_locked_off_kernel(): Add const qualifier to rec_t* parameter. row_vers_old_has_index_entry(): Add const qualifier to dtuple_t* parameter. row_sel_sec_rec_is_for_blob(): Add const qualifiers. Replace some occurrences of mtr_memo_contains_page() with mtr_memo_contains().
2006-10-23 21:14:36 +02:00
const byte* read_buf, /* in: a database page */
ulint zip_size) /* in: compressed page size, or
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
0 for uncompressed pages */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
dict_index_t* index;
ulint checksum;
ulint old_checksum;
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
ulint size = zip_size;
if (!size) {
size = UNIV_PAGE_SIZE;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_print_timestamp(stderr);
fprintf(stderr, " InnoDB: Page dump in ascii and hex (%lu bytes):\n",
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
(ulong) size);
ut_print_buf(stderr, read_buf, size);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fputs("InnoDB: End of page dump\n", stderr);
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
if (zip_size) {
/* Print compressed page. */
switch (fil_page_get_type(read_buf)) {
case FIL_PAGE_TYPE_ZBLOB:
checksum = srv_use_checksums
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
? page_zip_calc_checksum(read_buf, zip_size)
: BUF_NO_CHECKSUM_MAGIC;
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Compressed BLOB page"
" checksum %lu, stored %lu\n"
"InnoDB: Page lsn %lu %lu\n"
"InnoDB: Page number (if stored"
" to page already) %lu,\n"
"InnoDB: space id (if stored"
" to page already) %lu\n",
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
(ulong) checksum,
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_SPACE_OR_CHKSUM),
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_LSN),
(ulong) mach_read_from_4(
read_buf + (FIL_PAGE_LSN + 4)),
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_OFFSET),
(ulong) mach_read_from_4(
branches/zip: Replace FIL_PAGE_ZBLOB_SPACE_ID and FIL_PAGE_ZBLOB_DATA with FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID and FIL_PAGE_DATA. The doublewrite buffer needs to read the space_id in order to determine the type of the page. Because FIL_PAGE_TYPE could contain garbage in MySQL/InnoDB 5.0 and earlier versions, we cannot trust fil_page_get_type(). Instead, we have to always store the space_id at the same location. This modification wastes 12 bytes per compressed BLOB page (1.2% on 1-kilobyte pages).
2006-09-27 12:51:05 +02:00
read_buf
+ FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID));
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
return;
default:
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: unknown page type %lu,"
" assuming FIL_PAGE_INDEX\n",
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
fil_page_get_type(read_buf));
/* fall through */
case FIL_PAGE_INDEX:
checksum = srv_use_checksums
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
? page_zip_calc_checksum(read_buf, zip_size)
: BUF_NO_CHECKSUM_MAGIC;
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Compressed page checksum %lu,"
" stored %lu\n"
"InnoDB: Page lsn %lu %lu\n"
"InnoDB: Page number (if stored"
" to page already) %lu,\n"
"InnoDB: space id (if stored"
" to page already) %lu\n",
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
(ulong) checksum,
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_SPACE_OR_CHKSUM),
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_LSN),
(ulong) mach_read_from_4(
read_buf + (FIL_PAGE_LSN + 4)),
(ulong) mach_read_from_4(
read_buf + FIL_PAGE_OFFSET),
(ulong) mach_read_from_4(
read_buf
+ FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID));
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
return;
case FIL_PAGE_TYPE_XDES:
/* This is an uncompressed page. */
break;
}
}
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
checksum = srv_use_checksums
? buf_calc_page_new_checksum(read_buf) : BUF_NO_CHECKSUM_MAGIC;
old_checksum = srv_use_checksums
? buf_calc_page_old_checksum(read_buf) : BUF_NO_CHECKSUM_MAGIC;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_print_timestamp(stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Page checksum %lu, prior-to-4.0.14-form"
" checksum %lu\n"
"InnoDB: stored checksum %lu, prior-to-4.0.14-form"
" stored checksum %lu\n"
"InnoDB: Page lsn %lu %lu, low 4 bytes of lsn"
" at page end %lu\n"
"InnoDB: Page number (if stored to page already) %lu,\n"
"InnoDB: space id (if created with >= MySQL-4.1.1"
" and stored already) %lu\n",
(ulong) checksum, (ulong) old_checksum,
(ulong) mach_read_from_4(read_buf + FIL_PAGE_SPACE_OR_CHKSUM),
(ulong) mach_read_from_4(read_buf + UNIV_PAGE_SIZE
- FIL_PAGE_END_LSN_OLD_CHKSUM),
Import 5.0 code.
2005-10-27 09:29:40 +02:00
(ulong) mach_read_from_4(read_buf + FIL_PAGE_LSN),
(ulong) mach_read_from_4(read_buf + FIL_PAGE_LSN + 4),
(ulong) mach_read_from_4(read_buf + UNIV_PAGE_SIZE
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
- FIL_PAGE_END_LSN_OLD_CHKSUM + 4),
Import 5.0 code.
2005-10-27 09:29:40 +02:00
(ulong) mach_read_from_4(read_buf + FIL_PAGE_OFFSET),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
(ulong) mach_read_from_4(read_buf
+ FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (mach_read_from_2(read_buf + TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_TYPE)
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
== TRX_UNDO_INSERT) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
Import 5.0 code.
2005-10-27 09:29:40 +02:00
"InnoDB: Page may be an insert undo log page\n");
} else if (mach_read_from_2(read_buf + TRX_UNDO_PAGE_HDR
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ TRX_UNDO_PAGE_TYPE)
== TRX_UNDO_UPDATE) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
Import 5.0 code.
2005-10-27 09:29:40 +02:00
"InnoDB: Page may be an update undo log page\n");
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
switch (fil_page_get_type(read_buf)) {
case FIL_PAGE_INDEX:
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: Page may be an index page where"
" index id is %lu %lu\n",
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
(ulong) ut_dulint_get_high(
btr_page_get_index_id(read_buf)),
(ulong) ut_dulint_get_low(
btr_page_get_index_id(read_buf)));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
#ifdef UNIV_HOTBACKUP
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* If the code is in ibbackup, dict_sys may be uninitialized,
i.e., NULL */
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
if (dict_sys == NULL) {
break;
}
#endif /* UNIV_HOTBACKUP */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
index = dict_index_find_on_id_low(
btr_page_get_index_id(read_buf));
branches/zip: Prepare for writing compressed and reading pages. buf_page_print(): Print also compressed pages. Add parameter zip_size. buf_flush_init_for_writing(): Stamp the fields on a compressed B-tree index page. Add the header field FIL_PAGE_ZBLOB_SPACE_ID as an alias of FIL_PAGE_PREV. page_zip_calc_checksum(): New function. page_zip_compress(): Avoid copying the fields that are written in buf_flush_init_for_writing(). page_zip_header_cmp(): New function for comparing those fields of the page header that will not be written in buf_flush_init_for_writing().
2006-05-02 13:44:39 +02:00
if (index) {
fputs("InnoDB: (", stderr);
dict_index_name_print(stderr, NULL, index);
fputs(")\n", stderr);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_INODE:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fputs("InnoDB: Page may be an 'inode' page\n", stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_IBUF_FREE_LIST:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fputs("InnoDB: Page may be an insert buffer free list page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_ALLOCATED:
fputs("InnoDB: Page may be a freshly allocated page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_IBUF_BITMAP:
fputs("InnoDB: Page may be an insert buffer bitmap page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_SYS:
fputs("InnoDB: Page may be a system page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_TRX_SYS:
fputs("InnoDB: Page may be a transaction system page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_FSP_HDR:
fputs("InnoDB: Page may be a file space header page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_XDES:
fputs("InnoDB: Page may be an extent descriptor page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
case FIL_PAGE_TYPE_BLOB:
fputs("InnoDB: Page may be a BLOB page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
break;
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
case FIL_PAGE_TYPE_ZBLOB:
fputs("InnoDB: Page may be a compressed BLOB page\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
stderr);
branches/zip: Initial steps towards disk-based storage of compressed pages. dict_mem_table_create(): Account for DICT_TF_COMPRESSED in a debug assertion. btr_store_big_rec_extern_fields(), btr_free_externally_stored_field(), btr_copy_externally_stored_field(): Implement the disk format for compressed BLOB pages. btr_copy_externally_stored_field(): Improve error reporting and handling when decompressing BLOB pages. buf_flush_init_for_writing(), buf_page_is_corrupted(), buf_page_print(): Account for compressed BLOB pages (FIL_PAGE_TYPE_ZBLOB). buf_calc_zblob_page_checksum(): New function.
2006-04-05 15:41:12 +02:00
break;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
/************************************************************************
Initializes a buffer control block when the buf_pool is created. */
static
void
buf_block_init(
/*===========*/
buf_block_t* block, /* in: pointer to control block */
branches/zip: Remove support for AWE (Intel PAE on Win32). Removal of the variable innobase_buffer_pool_awe_mem_mb requires coordination with MySQL AB.
2006-10-30 14:48:08 +01:00
byte* frame) /* in: pointer to buffer frame */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Improve Valgrind instrumentation of allocated memory. UNIV_MEM_FREE(): Declare a memory area free. UNIV_MEM_ALLOC(): Declare a memory area allocated (but uninitialized). UNIV_MEM_DESC(): Associate a memory area with a control block. UNIV_MEM_UNDESC(): Unassociate a control block. trx_sys_create_doublewrite_buf(): Clear the buffer with memset(). buf_page_init(): Add a bogus UNIV_MEM_VALID(block->frame) to silence valid warnings about InnoDB data pages containing uninitialized data. buf_LRU_get_free_only(): Add UNIV_MEM_ALLOC(block->frame). buf_LRU_get_free_block(): Add UNIV_MEM_DESC(block->page.zip.data, block). buf_LRU_free_block(): Add UNIV_MEM_DESC(b->zip.data, b) when allocating a compressed-only control block for a compressed page. buf_LRU_block_free_non_file_page(): Replace UNIV_MEM_INVALID() with UNIV_MEM_FREE(). buf_LRU_block_remove_hashed_page(): Add UNIV_MEM_UNDESC(bpage) when deallocating a compressed-only control block. Add UNIV_MEM_INVALID(block->frame). (The frame should be flagged free by buf_LRU_block_free_non_file_page() moments later.) buf0buddy.c: Disable some extra checks in Valgrind-instrumented builds. Add UNIV_MEM_VALID(), UNIV_MEM_INVALID(), UNIV_MEM_FREE(), UNIV_MEM_ALLOC() as necessary.
2007-01-12 13:36:40 +01:00
UNIV_MEM_DESC(frame, UNIV_PAGE_SIZE, block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->frame = frame;
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.state = BUF_BLOCK_NOT_USED;
block->page.buf_fix_count = 0;
block->page.io_fix = BUF_IO_NONE;
branches/zip: Merge revisions 79:200 from trunk.
2006-02-17 15:19:39 +01:00
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
block->modify_clock = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block->page.file_page_was_freed = FALSE;
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#endif /* UNIV_DEBUG_FILE_ACCESSES */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->check_index_page_at_flush = FALSE;
block->index = NULL;
branches/zip: ha_create(): Remove parameter in_btr_search, which was passed as TRUE. Enclose hash_table_t::adaptive and buf_block_t::n_pointers in #ifdef UNIV_DEBUG. btr_search_drop_page_hash_index(): Enclose the corruption check (which depends on buf_block_t::n_pointers) in #ifdef UNIV_DEBUG.
2006-10-18 20:52:04 +02:00
#ifdef UNIV_DEBUG
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
block->page.in_page_hash = FALSE;
block->page.in_zip_hash = FALSE;
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
block->page.in_flush_list = FALSE;
branches/zip: Remove buf_block_t:free_list. Rename buf_page_t:flush_list to free_or_flush_list. The list node pointers can be shared, because a block can never be belong to both lists at the same time.
2006-12-04 13:44:06 +01:00
block->page.in_free_list = FALSE;
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.in_LRU_list = FALSE;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->n_pointers = 0;
branches/zip: ha_create(): Remove parameter in_btr_search, which was passed as TRUE. Enclose hash_table_t::adaptive and buf_block_t::n_pointers in #ifdef UNIV_DEBUG. btr_search_drop_page_hash_index(): Enclose the corruption check (which depends on buf_block_t::n_pointers) in #ifdef UNIV_DEBUG.
2006-10-18 20:52:04 +02:00
#endif /* UNIV_DEBUG */
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
page_zip_des_init(&block->page.zip);
branches/zip: Initial import of the zip-like compression of B-tree pages.
2005-10-27 13:48:10 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_create(&block->mutex, SYNC_BUF_BLOCK);
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
rw_lock_create(&block->lock, SYNC_LEVEL_VARYING);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(rw_lock_validate(&(block->lock)));
#ifdef UNIV_SYNC_DEBUG
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
rw_lock_create(&block->debug_latch, SYNC_NO_ORDER_CHECK);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif /* UNIV_SYNC_DEBUG */
}
/************************************************************************
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
Allocates a chunk of buffer frames. */
static
buf_chunk_t*
buf_chunk_init(
/*===========*/
/* out: chunk, or NULL on failure */
buf_chunk_t* chunk, /* out: chunk of buffers */
ulint mem_size) /* in: requested size in bytes */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_block_t* block;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
byte* frame;
ulint i;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Round down to a multiple of page size,
although it already should be. */
mem_size = ut_2pow_round(mem_size, UNIV_PAGE_SIZE);
/* Reserve space for the block descriptors. */
mem_size += ut_2pow_round((mem_size / UNIV_PAGE_SIZE) * (sizeof *block)
+ (UNIV_PAGE_SIZE - 1), UNIV_PAGE_SIZE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk->mem_size = mem_size;
chunk->mem = os_mem_alloc_large(&chunk->mem_size);
branches/zip: os_mem_alloc_large(), os_mem_free_large(): On POSIX, use mmap() and munmap(). On Windows, use VirtualAlloc() and VirtualFree(). Only on Netware, use ut_malloc_low() and ut_free(). The lower-level functions on POSIX and Windows allow InnoDB to return memory to the operating system when the buffer pool is shrunk.
2006-11-01 09:13:58 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
if (UNIV_UNLIKELY(chunk->mem == NULL)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(NULL);
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Allocate the block descriptors from
the start of the memory block. */
chunk->blocks = chunk->mem;
branches/zip: os_mem_alloc_large(), os_mem_free_large(): On POSIX, use mmap() and munmap(). On Windows, use VirtualAlloc() and VirtualFree(). Only on Netware, use ut_malloc_low() and ut_free(). The lower-level functions on POSIX and Windows allow InnoDB to return memory to the operating system when the buffer pool is shrunk.
2006-11-01 09:13:58 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Align pointer to the first frame */
branches/zip: os_mem_alloc_large(), os_mem_free_large(): On POSIX, use mmap() and munmap(). On Windows, use VirtualAlloc() and VirtualFree(). Only on Netware, use ut_malloc_low() and ut_free(). The lower-level functions on POSIX and Windows allow InnoDB to return memory to the operating system when the buffer pool is shrunk.
2006-11-01 09:13:58 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
frame = ut_align(chunk->mem, UNIV_PAGE_SIZE);
chunk->size = chunk->mem_size / UNIV_PAGE_SIZE
- (frame != chunk->mem);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Subtract the space needed for block descriptors. */
{
ulint size = chunk->size;
branches/zip: os_mem_alloc_large(), os_mem_free_large(): On POSIX, use mmap() and munmap(). On Windows, use VirtualAlloc() and VirtualFree(). Only on Netware, use ut_malloc_low() and ut_free(). The lower-level functions on POSIX and Windows allow InnoDB to return memory to the operating system when the buffer pool is shrunk.
2006-11-01 09:13:58 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
while (frame < (byte*) (chunk->blocks + size)) {
frame += UNIV_PAGE_SIZE;
size--;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk->size = size;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Remove support for AWE (Intel PAE on Win32). Removal of the variable innobase_buffer_pool_awe_mem_mb requires coordination with MySQL AB.
2006-10-30 14:48:08 +01:00
/* Init block structs and assign frames for them. Then we
assign the frames to the first blocks (we already mapped the
memory above). */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
block = chunk->blocks;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (i = chunk->size; i--; ) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_block_init(block, frame);
branches/zip: Remove buf_pool->frame_zero and buf_pool->high_end.
2006-10-31 10:59:15 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
#ifdef HAVE_purify
/* Wipe contents of frame to eliminate a Purify warning */
memset(block->frame, '\0', UNIV_PAGE_SIZE);
#endif
/* Add the block to the free list */
branches/zip: buf_page_t: Rename free_or_flush_list to list and document the possible list memberships.
2006-12-05 13:31:38 +01:00
UT_LIST_ADD_LAST(list, buf_pool->free, (&block->page));
branches/zip: Remove buf_block_t:free_list. Rename buf_page_t:flush_list to free_or_flush_list. The list node pointers can be shared, because a block can never be belong to both lists at the same time.
2006-12-04 13:44:06 +01:00
ut_d(block->page.in_free_list = TRUE);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
block++;
branches/zip: Remove buf_pool->frame_zero and buf_pool->high_end.
2006-10-31 10:59:15 +01:00
frame += UNIV_PAGE_SIZE;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
return(chunk);
}
branches/zip: Add the debug function buf_pool_contains_zip() #ifdef UNIV_DEBUG.
2006-12-19 23:12:44 +01:00
#ifdef UNIV_DEBUG
/*************************************************************************
Finds a block in the given buffer chunk that points to a
given compressed page. */
static
buf_block_t*
buf_chunk_contains_zip(
/*===================*/
/* out: buffer block pointing to
the compressed page, or NULL */
buf_chunk_t* chunk, /* in: chunk being checked */
const void* data) /* in: pointer to compressed page */
{
buf_block_t* block;
ulint i;
ut_ad(buf_pool);
branches/zip: Make mutex_own() work with UNIV_DEBUG, without UNIV_SYNC_DEBUG.
2007-01-18 19:29:12 +01:00
ut_ad(mutex_own(&buf_pool->mutex));
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
branches/zip: Add the debug function buf_pool_contains_zip() #ifdef UNIV_DEBUG.
2006-12-19 23:12:44 +01:00
block = chunk->blocks;
for (i = chunk->size; i--; block++) {
if (block->page.zip.data == data) {
return(block);
}
}
return(NULL);
}
/*************************************************************************
Finds a block in the buffer pool that points to a
given compressed page. */
buf_block_t*
buf_pool_contains_zip(
/*==================*/
/* out: buffer block pointing to
the compressed page, or NULL */
const void* data) /* in: pointer to compressed page */
{
ulint n;
buf_chunk_t* chunk = buf_pool->chunks;
for (n = buf_pool->n_chunks; n--; chunk++) {
buf_block_t* block = buf_chunk_contains_zip(chunk, data);
if (block) {
return(block);
}
}
return(NULL);
}
#endif /* UNIV_DEBUG */
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/*************************************************************************
Checks that all file pages in the buffer chunk are in a replaceable state. */
static
const buf_block_t*
buf_chunk_not_freed(
/*================*/
/* out: address of a non-free block,
or NULL if all freed */
buf_chunk_t* chunk) /* in: chunk being checked */
{
buf_block_t* block;
ulint i;
ut_ad(buf_pool);
ut_ad(mutex_own(&(buf_pool->mutex)));
block = chunk->blocks;
for (i = chunk->size; i--; block++) {
mutex_enter(&block->mutex);
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
if (buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
&& !buf_flush_ready_for_replace(&block->page)) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
mutex_exit(&block->mutex);
return(block);
}
mutex_exit(&block->mutex);
}
return(NULL);
}
/*************************************************************************
Checks that all blocks in the buffer chunk are in BUF_BLOCK_NOT_USED state. */
static
ibool
buf_chunk_all_free(
/*===============*/
/* out: TRUE if all freed */
const buf_chunk_t* chunk) /* in: chunk being checked */
{
const buf_block_t* block;
ulint i;
ut_ad(buf_pool);
ut_ad(mutex_own(&(buf_pool->mutex)));
block = chunk->blocks;
for (i = chunk->size; i--; block++) {
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
if (buf_block_get_state(block) != BUF_BLOCK_NOT_USED) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
return(FALSE);
}
}
return(TRUE);
}
/************************************************************************
Frees a chunk of buffer frames. */
static
void
buf_chunk_free(
/*===========*/
buf_chunk_t* chunk) /* out: chunk of buffers */
{
buf_block_t* block;
const buf_block_t* block_end;
ut_ad(mutex_own(&(buf_pool->mutex)));
block_end = chunk->blocks + chunk->size;
for (block = chunk->blocks; block < block_end; block++) {
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_NOT_USED);
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
ut_a(!block->page.zip.data);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
ut_ad(!block->page.in_LRU_list);
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
ut_ad(!block->page.in_flush_list);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Remove the block from the free list. */
branches/zip: Remove buf_block_t:free_list. Rename buf_page_t:flush_list to free_or_flush_list. The list node pointers can be shared, because a block can never be belong to both lists at the same time.
2006-12-04 13:44:06 +01:00
ut_ad(block->page.in_free_list);
branches/zip: buf_page_t: Rename free_or_flush_list to list and document the possible list memberships.
2006-12-05 13:31:38 +01:00
UT_LIST_REMOVE(list, buf_pool->free, (&block->page));
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Free the latches. */
mutex_free(&block->mutex);
rw_lock_free(&block->lock);
#ifdef UNIV_SYNC_DEBUG
rw_lock_free(&block->debug_latch);
#endif /* UNIV_SYNC_DEBUG */
branches/zip: Improve Valgrind instrumentation of allocated memory. UNIV_MEM_FREE(): Declare a memory area free. UNIV_MEM_ALLOC(): Declare a memory area allocated (but uninitialized). UNIV_MEM_DESC(): Associate a memory area with a control block. UNIV_MEM_UNDESC(): Unassociate a control block. trx_sys_create_doublewrite_buf(): Clear the buffer with memset(). buf_page_init(): Add a bogus UNIV_MEM_VALID(block->frame) to silence valid warnings about InnoDB data pages containing uninitialized data. buf_LRU_get_free_only(): Add UNIV_MEM_ALLOC(block->frame). buf_LRU_get_free_block(): Add UNIV_MEM_DESC(block->page.zip.data, block). buf_LRU_free_block(): Add UNIV_MEM_DESC(b->zip.data, b) when allocating a compressed-only control block for a compressed page. buf_LRU_block_free_non_file_page(): Replace UNIV_MEM_INVALID() with UNIV_MEM_FREE(). buf_LRU_block_remove_hashed_page(): Add UNIV_MEM_UNDESC(bpage) when deallocating a compressed-only control block. Add UNIV_MEM_INVALID(block->frame). (The frame should be flagged free by buf_LRU_block_free_non_file_page() moments later.) buf0buddy.c: Disable some extra checks in Valgrind-instrumented builds. Add UNIV_MEM_VALID(), UNIV_MEM_INVALID(), UNIV_MEM_FREE(), UNIV_MEM_ALLOC() as necessary.
2007-01-12 13:36:40 +01:00
UNIV_MEM_UNDESC(block);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
}
os_mem_free_large(chunk->mem, chunk->mem_size);
}
/************************************************************************
Creates the buffer pool. */
buf_pool_t*
buf_pool_init(void)
/*===============*/
/* out, own: buf_pool object, NULL if not
enough memory or error */
{
buf_chunk_t* chunk;
ulint i;
buf_pool = mem_alloc(sizeof(buf_pool_t));
/* 1. Initialize general fields
------------------------------- */
mutex_create(&buf_pool->mutex, SYNC_BUF_POOL);
mutex_enter(&(buf_pool->mutex));
branches/zip: Add the first bits of the binary buddy system for allocating compressed pages from the buffer pool. Makefile.am: Add buf0buddy.h, buf0buddy.ic. buf/Makefile.am: Add buf0buddy.c. Introduce the constants BUF_BUDDY_LOW and BUF_BUDDY_SIZES. buf_pool_t: Add zip_mutex and the lists zip_clean and zip_free[]. buf_page_get_mutex(): Return &buf_pool->zip_mutex instead of NULL. buf_buddy_get_offset(), buf_buddy_get(), buf_buddy_get_slot(), buf_buddy_alloc_free(), buf_buddy_alloc_free_low(): New functions.
2006-12-07 14:29:04 +01:00
mutex_create(&buf_pool->zip_mutex, SYNC_BUF_BLOCK);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_pool->n_chunks = 1;
buf_pool->chunks = chunk = mem_alloc(sizeof *chunk);
UT_LIST_INIT(buf_pool->free);
if (!buf_chunk_init(chunk, srv_buf_pool_size)) {
mem_free(chunk);
mem_free(buf_pool);
buf_pool = NULL;
return(NULL);
}
srv_buf_pool_old_size = srv_buf_pool_size;
buf_pool->curr_size = chunk->size;
srv_buf_pool_curr_size = buf_pool->curr_size * UNIV_PAGE_SIZE;
buf_pool->page_hash = hash_create(2 * buf_pool->curr_size);
branches/zip: Add buf_pool->zip_hash for keeping track on pages allocated to the buddy system for allocating compressed pages and their descriptors. buf_buddy_free_block(): New function: Deallocate the buffer frame. buf_buddy_free(), buf_buddy_free_low(): Return void instead of a pointer to a freed buffer frame.
2006-12-11 10:54:13 +01:00
buf_pool->zip_hash = hash_create(2 * buf_pool->curr_size);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->n_pend_reads = 0;
buf_pool->last_printout_time = time(NULL);
buf_pool->n_pages_read = 0;
buf_pool->n_pages_written = 0;
buf_pool->n_pages_created = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->n_page_gets = 0;
buf_pool->n_page_gets_old = 0;
buf_pool->n_pages_read_old = 0;
buf_pool->n_pages_written_old = 0;
buf_pool->n_pages_created_old = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* 2. Initialize flushing fields
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
-------------------------------- */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
UT_LIST_INIT(buf_pool->flush_list);
branches/zip: Move the page flushing fields from buf_block_t to buf_page_t and page_zip_des_t.
2006-11-27 15:12:09 +01:00
for (i = BUF_FLUSH_LRU; i < BUF_FLUSH_N_TYPES; i++) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->n_flush[i] = 0;
buf_pool->init_flush[i] = FALSE;
buf_pool->no_flush[i] = os_event_create(NULL);
}
buf_pool->LRU_flush_ended = 0;
buf_pool->ulint_clock = 1;
buf_pool->freed_page_clock = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* 3. Initialize LRU fields
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
--------------------------- */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
UT_LIST_INIT(buf_pool->LRU);
buf_pool->LRU_old = NULL;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
mutex_exit(&(buf_pool->mutex));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
btr_search_sys_create(buf_pool->curr_size
* UNIV_PAGE_SIZE / sizeof(void*) / 64);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Add the first bits of the binary buddy system for allocating compressed pages from the buffer pool. Makefile.am: Add buf0buddy.h, buf0buddy.ic. buf/Makefile.am: Add buf0buddy.c. Introduce the constants BUF_BUDDY_LOW and BUF_BUDDY_SIZES. buf_pool_t: Add zip_mutex and the lists zip_clean and zip_free[]. buf_page_get_mutex(): Return &buf_pool->zip_mutex instead of NULL. buf_buddy_get_offset(), buf_buddy_get(), buf_buddy_get_slot(), buf_buddy_alloc_free(), buf_buddy_alloc_free_low(): New functions.
2006-12-07 14:29:04 +01:00
/* 4. Initialize the buddy allocator fields */
UT_LIST_INIT(buf_pool->zip_clean);
for (i = 0; i < BUF_BUDDY_SIZES; i++) {
UT_LIST_INIT(buf_pool->zip_free[i]);
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
return(buf_pool);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
/************************************************************************
Relocate a buffer control block. Relocates the block on the LRU list
and in buf_pool->page_hash. Does not relocate bpage->list. */
void
buf_relocate(
/*=========*/
buf_page_t* bpage, /* control block being relocated */
buf_page_t* dpage) /* destination control block */
{
buf_page_t* b;
ulint fold;
branches/zip: Make mutex_own() work with UNIV_DEBUG, without UNIV_SYNC_DEBUG.
2007-01-18 19:29:12 +01:00
ut_ad(mutex_own(&buf_pool->mutex));
ut_ad(mutex_own(buf_page_get_mutex(bpage)));
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
ut_a(buf_page_get_io_fix(bpage) == BUF_IO_NONE);
ut_a(bpage->buf_fix_count == 0);
ut_a(buf_page_in_file(bpage));
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_ad(bpage->in_LRU_list);
ut_ad(!bpage->in_zip_hash);
ut_ad(bpage->in_page_hash);
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
ut_ad(bpage == buf_page_hash_get(bpage->space, bpage->offset));
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
memcpy(dpage, bpage, sizeof *dpage);
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_d(bpage->in_LRU_list = FALSE);
ut_d(bpage->in_page_hash = FALSE);
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
/* relocate buf_pool->LRU */
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
b = UT_LIST_GET_PREV(LRU, bpage);
UT_LIST_REMOVE(LRU, buf_pool->LRU, bpage);
if (b) {
UT_LIST_INSERT_AFTER(LRU, buf_pool->LRU, b, dpage);
} else {
UT_LIST_ADD_FIRST(LRU, buf_pool->LRU, dpage);
}
branches/zip: Fix two bugs in the relocation of block descriptors. Improve diagnostics. buf_buddy_relocate(): Do not assume all small blocks to contain only one buf_page_t object. Require an exact size match, as in the case of compressed pages. buf_relocate(): Relocate buf_pool->LRU_old if needed. Validate the LRU list in debug builds. buf_LRU_block_remove_hashed_page(): Do not overwrite FIL_PAGE_OFFSET or FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID. The memory will be invalidated in debug builds of buf_buddy_free(). buf0buddy.c: Add assertions for not holding buf_pool->zip_mutex. The calling thread should not be holding any block->mutex, but it might be too time-consuming to assert against that, even #ifdef UNIV_SYNC_DEBUG.
2007-01-04 16:45:42 +01:00
if (UNIV_UNLIKELY(buf_pool->LRU_old == bpage)) {
buf_pool->LRU_old = dpage;
}
ut_d(UT_LIST_VALIDATE(LRU, buf_page_t, buf_pool->LRU));
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
/* relocate buf_pool->page_hash */
fold = buf_page_address_fold(bpage->space, bpage->offset);
HASH_DELETE(buf_page_t, hash, buf_pool->page_hash, fold, bpage);
HASH_INSERT(buf_page_t, hash, buf_pool->page_hash, fold, dpage);
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
UNIV_MEM_INVALID(bpage, sizeof *bpage);
branches/zip: Refactor some code for relocating buffer control blocks. buf_relocate(): New function, split from buf_buddy_relocate().
2006-12-14 16:15:51 +01:00
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/************************************************************************
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
Shrinks the buffer pool. */
static
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
void
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
buf_pool_shrink(
/*============*/
/* out: TRUE if shrunk */
ulint chunk_size) /* in: number of pages to remove */
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
{
buf_chunk_t* chunks;
buf_chunk_t* chunk;
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
ulint max_size;
ulint max_free_size;
buf_chunk_t* max_chunk;
buf_chunk_t* max_free_chunk;
ut_ad(!mutex_own(&buf_pool->mutex));
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
try_again:
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
btr_search_disable(); /* Empty the adaptive hash index again */
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
mutex_enter(&buf_pool->mutex);
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
shrink_again:
if (buf_pool->n_chunks <= 1) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Cannot shrink if there is only one chunk */
goto func_done;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Search for the largest free chunk
not larger than the size difference */
chunks = buf_pool->chunks;
chunk = chunks + buf_pool->n_chunks;
max_size = max_free_size = 0;
max_chunk = max_free_chunk = NULL;
while (--chunk >= chunks) {
if (chunk->size <= chunk_size
&& chunk->size > max_free_size) {
if (chunk->size > max_size) {
max_size = chunk->size;
max_chunk = chunk;
}
if (buf_chunk_all_free(chunk)) {
max_free_size = chunk->size;
max_free_chunk = chunk;
}
}
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
if (!max_free_size) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
ulint dirty = 0;
ulint nonfree = 0;
buf_block_t* block;
buf_block_t* bend;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Cannot shrink: try again later
(do not assign srv_buf_pool_old_size) */
if (!max_chunk) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
goto func_exit;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
block = max_chunk->blocks;
bend = block + max_chunk->size;
/* Move the blocks of chunk to the end of the
LRU list and try to flush them. */
for (; block < bend; block++) {
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
switch (buf_block_get_state(block)) {
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
case BUF_BLOCK_NOT_USED:
continue;
case BUF_BLOCK_FILE_PAGE:
break;
default:
nonfree++;
continue;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
mutex_enter(&block->mutex);
/* The following calls will temporarily
release block->mutex and buf_pool->mutex.
Therefore, we have to always retry,
even if !dirty && !nonfree. */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
if (!buf_flush_ready_for_replace(&block->page)) {
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_LRU_make_block_old(&block->page);
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
dirty++;
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
} else if (!buf_LRU_free_block(&block->page, TRUE)) {
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
nonfree++;
}
mutex_exit(&block->mutex);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
mutex_exit(&buf_pool->mutex);
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Request for a flush of the chunk if it helps.
Do not flush if there are non-free blocks, since
flushing will not make the chunk freeable. */
if (nonfree) {
/* Avoid busy-waiting. */
os_thread_sleep(100000);
} else if (dirty
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
&& buf_flush_batch(BUF_FLUSH_LRU, dirty, 0)
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
== ULINT_UNDEFINED) {
buf_flush_wait_batch_end(BUF_FLUSH_LRU);
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
goto try_again;
}
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
max_size = max_free_size;
max_chunk = max_free_chunk;
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
srv_buf_pool_old_size = srv_buf_pool_size;
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Rewrite buf_pool->chunks. Copy everything but max_chunk. */
chunks = mem_alloc((buf_pool->n_chunks - 1) * sizeof *chunks);
memcpy(chunks, buf_pool->chunks,
(max_chunk - buf_pool->chunks) * sizeof *chunks);
memcpy(chunks + (max_chunk - buf_pool->chunks),
max_chunk + 1,
buf_pool->chunks + buf_pool->n_chunks
- (max_chunk + 1));
ut_a(buf_pool->curr_size > max_chunk->size);
buf_pool->curr_size -= max_chunk->size;
srv_buf_pool_curr_size = buf_pool->curr_size * UNIV_PAGE_SIZE;
chunk_size -= max_chunk->size;
buf_chunk_free(max_chunk);
mem_free(buf_pool->chunks);
buf_pool->chunks = chunks;
buf_pool->n_chunks--;
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/* Allow a slack of one megabyte. */
if (chunk_size > 1048576 / UNIV_PAGE_SIZE) {
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
goto shrink_again;
}
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
func_done:
srv_buf_pool_old_size = srv_buf_pool_size;
func_exit:
mutex_exit(&buf_pool->mutex);
btr_search_enable();
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
/************************************************************************
Rebuild buf_pool->page_hash. */
static
void
buf_pool_page_hash_rebuild(void)
/*============================*/
{
ulint i;
ulint n_chunks;
buf_chunk_t* chunk;
hash_table_t* page_hash;
branches/zip: buf_pool_page_hash_rebuild(): Rebuild also buf_pool->zip_hash and add compressed-only pages to buf_pool->page_hash. BUF_POOL_ZIP_FOLD(), BUF_POOL_ZIP_FOLD_PTR(), BUF_POOL_ZIP_FOLD_BPAGE(): Macros for computing buf_pool->zip_hash fold values. Make use of these in buf0buddy.c. buf0buddy.c: Replace some ptr == ut_align_down(ptr, ...) in assertions with !ut_align_offset(ptr, ...).
2007-01-05 11:49:56 +01:00
hash_table_t* zip_hash;
buf_page_t* b;
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
mutex_enter(&buf_pool->mutex);
/* Free, create, and populate the hash table. */
hash_table_free(buf_pool->page_hash);
buf_pool->page_hash = page_hash = hash_create(2 * buf_pool->curr_size);
branches/zip: buf_pool_page_hash_rebuild(): Rebuild also buf_pool->zip_hash and add compressed-only pages to buf_pool->page_hash. BUF_POOL_ZIP_FOLD(), BUF_POOL_ZIP_FOLD_PTR(), BUF_POOL_ZIP_FOLD_BPAGE(): Macros for computing buf_pool->zip_hash fold values. Make use of these in buf0buddy.c. buf0buddy.c: Replace some ptr == ut_align_down(ptr, ...) in assertions with !ut_align_offset(ptr, ...).
2007-01-05 11:49:56 +01:00
zip_hash = hash_create(2 * buf_pool->curr_size);
HASH_MIGRATE(buf_pool->zip_hash, zip_hash, buf_page_t, hash,
BUF_POOL_ZIP_FOLD_BPAGE);
hash_table_free(buf_pool->zip_hash);
buf_pool->zip_hash = zip_hash;
/* Insert the uncompressed file pages to buf_pool->page_hash. */
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
chunk = buf_pool->chunks;
n_chunks = buf_pool->n_chunks;
for (i = 0; i < n_chunks; i++, chunk++) {
ulint j;
buf_block_t* block = chunk->blocks;
for (j = 0; j < chunk->size; j++, block++) {
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
if (buf_block_get_state(block)
== BUF_BLOCK_FILE_PAGE) {
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_ad(!block->page.in_zip_hash);
ut_ad(block->page.in_page_hash);
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
HASH_INSERT(buf_page_t, hash, page_hash,
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
buf_page_address_fold(
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
block->page.space,
block->page.offset),
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
&block->page);
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
}
}
}
branches/zip: buf_pool_page_hash_rebuild(): Rebuild also buf_pool->zip_hash and add compressed-only pages to buf_pool->page_hash. BUF_POOL_ZIP_FOLD(), BUF_POOL_ZIP_FOLD_PTR(), BUF_POOL_ZIP_FOLD_BPAGE(): Macros for computing buf_pool->zip_hash fold values. Make use of these in buf0buddy.c. buf0buddy.c: Replace some ptr == ut_align_down(ptr, ...) in assertions with !ut_align_offset(ptr, ...).
2007-01-05 11:49:56 +01:00
/* Insert the compressed-only pages to buf_pool->page_hash.
All such blocks are either in buf_pool->zip_clean or
in buf_pool->flush_list. */
for (b = UT_LIST_GET_FIRST(buf_pool->zip_clean); b;
b = UT_LIST_GET_NEXT(list, b)) {
ut_a(buf_page_get_state(b) == BUF_BLOCK_ZIP_PAGE);
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
ut_ad(!b->in_flush_list);
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_ad(b->in_LRU_list);
ut_ad(b->in_page_hash);
ut_ad(!b->in_zip_hash);
branches/zip: buf_pool_page_hash_rebuild(): Rebuild also buf_pool->zip_hash and add compressed-only pages to buf_pool->page_hash. BUF_POOL_ZIP_FOLD(), BUF_POOL_ZIP_FOLD_PTR(), BUF_POOL_ZIP_FOLD_BPAGE(): Macros for computing buf_pool->zip_hash fold values. Make use of these in buf0buddy.c. buf0buddy.c: Replace some ptr == ut_align_down(ptr, ...) in assertions with !ut_align_offset(ptr, ...).
2007-01-05 11:49:56 +01:00
HASH_INSERT(buf_page_t, hash, page_hash,
buf_page_address_fold(b->space, b->offset), b);
}
for (b = UT_LIST_GET_FIRST(buf_pool->flush_list); b;
b = UT_LIST_GET_NEXT(list, b)) {
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
ut_ad(b->in_flush_list);
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_ad(b->in_LRU_list);
ut_ad(b->in_page_hash);
ut_ad(!b->in_zip_hash);
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
branches/zip: buf_pool_page_hash_rebuild(): Rebuild also buf_pool->zip_hash and add compressed-only pages to buf_pool->page_hash. BUF_POOL_ZIP_FOLD(), BUF_POOL_ZIP_FOLD_PTR(), BUF_POOL_ZIP_FOLD_BPAGE(): Macros for computing buf_pool->zip_hash fold values. Make use of these in buf0buddy.c. buf0buddy.c: Replace some ptr == ut_align_down(ptr, ...) in assertions with !ut_align_offset(ptr, ...).
2007-01-05 11:49:56 +01:00
switch (buf_page_get_state(b)) {
case BUF_BLOCK_ZIP_DIRTY:
HASH_INSERT(buf_page_t, hash, page_hash,
buf_page_address_fold(b->space,
b->offset), b);
break;
case BUF_BLOCK_FILE_PAGE:
/* uncompressed page */
break;
case BUF_BLOCK_ZIP_FREE:
case BUF_BLOCK_ZIP_PAGE:
case BUF_BLOCK_NOT_USED:
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
ut_error;
break;
}
}
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
mutex_exit(&buf_pool->mutex);
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
/************************************************************************
Resizes the buffer pool. */
branches/zip: Improve the shrinking of the buffer pool. buf_LRU_block_free_non_file_page(): Deallocate block->page_zip.data to avoid ut_a(!block->page_zip.data) in buf_chunk_free(). buf_chunk_free(): Add the assertion ut_a(!block->in_LRU_list). buf_pool_resize(): When shrinking the buffer pool and there are non-free blocks in the candidate chunk, free the clean blocks and move the dirty blocks to the end of the LRU list and request a flush. Proceed if the chunk becomes free, and retry otherwise.
2006-11-15 22:49:14 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
void
buf_pool_resize(void)
/*=================*/
{
mutex_enter(&buf_pool->mutex);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
if (srv_buf_pool_old_size == srv_buf_pool_size) {
mutex_exit(&buf_pool->mutex);
return;
}
if (srv_buf_pool_curr_size + 1048576 > srv_buf_pool_size) {
mutex_exit(&buf_pool->mutex);
/* Disable adaptive hash indexes and empty the index
in order to free up memory in the buffer pool chunks. */
buf_pool_shrink((srv_buf_pool_curr_size - srv_buf_pool_size)
/ UNIV_PAGE_SIZE);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
} else if (srv_buf_pool_curr_size + 1048576 < srv_buf_pool_size) {
/* Enlarge the buffer pool by at least one megabyte */
ulint mem_size
= srv_buf_pool_size - srv_buf_pool_curr_size;
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
buf_chunk_t* chunks;
buf_chunk_t* chunk;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunks = mem_alloc((buf_pool->n_chunks + 1) * sizeof *chunks);
memcpy(chunks, buf_pool->chunks, buf_pool->n_chunks
* sizeof *chunks);
chunk = &chunks[buf_pool->n_chunks];
if (!buf_chunk_init(chunk, mem_size)) {
mem_free(chunks);
} else {
buf_pool->curr_size += chunk->size;
srv_buf_pool_curr_size = buf_pool->curr_size
* UNIV_PAGE_SIZE;
mem_free(buf_pool->chunks);
buf_pool->chunks = chunks;
buf_pool->n_chunks++;
}
branches/zip: When shrinking the buffer pool, disable and purge the adaptive hash index, because it might occupy the chunk we would like to free. TODO: In btr_search_check_free_space_in_heap(), release the block if btr_search_latch is not immediately available. buf_pool_shrink(): Split from buf_pool_resize(). btr_search_disabled: New variable, similar to srv_use_adaptive_hash_indexes that was removed earlier. btr_search_disable(): New function: disable and purge the adaptive hash index. btr_search_enable(): New function: enable the adaptive hash index. ha_clear(): New function: Empty a hash table and free the memory heaps.
2006-11-21 15:40:14 +01:00
srv_buf_pool_old_size = srv_buf_pool_size;
mutex_exit(&buf_pool->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Rebuild buf_pool->page_hash after resizing the buffer pool. buf_pool_page_hash_rebuild(): New function.
2006-11-21 16:38:18 +01:00
buf_pool_page_hash_rebuild();
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Moves to 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(
/*=================*/
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_page_t* bpage) /* in: block to make younger */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
ut_ad(!mutex_own(&(buf_pool->mutex)));
/* Note that we read freed_page_clock's without holding any mutex:
this is allowed since the result is used only in heuristics */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
if (buf_pool->freed_page_clock
>= buf_page_get_freed_page_clock(bpage)
+ 1 + (buf_pool->curr_size / 4)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&buf_pool->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* There has been freeing activity in the LRU list:
best to move to the head of the LRU list */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_LRU_make_block_young(bpage);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&buf_pool->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
/************************************************************************
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
the buffer pool. */
void
buf_page_make_young(
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/*================*/
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_page_t* bpage) /* in: buffer block of a file page */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
mutex_enter(&(buf_pool->mutex));
branches/zip: Allow some functions to work on compressed-only pages (state == BUF_BLOCK_ZIP_PAGE). Make use of buf_page_in_file() and buf_page_get_mutex(). buf_block_get_newest_modification(): Rename to buf_page_get_newest_modification().
2006-11-30 16:09:33 +01:00
ut_a(buf_page_in_file(bpage));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_LRU_make_block_young(bpage);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
}
/************************************************************************
Resets the check_index_page_at_flush field of a page if found in the buffer
pool. */
void
buf_reset_check_index_page_at_flush(
/*================================*/
ulint space, /* in: space id */
ulint offset) /* in: page number */
{
buf_block_t* block;
mutex_enter_fast(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block = (buf_block_t*) buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
if (block && buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->check_index_page_at_flush = FALSE;
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
}
/************************************************************************
Returns the current state of is_hashed of a page. FALSE if the page is
not in the pool. NOTE that this operation does not fix the page in the
pool if it is found there. */
ibool
buf_page_peek_if_search_hashed(
/*===========================*/
/* out: TRUE if page hash index is built in search
system */
ulint space, /* in: space id */
ulint offset) /* in: page number */
{
buf_block_t* block;
ibool is_hashed;
mutex_enter_fast(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block = (buf_block_t*) buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
if (!block || buf_block_get_state(block) != BUF_BLOCK_FILE_PAGE) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
is_hashed = FALSE;
} else {
is_hashed = block->is_hashed;
}
mutex_exit(&(buf_pool->mutex));
return(is_hashed);
}
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Sets file_page_was_freed TRUE if the page is found in the buffer pool.
This function should be called when we free a file page and want the
debug version to check that it is not accessed any more unless
reallocated. */
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
buf_page_t*
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_page_set_file_page_was_freed(
/*=============================*/
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
/* out: control block if found in page hash table,
Import 5.0 code.
2005-10-27 09:29:40 +02:00
otherwise NULL */
ulint space, /* in: space id */
ulint offset) /* in: page number */
{
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
buf_page_t* bpage;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_enter_fast(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
bpage = buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
if (bpage) {
bpage->file_page_was_freed = TRUE;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
mutex_exit(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
return(bpage);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/************************************************************************
Sets file_page_was_freed FALSE if the page is found in the buffer pool.
This function should be called when we free a file page and want the
debug version to check that it is not accessed any more unless
reallocated. */
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
buf_page_t*
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_page_reset_file_page_was_freed(
/*===============================*/
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
/* out: control block if found in page hash table,
Import 5.0 code.
2005-10-27 09:29:40 +02:00
otherwise NULL */
ulint space, /* in: space id */
ulint offset) /* in: page number */
{
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
buf_page_t* bpage;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_enter_fast(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
bpage = buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
if (bpage) {
bpage->file_page_was_freed = FALSE;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
mutex_exit(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
return(bpage);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#endif /* UNIV_DEBUG_FILE_ACCESSES */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: When retrieving compressed BLOBs, avoid allocating uncompressed page frames. buf_page_get_zip(), buf_page_release_zip(): New functions, used by btr_copy_zblob_prefix(). btr_copy_zblob_prefix(): Do not call inflateEnd().
2007-01-19 00:10:49 +01:00
/************************************************************************
Get read access to a compressed page (usually FIL_PAGE_TYPE_ZBLOB). */
buf_page_t*
buf_page_get_zip(
/*=============*/
/* out: pointer to the block */
ulint space, /* in: space id */
ulint zip_size,/* in: compressed page size */
ulint offset) /* in: page number */
{
buf_page_t* bpage;
mutex_t* block_mutex;
ibool must_read;
#ifndef UNIV_LOG_DEBUG
ut_ad(!ibuf_inside());
#endif
buf_pool->n_page_gets++;
for (;;) {
mutex_enter_fast(&buf_pool->mutex);
lookup:
bpage = buf_page_hash_get(space, offset);
if (bpage) {
break;
}
/* Page not in buf_pool: needs to be read from file */
mutex_exit(&buf_pool->mutex);
buf_read_page(space, zip_size, offset);
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 37 || buf_validate());
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
}
if (UNIV_UNLIKELY(!bpage->zip.data)) {
/* There is no compressed page. */
mutex_exit(&buf_pool->mutex);
return(NULL);
}
block_mutex = buf_page_get_mutex(bpage);
mutex_enter(block_mutex);
switch (buf_page_get_state(bpage)) {
case BUF_BLOCK_NOT_USED:
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
case BUF_BLOCK_ZIP_FREE:
ut_error;
break;
case BUF_BLOCK_ZIP_PAGE:
case BUF_BLOCK_ZIP_DIRTY:
bpage->buf_fix_count++;
break;
case BUF_BLOCK_FILE_PAGE:
/* Discard the uncompressed page frame if possible. */
if (buf_LRU_free_block(bpage, FALSE)) {
mutex_exit(block_mutex);
goto lookup;
}
buf_block_buf_fix_inc((buf_block_t*) bpage,
__FILE__, __LINE__);
break;
}
must_read = buf_page_get_io_fix(bpage) == BUF_IO_READ;
mutex_exit(&buf_pool->mutex);
buf_page_set_accessed(bpage, TRUE);
mutex_exit(block_mutex);
buf_block_make_young(bpage);
#ifdef UNIV_DEBUG_FILE_ACCESSES
ut_a(!bpage->file_page_was_freed);
#endif
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 5771 || buf_validate());
ut_a(bpage->buf_fix_count > 0);
ut_a(buf_page_in_file(bpage));
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
if (must_read) {
/* Let us wait until the read operation
completes */
for (;;) {
enum buf_io_fix io_fix;
mutex_enter(block_mutex);
io_fix = buf_page_get_io_fix(bpage);
mutex_exit(block_mutex);
if (io_fix == BUF_IO_READ) {
os_thread_sleep(WAIT_FOR_READ);
} else {
break;
}
}
}
#ifdef UNIV_IBUF_DEBUG
ut_a(ibuf_count_get(buf_page_get_space(bpage),
buf_page_get_page_no(bpage)) == 0);
#endif
return(bpage);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
This is the general function used to get access to a database page. */
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
buf_block_t*
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_page_get_gen(
/*=============*/
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
/* out: pointer to the block or NULL */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint space, /* in: space id */
branches/zip: Remove the fil_space_get_zip_size() call from buf_page_get_gen(). This saves one mutex operation per block request. buf_page_get_gen(), various macros and functions: Add parameter zip_size. btr_node_ptr_get_child(): Add parameter index. fil_space_get_latch(): Add optional output parameter zip_size. fil_space_get_zip_size(): Return 0 for space id==0, because the system tablespace is never compressed. fsp_header_init(): Remove the parameter zip_size. ibuf_free_excess_pages(): Remove the parameter zip_size. trx_rseg_t, trx_undo_t: Add field zip_size. xdes_lst_get_next(): Remove, unused.
2007-01-18 10:59:00 +01:00
ulint zip_size,/* in: compressed page size in bytes
or 0 for uncompressed pages */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint offset, /* in: page number */
ulint rw_latch,/* in: RW_S_LATCH, RW_X_LATCH, RW_NO_LATCH */
branches/zip: buf_page_get_release_on_io(), buf_page_get_gen(): Replace buf_frame_t* guess with buf_block_t* guess in order to avoid a buf_block_align() call. trx_undo_t: Replace page_t* guess_page with buf_block_t* guess_block. btr_search_t: Replace page_t* root_guess with buf_block_t* root_guess.
2006-10-23 21:34:45 +02:00
buf_block_t* guess, /* in: guessed block or NULL */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint mode, /* in: BUF_GET, BUF_GET_IF_IN_POOL,
BUF_GET_NO_LATCH, BUF_GET_NOWAIT */
const char* file, /* in: file name */
ulint line, /* in: line where called */
mtr_t* mtr) /* in: mini-transaction */
{
buf_block_t* block;
ibool accessed;
ulint fix_type;
ibool success;
ibool must_read;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(mtr);
ut_ad((rw_latch == RW_S_LATCH)
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
|| (rw_latch == RW_X_LATCH)
|| (rw_latch == RW_NO_LATCH));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad((mode != BUF_GET_NO_LATCH) || (rw_latch == RW_NO_LATCH));
ut_ad((mode == BUF_GET) || (mode == BUF_GET_IF_IN_POOL)
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
|| (mode == BUF_GET_NO_LATCH) || (mode == BUF_GET_NOWAIT));
branches/zip: Remove the fil_space_get_zip_size() call from buf_page_get_gen(). This saves one mutex operation per block request. buf_page_get_gen(), various macros and functions: Add parameter zip_size. btr_node_ptr_get_child(): Add parameter index. fil_space_get_latch(): Add optional output parameter zip_size. fil_space_get_zip_size(): Return 0 for space id==0, because the system tablespace is never compressed. fsp_header_init(): Remove the parameter zip_size. ibuf_free_excess_pages(): Remove the parameter zip_size. trx_rseg_t, trx_undo_t: Add field zip_size. xdes_lst_get_next(): Remove, unused.
2007-01-18 10:59:00 +01:00
ut_ad(zip_size == fil_space_get_zip_size(space));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifndef UNIV_LOG_DEBUG
branches/zip: Minor cleanup. buf_page_get_gen(): Cache the result of fil_space_get_zip_size(). trx_commit_off_kernel(): Eliminate the flag must_flush_log. Initialize lsn = 0 to signify must_flush_log == FALSE. log_flush_margin(): Eliminate the flag do_flush. Initialize lsn = 0 to signify do_flush == FALSE.
2006-12-29 10:07:01 +01:00
ut_ad(!ibuf_inside() || ibuf_page(space, zip_size, offset));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
buf_pool->n_page_gets++;
loop:
block = NULL;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter_fast(&(buf_pool->mutex));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (guess) {
branches/zip: buf_page_get_release_on_io(), buf_page_get_gen(): Replace buf_frame_t* guess with buf_block_t* guess in order to avoid a buf_block_align() call. trx_undo_t: Replace page_t* guess_page with buf_block_t* guess_block. btr_search_t: Replace page_t* root_guess with buf_block_t* root_guess.
2006-10-23 21:34:45 +02:00
block = guess;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
if (offset != block->page.offset
branches/zip: buf_page_get_gen(): When checking a guessed block, also check buf_page_in_file(). This error was introduced in r1082.
2007-01-02 22:25:40 +01:00
|| space != block->page.space
|| !buf_page_in_file(&block->page)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_page_get_gen(): When checking a guessed block, also check buf_page_in_file(). This error was introduced in r1082.
2007-01-02 22:25:40 +01:00
block = guess = NULL;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
if (block == NULL) {
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block = (buf_block_t*) buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
if (block == NULL
|| UNIV_UNLIKELY(buf_block_get_state(block)
!= BUF_BLOCK_FILE_PAGE)) {
branches/zip: Support compressed-only pages in buf_page_create(). buf_page_get_gen(): Note that buf_read_page() may uncompress a compressed-only page. buf_page_create(): Test buf_page_in_file() instead of BUF_BLOCK_FILE_PAGE.
2007-01-05 15:03:44 +01:00
/* Uncompressed page not in buf_pool: needs to be
decompressed or read from file */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
branches/zip: Support compressed-only pages in buf_page_create(). buf_page_get_gen(): Note that buf_read_page() may uncompress a compressed-only page. buf_page_create(): Test buf_page_in_file() instead of BUF_BLOCK_FILE_PAGE.
2007-01-05 15:03:44 +01:00
if (!block && mode == BUF_GET_IF_IN_POOL) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(NULL);
}
branches/zip: Minor cleanup. buf_page_get_gen(): Cache the result of fil_space_get_zip_size(). trx_commit_off_kernel(): Eliminate the flag must_flush_log. Initialize lsn = 0 to signify must_flush_log == FALSE. log_flush_margin(): Eliminate the flag do_flush. Initialize lsn = 0 to signify do_flush == FALSE.
2006-12-29 10:07:01 +01:00
buf_read_page(space, zip_size, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 37 || buf_validate());
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
goto loop;
}
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
must_read = buf_block_get_io_fix(block) == BUF_IO_READ;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
if (must_read && mode == BUF_GET_IF_IN_POOL) {
/* The page is only being read to buffer */
mutex_exit(&buf_pool->mutex);
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
return(NULL);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_block_buf_fix_inc() a macro wrapper for buf_block_buf_fix_inc_func() to eliminate an #ifdef in every invocation.
2006-11-30 10:24:57 +01:00
buf_block_buf_fix_inc(block, file, line);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&buf_pool->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Check if this is the first access to the page */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
accessed = buf_page_is_accessed(&block->page);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_page_set_accessed(&block->page, TRUE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_block_make_young(&block->page);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
ut_a(!block->page.file_page_was_freed);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
#endif
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 5771 || buf_validate());
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
ut_a(block->page.buf_fix_count > 0);
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (mode == BUF_GET_NOWAIT) {
if (rw_latch == RW_S_LATCH) {
success = rw_lock_s_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_S_FIX;
} else {
ut_ad(rw_latch == RW_X_LATCH);
success = rw_lock_x_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_X_FIX;
}
if (!success) {
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.buf_fix_count--;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_SYNC_DEBUG
rw_lock_s_unlock(&(block->debug_latch));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
#endif
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(NULL);
}
} else if (rw_latch == RW_NO_LATCH) {
if (must_read) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* Let us wait until the read operation
Import 5.0 code.
2005-10-27 09:29:40 +02:00
completes */
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
for (;;) {
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
enum buf_io_fix io_fix;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
mutex_enter(&block->mutex);
io_fix = buf_block_get_io_fix(block);
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
if (io_fix == BUF_IO_READ) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
os_thread_sleep(WAIT_FOR_READ);
branches/zip: buf_page_get_gen(): Restore the "break" statement to a "for (;;)" loop that was accidentally removed in r1090.
2006-12-20 15:07:08 +01:00
} else {
break;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
}
fix_type = MTR_MEMO_BUF_FIX;
} else if (rw_latch == RW_S_LATCH) {
rw_lock_s_lock_func(&(block->lock), 0, file, line);
fix_type = MTR_MEMO_PAGE_S_FIX;
} else {
rw_lock_x_lock_func(&(block->lock), 0, file, line);
fix_type = MTR_MEMO_PAGE_X_FIX;
}
mtr_memo_push(mtr, block, fix_type);
if (!accessed) {
/* In the case of a first access, try to apply linear
read-ahead */
branches/zip: Minor cleanup. buf_page_get_gen(): Cache the result of fil_space_get_zip_size(). trx_commit_off_kernel(): Eliminate the flag must_flush_log. Initialize lsn = 0 to signify must_flush_log == FALSE. log_flush_margin(): Eliminate the flag do_flush. Initialize lsn = 0 to signify do_flush == FALSE.
2006-12-29 10:07:01 +01:00
buf_read_ahead_linear(space, zip_size, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
#ifdef UNIV_IBUF_DEBUG
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
ut_a(ibuf_count_get(buf_block_get_space(block),
buf_block_get_page_no(block)) == 0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
return(block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/************************************************************************
This is the general function used to get optimistic access to a database
page. */
ibool
buf_page_optimistic_get_func(
/*=========================*/
/* out: TRUE if success */
ulint rw_latch,/* in: RW_S_LATCH, RW_X_LATCH */
buf_block_t* block, /* in: guessed buffer block */
branches/zip: Replace ib_ulonglong with ib_uint64_t, since we want exactly 64 bits. "long long" might be longer than 64 bits on some systems.
2006-11-29 15:52:16 +01:00
ib_uint64_t modify_clock,/* in: modify clock value if mode is
Import 5.0 code.
2005-10-27 09:29:40 +02:00
..._GUESS_ON_CLOCK */
const char* file, /* in: file name */
ulint line, /* in: line where called */
mtr_t* mtr) /* in: mini-transaction */
{
ibool accessed;
ibool success;
ulint fix_type;
ut_ad(mtr && block);
ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
if (UNIV_UNLIKELY(buf_block_get_state(block) != BUF_BLOCK_FILE_PAGE)) {
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(FALSE);
}
branches/zip: Make buf_block_buf_fix_inc() a macro wrapper for buf_block_buf_fix_inc_func() to eliminate an #ifdef in every invocation.
2006-11-30 10:24:57 +01:00
buf_block_buf_fix_inc(block, file, line);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
accessed = buf_page_is_accessed(&block->page);
buf_page_set_accessed(&block->page, TRUE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_block_make_young(&block->page);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
/* Check if this is the first access to the page */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Remove many fil_space_get_zip_size() calls. ibuf_page(), ibuf_page_low(), ibuf_free_excess_pages(), ibuf_insert(), buf_read_page(), buf_read_ahead_linear(), buf_read_recv_pages(): Add parameter zip_size.
2006-10-19 13:07:50 +02:00
ut_ad(!ibuf_inside()
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
|| ibuf_page(buf_block_get_space(block),
buf_block_get_zip_size(block),
buf_block_get_page_no(block)));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (rw_latch == RW_S_LATCH) {
success = rw_lock_s_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_S_FIX;
} else {
success = rw_lock_x_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_X_FIX;
}
if (UNIV_UNLIKELY(!success)) {
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.buf_fix_count--;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_SYNC_DEBUG
rw_lock_s_unlock(&(block->debug_latch));
#endif
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
return(FALSE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Introduce the unsigned 64-bit data type ib_ulonglong for log sequence numbers, to replace dulint.
2006-11-24 14:05:01 +01:00
if (UNIV_UNLIKELY(modify_clock != block->modify_clock)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_SYNC_DEBUG
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
buf_block_dbg_add_level(block, SYNC_NO_ORDER_CHECK);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif /* UNIV_SYNC_DEBUG */
if (rw_latch == RW_S_LATCH) {
rw_lock_s_unlock(&(block->lock));
} else {
rw_lock_x_unlock(&(block->lock));
}
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.buf_fix_count--;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_SYNC_DEBUG
rw_lock_s_unlock(&(block->debug_latch));
#endif
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
return(FALSE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
mtr_memo_push(mtr, block, fix_type);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 5771 || buf_validate());
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
ut_a(block->page.buf_fix_count > 0);
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
ut_a(block->page.file_page_was_freed == FALSE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
if (UNIV_UNLIKELY(!accessed)) {
/* In the case of a first access, try to apply linear
read-ahead */
branches/zip: Remove many fil_space_get_zip_size() calls. ibuf_page(), ibuf_page_low(), ibuf_free_excess_pages(), ibuf_insert(), buf_read_page(), buf_read_ahead_linear(), buf_read_recv_pages(): Add parameter zip_size.
2006-10-19 13:07:50 +02:00
buf_read_ahead_linear(buf_block_get_space(block),
buf_block_get_zip_size(block),
buf_block_get_page_no(block));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
#ifdef UNIV_IBUF_DEBUG
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
ut_a(ibuf_count_get(buf_block_get_space(block),
buf_block_get_page_no(block)) == 0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
buf_pool->n_page_gets++;
return(TRUE);
}
/************************************************************************
This is used to get access to a known database page, when no waiting can be
done. For example, if a search in an adaptive hash index leads us to this
frame. */
ibool
buf_page_get_known_nowait(
/*======================*/
/* out: TRUE if success */
ulint rw_latch,/* in: RW_S_LATCH, RW_X_LATCH */
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
buf_block_t* block, /* in: the known page */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint mode, /* in: BUF_MAKE_YOUNG or BUF_KEEP_OLD */
const char* file, /* in: file name */
ulint line, /* in: line where called */
mtr_t* mtr) /* in: mini-transaction */
{
ibool success;
ulint fix_type;
ut_ad(mtr);
ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
if (buf_block_get_state(block) == BUF_BLOCK_REMOVE_HASH) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* Another thread is just freeing the block from the LRU list
of the buffer pool: do not try to access this page; this
Import 5.0 code.
2005-10-27 09:29:40 +02:00
attempt to access the page can only come through the hash
index because when the buffer block state is ..._REMOVE_HASH,
we have already removed it from the page address hash table
of the buffer pool. */
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(FALSE);
}
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Make buf_block_buf_fix_inc() a macro wrapper for buf_block_buf_fix_inc_func() to eliminate an #ifdef in every invocation.
2006-11-30 10:24:57 +01:00
buf_block_buf_fix_inc(block, file, line);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (mode == BUF_MAKE_YOUNG) {
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_block_make_young(&block->page);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
ut_ad(!ibuf_inside() || (mode == BUF_KEEP_OLD));
if (rw_latch == RW_S_LATCH) {
success = rw_lock_s_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_S_FIX;
} else {
success = rw_lock_x_lock_func_nowait(&(block->lock),
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file, line);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fix_type = MTR_MEMO_PAGE_X_FIX;
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (!success) {
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.buf_fix_count--;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_SYNC_DEBUG
rw_lock_s_unlock(&(block->debug_latch));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
#endif
Import 5.0 code.
2005-10-27 09:29:40 +02:00
return(FALSE);
}
mtr_memo_push(mtr, block, fix_type);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 5771 || buf_validate());
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
ut_a(block->page.buf_fix_count > 0);
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
ut_a(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
ut_a(block->page.file_page_was_freed == FALSE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
#ifdef UNIV_IBUF_DEBUG
ut_a((mode == BUF_KEEP_OLD)
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
|| (ibuf_count_get(buf_block_get_space(block),
buf_block_get_page_no(block)) == 0));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
buf_pool->n_page_gets++;
return(TRUE);
}
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
/************************************************************************
Initialize some fields of a control block. */
UNIV_INLINE
void
buf_block_init_low(
/*===============*/
buf_block_t* block) /* in: block to init */
{
block->check_index_page_at_flush = FALSE;
block->index = NULL;
block->n_hash_helps = 0;
block->is_hashed = FALSE;
block->n_fields = 1;
block->n_bytes = 0;
block->left_side = TRUE;
}
/************************************************************************
Initialize some fields of a control block. */
UNIV_INLINE
void
buf_page_init_low(
/*==============*/
buf_page_t* bpage) /* in: block to init */
{
bpage->accessed = FALSE;
bpage->io_fix = BUF_IO_NONE;
bpage->buf_fix_count = 0;
bpage->freed_page_clock = 0;
// bpage->hash = (buf_page_t*) -1;
bpage->newest_modification = 0;
bpage->oldest_modification = 0;
#ifdef UNIV_DEBUG_FILE_ACCESSES
bpage->file_page_was_freed = FALSE;
#endif /* UNIV_DEBUG_FILE_ACCESSES */
}
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
#ifdef UNIV_HOTBACKUP
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Inits a page to the buffer buf_pool, for use in ibbackup --restore. */
void
buf_page_init_for_backup_restore(
/*=============================*/
ulint space, /* in: space id */
ulint offset, /* in: offset of the page within space
in units of a page */
branches/zip: Remove compilation errors with -DUNIV_HOTBACKUP. dict_load_foreigns(): Enclose in #ifndef UNIV_HOTBACKUP. fil_extend_tablespaces_to_stored_len(): Pass zip_size to fil_read(). buf_page_init_for_backup_restore(): Add parameter zip_size. Enclose the declaration in buf0buf.h in #ifdef UNIV_HOTBACKUP. recv_apply_log_recs_for_backup(): Replace the local variable "page" with the local variable "block". Add local variable zip_size.
2006-10-06 13:23:19 +02:00
ulint zip_size,/* in: compressed page size in bytes
or 0 for uncompressed pages */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_block_t* block) /* in: block to init */
{
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
buf_block_init_low(block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->lock_hash_val = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
buf_page_init_low(&block->page);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
block->page.state = BUF_BLOCK_FILE_PAGE;
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
block->page.space = space;
block->page.offset = offset;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Fix compilation errors when UNIV_HOTBACKUP is defined.
2006-11-30 16:25:52 +01:00
page_zip_des_init(&block->page.zip);
branches/zip: Since r1057, page_zip_des_init() affects buf_block_get_state(). buf_block_init(), buf_page_init_for_backup_restore(): Assign block->page.zip.state after invoking page_zip_des_init(&block->page).
2006-11-24 14:02:31 +01:00
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
/* We assume that block->page.data has been allocated
branches/zip: Remove compilation errors with -DUNIV_HOTBACKUP. dict_load_foreigns(): Enclose in #ifndef UNIV_HOTBACKUP. fil_extend_tablespaces_to_stored_len(): Pass zip_size to fil_read(). buf_page_init_for_backup_restore(): Add parameter zip_size. Enclose the declaration in buf0buf.h in #ifdef UNIV_HOTBACKUP. recv_apply_log_recs_for_backup(): Replace the local variable "page" with the local variable "block". Add local variable zip_size.
2006-10-06 13:23:19 +02:00
with zip_size == UNIV_PAGE_SIZE. */
ut_ad(zip_size <= UNIV_PAGE_SIZE);
ut_ad(ut_is_2pow(zip_size));
branches/zip: Fix compilation errors when UNIV_HOTBACKUP is defined.
2006-11-30 16:25:52 +01:00
page_zip_set_size(&block->page.zip, zip_size);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
#endif /* UNIV_HOTBACKUP */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Inits a page to the buffer buf_pool. */
static
void
buf_page_init(
/*==========*/
ulint space, /* in: space id */
ulint offset, /* in: offset of the page within space
in units of a page */
buf_block_t* block) /* in: block to init */
{
branches/zip: buf_page_init(): Improve diagnostic output in case of buf_pool->page_hash corruption.
2007-01-10 15:36:15 +01:00
buf_page_t* hash_page;
branches/zip: Make mutex_own() work with UNIV_DEBUG, without UNIV_SYNC_DEBUG.
2007-01-18 19:29:12 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(mutex_own(&(buf_pool->mutex)));
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
ut_ad(mutex_own(&(block->mutex)));
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
ut_a(buf_block_get_state(block) != BUF_BLOCK_FILE_PAGE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Set the state of the block */
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
buf_block_set_file_page(block, space, offset);
branches/zip: Make Valgrind diagnostics more accurate. buf_page_init(): Declare block->frame valid only in the system tablespace. trx_sysf_create(): Declare the rest of the page valid, although it is uninitialized.
2007-01-13 22:15:55 +01:00
#ifdef UNIV_DEBUG_VALGRIND
if (!space) {
/* Silence valid Valgrind warnings about uninitialized
data being written to data files. There are some unused
bytes on some pages that InnoDB does not initialize. */
UNIV_MEM_VALID(block->frame, UNIV_PAGE_SIZE);
}
#endif /* UNIV_DEBUG_VALGRIND */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
buf_block_init_low(block);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
block->lock_hash_val = lock_rec_hash(space, offset);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Insert into the hash table of file pages */
branches/zip: buf_page_init(): Improve diagnostic output in case of buf_pool->page_hash corruption.
2007-01-10 15:36:15 +01:00
hash_page = buf_page_hash_get(space, offset);
if (UNIV_LIKELY_NULL(hash_page)) {
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: Error: page %lu %lu already found"
branches/zip: buf_page_init(): Improve diagnostic output in case of buf_pool->page_hash corruption.
2007-01-10 15:36:15 +01:00
" in the hash table: %p, %p\n",
Import 5.0 code.
2005-10-27 09:29:40 +02:00
(ulong) space,
branches/zip: buf_page_init(): Improve diagnostic output in case of buf_pool->page_hash corruption.
2007-01-10 15:36:15 +01:00
(ulong) offset,
hash_page, block);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
branches/zip: Before calling buf_buddy_alloc() or buf_buddy_free(), release any block mutexes but hold the buf_pool->mutex.
2007-01-03 16:54:05 +01:00
mutex_exit(&block->mutex);
mutex_exit(&buf_pool->mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
buf_print();
buf_LRU_print();
buf_validate();
buf_LRU_validate();
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
ut_error;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_page_t: Add the in_page_hash and in_zip_hash for extra validation in UNIV_DEBUG builds.
2007-01-24 11:36:05 +01:00
ut_ad(!block->page.in_zip_hash);
ut_ad(!block->page.in_page_hash);
ut_d(block->page.in_page_hash = TRUE);
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
HASH_INSERT(buf_page_t, hash, buf_pool->page_hash,
buf_page_address_fold(space, offset), &block->page);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
buf_page_init_low(&block->page);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
/************************************************************************
Decompress a block. */
static
ibool
buf_zip_decompress(
/*===============*/
/* out: TRUE if successful */
branches/zip: Maintain a valid checksum for compressed-only pages kept in the buffer pool. buf_LRU_free_block(): When preserving the compressed page, compute the checksum. Also assert page_zip_validate(). buf_zip_decompress(): Add parameter "check" for enabling checksum tests. Before decompressing a compressed-only page from the buffer pool, verify the compressed page checksum.
2007-01-08 18:08:57 +01:00
buf_block_t* block, /* in/out: block */
ibool check) /* in: TRUE=verify the page checksum */
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
{
const byte* frame = block->page.zip.data;
ut_ad(buf_block_get_zip_size(block));
ut_a(buf_block_get_space(block) != 0);
branches/zip: Maintain a valid checksum for compressed-only pages kept in the buffer pool. buf_LRU_free_block(): When preserving the compressed page, compute the checksum. Also assert page_zip_validate(). buf_zip_decompress(): Add parameter "check" for enabling checksum tests. Before decompressing a compressed-only page from the buffer pool, verify the compressed page checksum.
2007-01-08 18:08:57 +01:00
if (UNIV_LIKELY(check)) {
ulint stamp_checksum = mach_read_from_4(
frame + FIL_PAGE_SPACE_OR_CHKSUM);
ulint calc_checksum = page_zip_calc_checksum(
frame, page_zip_get_size(&block->page.zip));
if (UNIV_UNLIKELY(stamp_checksum != calc_checksum)) {
ut_print_timestamp(stderr);
fprintf(stderr,
" InnoDB: compressed page checksum mismatch"
" (space %u page %u): %lu != %lu\n",
block->page.space, block->page.offset,
stamp_checksum, calc_checksum);
return(FALSE);
}
}
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
switch (fil_page_get_type(frame)) {
case FIL_PAGE_INDEX:
if (page_zip_decompress(&block->page.zip,
block->frame)) {
return(TRUE);
}
fprintf(stderr,
"InnoDB: unable to decompress space %lu page %lu\n",
(ulong) block->page.space,
(ulong) block->page.offset);
return(FALSE);
case FIL_PAGE_TYPE_ALLOCATED:
case FIL_PAGE_INODE:
case FIL_PAGE_IBUF_BITMAP:
case FIL_PAGE_TYPE_FSP_HDR:
case FIL_PAGE_TYPE_XDES:
case FIL_PAGE_TYPE_ZBLOB:
/* Copy to uncompressed storage. */
memcpy(block->frame, frame,
buf_block_get_zip_size(block));
return(TRUE);
}
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Maintain a valid checksum for compressed-only pages kept in the buffer pool. buf_LRU_free_block(): When preserving the compressed page, compute the checksum. Also assert page_zip_validate(). buf_zip_decompress(): Add parameter "check" for enabling checksum tests. Before decompressing a compressed-only page from the buffer pool, verify the compressed page checksum.
2007-01-08 18:08:57 +01:00
" InnoDB: unknown compressed page"
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
" type %lu\n",
fil_page_get_type(frame));
return(FALSE);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Function which inits a page for read to the buffer buf_pool. If the page is
(1) already in buf_pool, or
(2) if we specify to read only ibuf pages and the page is not an ibuf page, or
(3) if the space is deleted or being deleted,
then this function does nothing.
Sets the io_fix flag to BUF_IO_READ and sets a non-recursive exclusive lock
on the buffer frame. The io-handler must take care that the flag is cleared
and the lock released later. This is one of the functions which perform the
state transition NOT_USED => FILE_PAGE to a block (the other is
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
buf_page_create). */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_block_t*
buf_page_init_for_read(
/*===================*/
/* out: pointer to the block or NULL */
ulint* err, /* out: DB_SUCCESS or DB_TABLESPACE_DELETED */
ulint mode, /* in: BUF_READ_IBUF_PAGES_ONLY, ... */
ulint space, /* in: space id */
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
ulint zip_size,/* in: compressed page size, or 0 */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ib_longlong tablespace_version,/* in: prevents reading from a wrong
version of the tablespace in case we have done
DISCARD + IMPORT */
ulint offset) /* in: page number */
{
buf_block_t* block;
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
buf_page_t* bpage;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mtr_t mtr;
ut_ad(buf_pool);
*err = DB_SUCCESS;
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
/* It is a read-ahead within an ibuf routine */
branches/zip: When identifying insert buffer bitmap pages, note that the pages occur every zip_size pages in compressed tablespaces, instead of UNIV_PAGESIZE (or XDES_DESCRIBED_PER_PAGE). ibuf_bitmap_page(): Add parameter zip_size. ibuf_fixed_addr_page(): Add parameter space.
2006-06-05 11:04:19 +02:00
ut_ad(!ibuf_bitmap_page(zip_size, offset));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(ibuf_inside());
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mtr_start(&mtr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Remove many fil_space_get_zip_size() calls. ibuf_page(), ibuf_page_low(), ibuf_free_excess_pages(), ibuf_insert(), buf_read_page(), buf_read_ahead_linear(), buf_read_recv_pages(): Add parameter zip_size.
2006-10-19 13:07:50 +02:00
if (!ibuf_page_low(space, zip_size, offset, &mtr)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mtr_commit(&mtr);
return(NULL);
}
} else {
ut_ad(mode == BUF_READ_ANY_PAGE);
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
block = buf_LRU_get_free_block(0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_a(block);
mutex_enter(&(buf_pool->mutex));
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
if (fil_tablespace_deleted_or_being_deleted_in_mem(
space, tablespace_version)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
*err = DB_TABLESPACE_DELETED;
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
goto err_exit;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
bpage = buf_page_hash_get(space, offset);
if (UNIV_LIKELY_NULL(bpage)) {
switch (buf_page_get_state(bpage)) {
case BUF_BLOCK_ZIP_PAGE:
case BUF_BLOCK_ZIP_DIRTY:
ut_a(page_zip_get_size(&bpage->zip) == zip_size);
if (bpage->buf_fix_count
|| buf_page_get_io_fix(bpage)
!= BUF_IO_NONE) {
goto err_exit;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
/* Move the compressed page from bpage to block,
and uncompress it. */
mutex_enter(&buf_pool->zip_mutex);
buf_relocate(bpage, &block->page);
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
buf_block_init_low(block);
block->lock_hash_val = lock_rec_hash(space, offset);
branches/zip: Improve Valgrind instrumentation of allocated memory. UNIV_MEM_FREE(): Declare a memory area free. UNIV_MEM_ALLOC(): Declare a memory area allocated (but uninitialized). UNIV_MEM_DESC(): Associate a memory area with a control block. UNIV_MEM_UNDESC(): Unassociate a control block. trx_sys_create_doublewrite_buf(): Clear the buffer with memset(). buf_page_init(): Add a bogus UNIV_MEM_VALID(block->frame) to silence valid warnings about InnoDB data pages containing uninitialized data. buf_LRU_get_free_only(): Add UNIV_MEM_ALLOC(block->frame). buf_LRU_get_free_block(): Add UNIV_MEM_DESC(block->page.zip.data, block). buf_LRU_free_block(): Add UNIV_MEM_DESC(b->zip.data, b) when allocating a compressed-only control block for a compressed page. buf_LRU_block_free_non_file_page(): Replace UNIV_MEM_INVALID() with UNIV_MEM_FREE(). buf_LRU_block_remove_hashed_page(): Add UNIV_MEM_UNDESC(bpage) when deallocating a compressed-only control block. Add UNIV_MEM_INVALID(block->frame). (The frame should be flagged free by buf_LRU_block_free_non_file_page() moments later.) buf0buddy.c: Disable some extra checks in Valgrind-instrumented builds. Add UNIV_MEM_VALID(), UNIV_MEM_INVALID(), UNIV_MEM_FREE(), UNIV_MEM_ALLOC() as necessary.
2007-01-12 13:36:40 +01:00
UNIV_MEM_DESC(bpage->zip.data, zip_size, block);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
if (buf_page_get_state(&block->page)
== BUF_BLOCK_ZIP_PAGE) {
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
UT_LIST_REMOVE(list, buf_pool->zip_clean,
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
&block->page);
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
ut_ad(!block->page.in_flush_list);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
} else {
/* Relocate buf_pool->flush_list. */
buf_page_t* b;
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
b = UT_LIST_GET_PREV(list, &block->page);
ut_ad(block->page.in_flush_list);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
UT_LIST_REMOVE(list, buf_pool->flush_list,
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
&block->page);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
if (b) {
UT_LIST_INSERT_AFTER(
list, buf_pool->flush_list, b,
&block->page);
} else {
UT_LIST_ADD_FIRST(
list, buf_pool->flush_list,
&block->page);
}
}
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
/* Buffer-fix, I/O-fix, and X-latch the block
for the duration of the decompression. */
branches/zip: buf_relocate(): Copy the block. Invalidate the source block. buf_buddy_block_free(), buf_buddy_free_low(): Flag the freed bytes uninitialized. buf_page_init_for_read(), buf_buddy_relocate(): Do not refer to bpage after invoking buf_relocate(). buf_buddy_relocate(): Flag src uninitialized.
2007-01-12 10:14:15 +01:00
block->page.state = BUF_BLOCK_FILE_PAGE;
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
block->page.buf_fix_count = 1;
branches/zip: buf_page_init_for_read(): IO-fix the block during buf_zip_decompress(). This will prevent btr_cur_search_to_nth_level() from accessing a half-decompressed leaf page. Internal B-tree nodes (non-leaf pages) in B-tree searches are only protected by the tree latch (index->lock) and by buffer-fixing.
2007-01-16 13:40:15 +01:00
buf_block_set_io_fix(block, BUF_IO_READ);
branches/zip: Remove a race condition. buf_page_init_for_read(): Protect the buf_zip_decompress() call with an X-lock on the block. page_zip_validate(): Move the page_is_comp() assertion after the page header comparison.
2007-01-10 11:08:34 +01:00
rw_lock_x_lock(&block->lock);
branches/zip: Before calling buf_buddy_alloc() or buf_buddy_free(), release any block mutexes but hold the buf_pool->mutex.
2007-01-03 16:54:05 +01:00
mutex_exit(&block->mutex);
mutex_exit(&buf_pool->zip_mutex);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
buf_buddy_free(bpage, sizeof *bpage);
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
mutex_exit(&buf_pool->mutex);
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
mtr_commit(&mtr);
}
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
/* Decompress the page while not holding
buf_pool->mutex or block->mutex. */
branches/zip: Maintain a valid checksum for compressed-only pages kept in the buffer pool. buf_LRU_free_block(): When preserving the compressed page, compute the checksum. Also assert page_zip_validate(). buf_zip_decompress(): Add parameter "check" for enabling checksum tests. Before decompressing a compressed-only page from the buffer pool, verify the compressed page checksum.
2007-01-08 18:08:57 +01:00
buf_zip_decompress(block, srv_use_checksums);
branches/zip: buf_page_init_for_read(): When decompressing a compressed block, initialize all the fields of the control block. buf_block_init_low(), buf_page_init_low(): New functions, used in buf_page_init_for_backup_restore(), buf_page_init(), and buf_page_init_for_read().
2007-01-23 15:08:37 +01:00
/* Unfix and unlatch the block. */
branches/zip: buf_page_init_for_read(): IO-fix the block during buf_zip_decompress(). This will prevent btr_cur_search_to_nth_level() from accessing a half-decompressed leaf page. Internal B-tree nodes (non-leaf pages) in B-tree searches are only protected by the tree latch (index->lock) and by buffer-fixing.
2007-01-16 13:40:15 +01:00
mutex_enter(&buf_pool->mutex);
branches/zip: buf_page_init_for_read(): Increment the buf_fix_count during the buf_zip_decompress() call. Otherwise, the block may end up being freed by another thread. This bug was caught by the page_zip_validate() check in buf_LRU_block_remove_hashed_page().
2007-01-10 13:50:10 +01:00
mutex_enter(&block->mutex);
branches/zip: buf_page_init_for_read(): Decrement, do not reset block->page.buf_fix_count. This mistake was made in r1255. When access to the block is requested by some other thread, buf_fix_count can be incremented during the execution of buf_zip_decompress().
2007-01-23 17:12:14 +01:00
block->page.buf_fix_count--;
branches/zip: buf_page_init_for_read(): IO-fix the block during buf_zip_decompress(). This will prevent btr_cur_search_to_nth_level() from accessing a half-decompressed leaf page. Internal B-tree nodes (non-leaf pages) in B-tree searches are only protected by the tree latch (index->lock) and by buffer-fixing.
2007-01-16 13:40:15 +01:00
buf_block_set_io_fix(block, BUF_IO_NONE);
mutex_exit(&buf_pool->mutex);
branches/zip: buf_page_init_for_read(): Increment the buf_fix_count during the buf_zip_decompress() call. Otherwise, the block may end up being freed by another thread. This bug was caught by the page_zip_validate() check in buf_LRU_block_remove_hashed_page().
2007-01-10 13:50:10 +01:00
mutex_exit(&block->mutex);
branches/zip: Remove a race condition. buf_page_init_for_read(): Protect the buf_zip_decompress() call with an X-lock on the block. page_zip_validate(): Move the page_is_comp() assertion after the page header comparison.
2007-01-10 11:08:34 +01:00
rw_lock_x_unlock(&block->lock);
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
return(NULL);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
case BUF_BLOCK_FILE_PAGE:
break;
case BUF_BLOCK_ZIP_FREE:
case BUF_BLOCK_NOT_USED:
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
ut_error;
break;
}
err_exit:
/* The page belongs to a space which has been
deleted or is being deleted, or the page is
already in buf_pool, return */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
mutex_exit(&(buf_pool->mutex));
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
buf_block_free(block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
mtr_commit(&mtr);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Minor improvements. buf_LRU_block_remove_hashed_page(): Return the new state of the block. Only call buf_LRU_block_free_hashed_page() if buf_LRU_block_remove_hashed_page() did not return BUF_BLOCK_ZIP_FREE, that is, the control block was not freed. buf_LRU_insert_zip_clean(): New function, for inserting a compressed-only block into buf_pool->zip_clean in the LRU order. buf_LRU_block_remove_hashed_page(), buf_LRU_free_block(): Add the flag "zip" for controlling if the compressed page of an uncompressed page should be removed. For now, assume zip==TRUE. buf_LRU_get_free_block(): Replace the test for UT_LIST_GET_LEN(buf_pool->free) with a test for the return value of buf_LRU_get_free_only(). Do not free zip.data, as it must already have been freed. buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove compressed-only blocks from the buf_pool->zip_clean list. buf_flush_remove(): Restore compressed-only blocks to the buf_pool->zip_clean list. buf_page_init_for_read(): Uncompress compressed-only blocks when possible. Currently, there cannot be any compressed-only blocks in the buffer pool; they would be inserted by buf_LRU_free_block(bpage, zip=FALSE).
2006-12-15 16:05:18 +01:00
return(NULL);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(block);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Minor improvements. Introduce FIL_PAGE_ZBLOB_DATA as a synonym for FIL_PAGE_FILE_FLUSH_LSN. btr_store_big_rec_extern_fields(): Make the assertion about dict_table_zip_size() more accurate. buf_LRU_get_free_block(), buf_block_alloc(): Add parameter zip_size. buf_calc_zblob_page_checksum(): Remove. Replace with page_zip_calc_checksum(). buf_page_init(): Remove parameter zip_size. buf_page_io_complete(): Add a placeholder for handling compressed pages. trx_doublewrite_page_inside(): Remove redundant function. page_zip_write_rec(): Relax an overly tight assertion about blob_no.
2006-05-04 13:44:49 +02:00
buf_page_init(space, offset, block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* The block must be put to the LRU list, to the old blocks */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_LRU_add_block(&block->page, TRUE/* to old blocks */);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_page_set_io_fix(&block->page, BUF_IO_READ);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->n_pend_reads++;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: buf_page_init_for_read(): X-latch the block before invoking buf_buddy_alloc(). An IO-fixed block must also be x-latched when block->mutex and buf_pool->mutex are released. buf_page_create(): Protect the block during buf_buddy_alloc() by IO-fixing and X-latching it.
2007-01-16 14:45:52 +01:00
/* We set a pass-type x-lock on the frame because then the same
thread which called for the read operation (and is running now at
this point of code) can wait for the read to complete by waiting
for the x-lock on the frame; if the x-lock were recursive, the
same thread would illegally get the x-lock before the page read
is completed. The x-lock is cleared by the io-handler thread. */
rw_lock_x_lock_gen(&(block->lock), BUF_IO_READ);
branches/zip: When adding a page to the buffer pool, add it to buf_pool->page_hash and buf_pool->LRU before releasing buf_pool->mutex. buf_page_init_for_read(), buf_page_create(): Allocate the compressed page after the block has been added to the buffer pool. Document the reason for this.
2007-01-15 17:51:39 +01:00
if (zip_size) {
void* data;
page_zip_set_size(&block->page.zip, zip_size);
mutex_exit(&block->mutex);
/* buf_pool->mutex may be released and reacquired by
buf_buddy_alloc(). Thus, we must release block->mutex
in order not to break the latching order in
the reacquisition of buf_pool->mutex. We also must
defer this operation until after the block descriptor
has been added to buf_pool->LRU and buf_pool->page_hash. */
data = buf_buddy_alloc(zip_size, TRUE);
mutex_enter(&block->mutex);
block->page.zip.data = data;
}
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
mutex_exit(&(buf_pool->mutex));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (mode == BUF_READ_IBUF_PAGES_ONLY) {
mtr_commit(&mtr);
}
return(block);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/************************************************************************
Initializes a page to the buffer buf_pool. The page is usually not read
from a file even if it cannot be found in the buffer buf_pool. This is one
of the functions which perform to a block a state transition NOT_USED =>
FILE_PAGE (the other is buf_page_init_for_read above). */
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
buf_block_t*
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_page_create(
/*============*/
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
/* out: pointer to the block, page bufferfixed */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint space, /* in: space id */
ulint offset, /* in: offset of the page within space in units of
a page */
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
ulint zip_size,/* in: compressed page size, or 0 */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mtr_t* mtr) /* in: mini-transaction handle */
{
buf_frame_t* frame;
buf_block_t* block;
buf_block_t* free_block = NULL;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(mtr);
branches/zip: Minor cleanup and bug fixes btr_page_reorganize_low(): Rename new_page to temp_page. btr_store_big_rec_extern_fields(): FIL_PAGE_TYPE is 2 bytes, not 4. buf_page_init(), buf_page_create(), buf_read_page_low(), buf_page_init_for_read(): Add parameter zip_size. buf_page_init_for_backup_restore(), recv_apply_log_recs_for_backup(): Enclose in #ifdef UNIV_HOTBACKUP. Enclose some debug code in #ifdef UNIV_LOG_REPLICATE. page_zip_write_header_log(): Replace page_zip with a pointer to the uncompressed page. page_zip_write_rec(): Relax an assertion about blob_no + n_ext. page_copy_rec_list_to_created_page_write_log(): Allow logging to be disabled.
2006-04-26 11:35:18 +02:00
ut_ad(space || !zip_size);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
free_block = buf_LRU_get_free_block(0);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_enter(&(buf_pool->mutex));
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block = (buf_block_t*) buf_page_hash_get(space, offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Support compressed-only pages in buf_page_create(). buf_page_get_gen(): Note that buf_read_page() may uncompress a compressed-only page. buf_page_create(): Test buf_page_in_file() instead of BUF_BLOCK_FILE_PAGE.
2007-01-05 15:03:44 +01:00
if (block && buf_page_in_file(&block->page)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_IBUF_DEBUG
branches/zip: Support compressed-only pages in buf_page_create(). buf_page_get_gen(): Note that buf_read_page() may uncompress a compressed-only page. buf_page_create(): Test buf_page_in_file() instead of BUF_BLOCK_FILE_PAGE.
2007-01-05 15:03:44 +01:00
ut_a(ibuf_count_get(space, offset) == 0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#ifdef UNIV_DEBUG_FILE_ACCESSES
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
block->page.file_page_was_freed = FALSE;
branches/zip: Add #define UNIV_DEBUG_FILE_ACCESSES to univ.i and enclose all related debug code in #ifdef UNIV_DEBUG_FILE_ACCESSES. This should have no effect on the behaviour, as the symbol is not defined by default. It only reduces the size of buf_block_t and removes some assignments and debug functions.
2006-11-16 09:48:28 +01:00
#endif /* UNIV_DEBUG_FILE_ACCESSES */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Page can be found in buf_pool */
mutex_exit(&(buf_pool->mutex));
buf_block_free(free_block);
branches/zip: Remove the fil_space_get_zip_size() call from buf_page_get_gen(). This saves one mutex operation per block request. buf_page_get_gen(), various macros and functions: Add parameter zip_size. btr_node_ptr_get_child(): Add parameter index. fil_space_get_latch(): Add optional output parameter zip_size. fil_space_get_zip_size(): Return 0 for space id==0, because the system tablespace is never compressed. fsp_header_init(): Remove the parameter zip_size. ibuf_free_excess_pages(): Remove the parameter zip_size. trx_rseg_t, trx_undo_t: Add field zip_size. xdes_lst_get_next(): Remove, unused.
2007-01-18 10:59:00 +01:00
return(buf_page_get_with_no_latch(space, zip_size,
offset, mtr));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/* If we get here, the page was not in buf_pool: init it there */
#ifdef UNIV_DEBUG
if (buf_debug_prints) {
fprintf(stderr, "Creating space %lu page %lu to buffer\n",
(ulong) space, (ulong) offset);
}
#endif /* UNIV_DEBUG */
block = free_block;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Minor improvements. Introduce FIL_PAGE_ZBLOB_DATA as a synonym for FIL_PAGE_FILE_FLUSH_LSN. btr_store_big_rec_extern_fields(): Make the assertion about dict_table_zip_size() more accurate. buf_LRU_get_free_block(), buf_block_alloc(): Add parameter zip_size. buf_calc_zblob_page_checksum(): Remove. Replace with page_zip_calc_checksum(). buf_page_init(): Remove parameter zip_size. buf_page_io_complete(): Add a placeholder for handling compressed pages. trx_doublewrite_page_inside(): Remove redundant function. page_zip_write_rec(): Relax an overly tight assertion about blob_no.
2006-05-04 13:44:49 +02:00
buf_page_init(space, offset, block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* The block must be put to the LRU list */
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_LRU_add_block(&block->page, FALSE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Make buf_block_buf_fix_inc() a macro wrapper for buf_block_buf_fix_inc_func() to eliminate an #ifdef in every invocation.
2006-11-30 10:24:57 +01:00
buf_block_buf_fix_inc(block, __FILE__, __LINE__);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
buf_pool->n_pages_created++;
branches/zip: buf_page_init_for_read(): X-latch the block before invoking buf_buddy_alloc(). An IO-fixed block must also be x-latched when block->mutex and buf_pool->mutex are released. buf_page_create(): Protect the block during buf_buddy_alloc() by IO-fixing and X-latching it.
2007-01-16 14:45:52 +01:00
/* Prevent race conditions during buf_buddy_alloc(),
which may release and reacquire buf_pool->mutex,
by IO-fixing and X-latching the block. */
buf_page_set_io_fix(&block->page, BUF_IO_READ);
rw_lock_x_lock(&block->lock);
branches/zip: When adding a page to the buffer pool, add it to buf_pool->page_hash and buf_pool->LRU before releasing buf_pool->mutex. buf_page_init_for_read(), buf_page_create(): Allocate the compressed page after the block has been added to the buffer pool. Document the reason for this.
2007-01-15 17:51:39 +01:00
if (zip_size) {
void* data;
page_zip_set_size(&block->page.zip, zip_size);
mutex_exit(&block->mutex);
/* buf_pool->mutex may be released and reacquired by
buf_buddy_alloc(). Thus, we must release block->mutex
in order not to break the latching order in
the reacquisition of buf_pool->mutex. We also must
defer this operation until after the block descriptor
has been added to buf_pool->LRU and buf_pool->page_hash. */
data = buf_buddy_alloc(zip_size, TRUE);
mutex_enter(&block->mutex);
block->page.zip.data = data;
}
branches/zip: buf_page_init_for_read(): X-latch the block before invoking buf_buddy_alloc(). An IO-fixed block must also be x-latched when block->mutex and buf_pool->mutex are released. buf_page_create(): Protect the block during buf_buddy_alloc() by IO-fixing and X-latching it.
2007-01-16 14:45:52 +01:00
buf_page_set_io_fix(&block->page, BUF_IO_NONE);
rw_lock_x_unlock(&block->lock);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&(buf_pool->mutex));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mtr_memo_push(mtr, block, MTR_MEMO_BUF_FIX);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
buf_page_set_accessed(&block->page, TRUE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Delete possible entries for the page from the insert buffer:
such can exist if the page belonged to an index which was dropped */
branches/zip: Remove some more buf_block_align() calls. btr_search_build_page_hash_index(), fsp_parse_init_file_page(), ibuf_parse_bitmap_init(): Replace page_t* with buf_block_t*. buf_read_ibuf_merge_pages(): Add const qualifiers. Retrieve zip_size. ibuf_merge_or_delete_for_page(): Replace page_t* with buf_block_t*. Add parameter zip_size.
2006-10-19 10:27:34 +02:00
ibuf_merge_or_delete_for_page(NULL, space, offset, zip_size, TRUE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Flush pages from the end of the LRU list if necessary */
buf_flush_free_margin();
frame = block->frame;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
memset(frame + FIL_PAGE_PREV, 0xff, 4);
memset(frame + FIL_PAGE_NEXT, 0xff, 4);
mach_write_to_2(frame + FIL_PAGE_TYPE, FIL_PAGE_TYPE_ALLOCATED);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Reset to zero the file flush lsn field in the page; if the first
page of an ibdata file is 'created' in this function into the buffer
pool then we lose the original contents of the file flush lsn stamp.
Then InnoDB could in a crash recovery print a big, false, corruption
warning if the stamp contains an lsn bigger than the ib_logfile lsn. */
memset(frame + FIL_PAGE_FILE_FLUSH_LSN, 0, 8);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
ut_a(++buf_dbg_counter % 357 || buf_validate());
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_IBUF_DEBUG
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
ut_a(ibuf_count_get(buf_block_get_space(block),
buf_block_get_page_no(block)) == 0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
branches/zip: Reduce the number of buf_block_align() calls. btr_block_get(): New function to return buf_block_t. btr_page_alloc(), buf_page_get_release_on_io(), buf_page_get_gen(), buf_page_create(), fseg_create(), fseg_create_general(): Return buf_block_t. buf_page_get_known_nowait(): Expect buf_block_t instead of buf_frame_t. buf_frame_get_newest_modification(): Replace with buf_block_get_newest_modification(). buf_page_dbg_add_level(): Replace with buf_block_dbg_add_level(). buf_block_get_zip_size(): New function. buf_block_get_page_zip(): Reintroduce. recv_recover_page(): Replace page, space, page_no with block. ibuf_bitmap_page_init(): Replace page, zip_size with block. ibuf_parse_bitmap_init(): Remove the parameter zip_size. btr_search_drop_page_hash_index(): Replace page with block.
2006-10-12 13:05:22 +02:00
return(block);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/************************************************************************
Completes an asynchronous read or write request of a file page to or from
the buffer pool. */
void
buf_page_io_complete(
/*=================*/
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
buf_page_t* bpage) /* in: pointer to the block in question */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
enum buf_io_fix io_type;
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
const ibool uncompressed = (buf_page_get_state(bpage)
== BUF_BLOCK_FILE_PAGE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
ut_a(buf_page_in_file(bpage));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
/* We do not need protect io_fix here by mutex to read
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
it because this is the only function where we can change the value
from BUF_IO_READ or BUF_IO_WRITE to some other value, and our code
ensures that this is the only thread that handles the i/o for this
block. */
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
io_type = buf_page_get_io_fix(bpage);
ut_ad(io_type == BUF_IO_READ || io_type == BUF_IO_WRITE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (io_type == BUF_IO_READ) {
branches/zip: Remove GCC 4.0.4 warnings from make AM_CFLAGS='-std=c89 -Wall -W -pedantic'
2006-05-10 15:58:35 +02:00
ulint read_page_no;
ulint read_space_id;
branches/zip: Write compressed pages to disk. os_aio_simulated_handle(): Temporarily disable os_file_check_page_trailers(), which cannot be invoked on compressed pages. dict_table_add_system_columns(): New function, split from dict_table_add_to_cache(). mlog_parse_index(): Add system columns to the dummy table and identify DB_TRX_ID and DB_ROLL_PTR in the dummy index. buf_LRU_get_free_block(): Note that page_zip->data should be allocated from an aligned memory pool. buf_flush_buffered_writes(): Write compressed pages to disk. buf_flush_post_to_doublewrite_buf(): Copy compressed pages to the doublewrite buffer. Zero fill any excess space. buf_flush_init_for_writing(): Treat all compressed pages the same. buf_read_page_low(): Read compressed pages from disk. buf_page_io_complete(): Process compressed pages. trx_sys_doublewrite_init_or_restore_page(): Process compressed pages. mlog_write_initial_log_record_fast(): Enable a debug printout #ifdef UNIV_LOG_DEBUG. fsp_header_init(), fsp_fill_free_list(): Pass the compressed page size to buf_page_create(). page_zip_compress_write_log(): Flatten the if-else if-else logic. page_zip_parse_write_blob_ptr(): Do not test page_zip if page==NULL. page_zip_parse_write_node_ptr(): Do not test page_zip if page==NULL. Invoke mlog_close() correctly. row_sel_store_row_id_to_prebuilt(): Add UNIV_UNLIKELY hint to an assertion-like test.
2006-05-30 11:04:57 +02:00
byte* frame;
branches/zip: Remove GCC 4.0.4 warnings from make AM_CFLAGS='-std=c89 -Wall -W -pedantic'
2006-05-10 15:58:35 +02:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if (buf_page_get_zip_size(bpage)) {
frame = bpage->zip.data;
if (uncompressed
&& !buf_zip_decompress((buf_block_t*) bpage,
FALSE)) {
branches/zip: Use the buddy allocator for managing compressed pages. There is something wrong with the management of compressed-only blocks (BUF_BLOCK_ZIP_PAGE). To disable the creation of such blocks, set zip=TRUE in buf_LRU_block_remove_hashed_page(). buf_LRU_block_remove_hashed_page(): Release buf_pool->zip_mutex when freeing a compressed-only page and its control block, with buf_buddy_free(). Adapt callers. buf_LRU_block_free_hashed_page(): Change the parameter type from buf_page_t* to buf_block_t*. buf_LRU_free_block(): Move below the definition of buf_LRU_add_block_low(). Allocate block descriptor for compressed-only blocks. buf_LRU_block_free_non_file_page(): Replace ut_free() with buf_buddy_free(). buf_zip_decompress(): New function, split from buf_page_io_complete(). buf_page_init_for_read(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. Decompress compressed-only blocks that are needed again. buf_page_create(): Do not allocate the compressed page until it is really needed, to avoid the overhead of the buddy allocator. buf_validate(): Replace some equality tests on the lengths of the LRU list and the flush lists with greater-or-equal tests, since the counted numbers do not include control blocks for compressed-only pages.
2007-01-03 14:10:46 +01:00
branches/zip: Write compressed pages to disk. os_aio_simulated_handle(): Temporarily disable os_file_check_page_trailers(), which cannot be invoked on compressed pages. dict_table_add_system_columns(): New function, split from dict_table_add_to_cache(). mlog_parse_index(): Add system columns to the dummy table and identify DB_TRX_ID and DB_ROLL_PTR in the dummy index. buf_LRU_get_free_block(): Note that page_zip->data should be allocated from an aligned memory pool. buf_flush_buffered_writes(): Write compressed pages to disk. buf_flush_post_to_doublewrite_buf(): Copy compressed pages to the doublewrite buffer. Zero fill any excess space. buf_flush_init_for_writing(): Treat all compressed pages the same. buf_read_page_low(): Read compressed pages from disk. buf_page_io_complete(): Process compressed pages. trx_sys_doublewrite_init_or_restore_page(): Process compressed pages. mlog_write_initial_log_record_fast(): Enable a debug printout #ifdef UNIV_LOG_DEBUG. fsp_header_init(), fsp_fill_free_list(): Pass the compressed page size to buf_page_create(). page_zip_compress_write_log(): Flatten the if-else if-else logic. page_zip_parse_write_blob_ptr(): Do not test page_zip if page==NULL. page_zip_parse_write_node_ptr(): Do not test page_zip if page==NULL. Invoke mlog_close() correctly. row_sel_store_row_id_to_prebuilt(): Add UNIV_UNLIKELY hint to an assertion-like test.
2006-05-30 11:04:57 +02:00
goto corrupt;
branches/zip: Minor improvements. Introduce FIL_PAGE_ZBLOB_DATA as a synonym for FIL_PAGE_FILE_FLUSH_LSN. btr_store_big_rec_extern_fields(): Make the assertion about dict_table_zip_size() more accurate. buf_LRU_get_free_block(), buf_block_alloc(): Add parameter zip_size. buf_calc_zblob_page_checksum(): Remove. Replace with page_zip_calc_checksum(). buf_page_init(): Remove parameter zip_size. buf_page_io_complete(): Add a placeholder for handling compressed pages. trx_doublewrite_page_inside(): Remove redundant function. page_zip_write_rec(): Relax an overly tight assertion about blob_no.
2006-05-04 13:44:49 +02:00
}
branches/zip: Write compressed pages to disk. os_aio_simulated_handle(): Temporarily disable os_file_check_page_trailers(), which cannot be invoked on compressed pages. dict_table_add_system_columns(): New function, split from dict_table_add_to_cache(). mlog_parse_index(): Add system columns to the dummy table and identify DB_TRX_ID and DB_ROLL_PTR in the dummy index. buf_LRU_get_free_block(): Note that page_zip->data should be allocated from an aligned memory pool. buf_flush_buffered_writes(): Write compressed pages to disk. buf_flush_post_to_doublewrite_buf(): Copy compressed pages to the doublewrite buffer. Zero fill any excess space. buf_flush_init_for_writing(): Treat all compressed pages the same. buf_read_page_low(): Read compressed pages from disk. buf_page_io_complete(): Process compressed pages. trx_sys_doublewrite_init_or_restore_page(): Process compressed pages. mlog_write_initial_log_record_fast(): Enable a debug printout #ifdef UNIV_LOG_DEBUG. fsp_header_init(), fsp_fill_free_list(): Pass the compressed page size to buf_page_create(). page_zip_compress_write_log(): Flatten the if-else if-else logic. page_zip_parse_write_blob_ptr(): Do not test page_zip if page==NULL. page_zip_parse_write_node_ptr(): Do not test page_zip if page==NULL. Invoke mlog_close() correctly. row_sel_store_row_id_to_prebuilt(): Add UNIV_UNLIKELY hint to an assertion-like test.
2006-05-30 11:04:57 +02:00
} else {
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
ut_a(uncompressed);
frame = ((buf_block_t*) bpage)->frame;
branches/zip: Minor improvements. Introduce FIL_PAGE_ZBLOB_DATA as a synonym for FIL_PAGE_FILE_FLUSH_LSN. btr_store_big_rec_extern_fields(): Make the assertion about dict_table_zip_size() more accurate. buf_LRU_get_free_block(), buf_block_alloc(): Add parameter zip_size. buf_calc_zblob_page_checksum(): Remove. Replace with page_zip_calc_checksum(). buf_page_init(): Remove parameter zip_size. buf_page_io_complete(): Add a placeholder for handling compressed pages. trx_doublewrite_page_inside(): Remove redundant function. page_zip_write_rec(): Relax an overly tight assertion about blob_no.
2006-05-04 13:44:49 +02:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* If this page is not uninitialized and not in the
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
doublewrite buffer, then the page number and space id
should be the same as in block. */
branches/zip: Fix some crash recovery bugs. dict_load_table(): Initialize table->flags with zip_size. mlog_parse_nbytes(), mlog_parse_string(): Add parameter page_zip and write the changes also to the compressed page if one is specified. Assert that these functions are not called on FIL_PAGE_INDEX pages. buf_page_io_complete(): Replace block->frame with frame where appropriate. recv_parse_or_apply_log_rec_body(): Add ut_a(!page_zip) where appropriate. page_parse_delete_rec_list(): Add parameter page_zip.
2006-08-02 07:49:15 +02:00
read_page_no = mach_read_from_4(frame + FIL_PAGE_OFFSET);
branches/zip: Replace FIL_PAGE_ZBLOB_SPACE_ID and FIL_PAGE_ZBLOB_DATA with FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID and FIL_PAGE_DATA. The doublewrite buffer needs to read the space_id in order to determine the type of the page. Because FIL_PAGE_TYPE could contain garbage in MySQL/InnoDB 5.0 and earlier versions, we cannot trust fil_page_get_type(). Instead, we have to always store the space_id at the same location. This modification wastes 12 bytes per compressed BLOB page (1.2% on 1-kilobyte pages).
2006-09-27 12:51:05 +02:00
read_space_id = mach_read_from_4(
frame + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID);
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if (bpage->space == TRX_SYS_SPACE
&& trx_doublewrite_page_inside(bpage->offset)) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
ut_print_timestamp(stderr);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Error: reading page %lu\n"
"InnoDB: which is in the"
" doublewrite buffer!\n",
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
(ulong) bpage->offset);
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
} else if (!read_space_id && !read_page_no) {
/* This is likely an uninitialized page. */
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
} else if ((bpage->space
&& bpage->space != read_space_id)
|| bpage->offset != read_page_no) {
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
/* We did not compare space_id to read_space_id
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if bpage->space == 0, because the field on the
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
page may contain garbage in MySQL < 4.1.1,
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
which only supported bpage->space == 0. */
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
ut_print_timestamp(stderr);
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
" InnoDB: Error: space id and page n:o"
" stored in the page\n"
"InnoDB: read in are %lu:%lu,"
" should be %lu:%lu!\n",
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
(ulong) read_space_id, (ulong) read_page_no,
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
(ulong) bpage->space,
(ulong) bpage->offset);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Minor improvements. Introduce FIL_PAGE_ZBLOB_DATA as a synonym for FIL_PAGE_FILE_FLUSH_LSN. btr_store_big_rec_extern_fields(): Make the assertion about dict_table_zip_size() more accurate. buf_LRU_get_free_block(), buf_block_alloc(): Add parameter zip_size. buf_calc_zblob_page_checksum(): Remove. Replace with page_zip_calc_checksum(). buf_page_init(): Remove parameter zip_size. buf_page_io_complete(): Add a placeholder for handling compressed pages. trx_doublewrite_page_inside(): Remove redundant function. page_zip_write_rec(): Relax an overly tight assertion about blob_no.
2006-05-04 13:44:49 +02:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* From version 3.23.38 up we store the page checksum
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
to the 4 first bytes of the page end lsn field */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: page_zip_des_t: Rename "size" to "ssize" and reduce the storage size from 16 to 3 bits. page_zip_get_size(), page_zip_set_size(): New functions. Replace direct references to page_zip_des_t:size with calls to buf_block_get_zip_size(), page_zip_get_size(), and page_zip_set_size().
2006-11-27 14:44:32 +01:00
if (buf_page_is_corrupted(frame,
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
buf_page_get_zip_size(bpage))) {
branches/zip: Write compressed pages to disk. os_aio_simulated_handle(): Temporarily disable os_file_check_page_trailers(), which cannot be invoked on compressed pages. dict_table_add_system_columns(): New function, split from dict_table_add_to_cache(). mlog_parse_index(): Add system columns to the dummy table and identify DB_TRX_ID and DB_ROLL_PTR in the dummy index. buf_LRU_get_free_block(): Note that page_zip->data should be allocated from an aligned memory pool. buf_flush_buffered_writes(): Write compressed pages to disk. buf_flush_post_to_doublewrite_buf(): Copy compressed pages to the doublewrite buffer. Zero fill any excess space. buf_flush_init_for_writing(): Treat all compressed pages the same. buf_read_page_low(): Read compressed pages from disk. buf_page_io_complete(): Process compressed pages. trx_sys_doublewrite_init_or_restore_page(): Process compressed pages. mlog_write_initial_log_record_fast(): Enable a debug printout #ifdef UNIV_LOG_DEBUG. fsp_header_init(), fsp_fill_free_list(): Pass the compressed page size to buf_page_create(). page_zip_compress_write_log(): Flatten the if-else if-else logic. page_zip_parse_write_blob_ptr(): Do not test page_zip if page==NULL. page_zip_parse_write_node_ptr(): Do not test page_zip if page==NULL. Invoke mlog_close() correctly. row_sel_store_row_id_to_prebuilt(): Add UNIV_UNLIKELY hint to an assertion-like test.
2006-05-30 11:04:57 +02:00
corrupt:
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: Database page corruption on disk"
" or a failed\n"
"InnoDB: file read of page %lu.\n"
"InnoDB: You may have to recover"
" from a backup.\n",
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
(ulong) bpage->offset);
buf_page_print(frame, buf_page_get_zip_size(bpage));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
fprintf(stderr,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
"InnoDB: Database page corruption on disk"
" or a failed\n"
"InnoDB: file read of page %lu.\n"
"InnoDB: You may have to recover"
" from a backup.\n",
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
(ulong) bpage->offset);
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
fputs("InnoDB: It is also possible that"
" your operating\n"
"InnoDB: system has corrupted its"
" own file cache\n"
"InnoDB: and rebooting your computer"
" removes the\n"
"InnoDB: error.\n"
"InnoDB: If the corrupt page is an index page\n"
"InnoDB: you can also try to"
" fix the corruption\n"
"InnoDB: by dumping, dropping,"
" and reimporting\n"
"InnoDB: the corrupt table."
" You can use CHECK\n"
"InnoDB: TABLE to scan your"
" table for corruption.\n"
"InnoDB: See also"
" http://dev.mysql.com/doc/refman/5.1/en/"
"forcing-recovery.html\n"
"InnoDB: about forcing recovery.\n", stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (srv_force_recovery < SRV_FORCE_IGNORE_CORRUPT) {
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
fputs("InnoDB: Ending processing because of"
" a corrupt database page.\n",
stderr);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
exit(1);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
if (recv_recovery_is_on()) {
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
/* Pages must be uncompressed for crash recovery. */
ut_a(uncompressed);
recv_recover_page(FALSE, TRUE, (buf_block_t*) bpage);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if (uncompressed && !recv_no_ibuf_operations) {
branches/zip: Merge revisions 804:838 from trunk.
2006-09-19 12:14:07 +02:00
ibuf_merge_or_delete_for_page(
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
(buf_block_t*) bpage, bpage->space,
bpage->offset, buf_page_get_zip_size(bpage),
TRUE);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&(buf_pool->mutex));
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
mutex_enter(buf_page_get_mutex(bpage));
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_IBUF_DEBUG
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
ut_a(ibuf_count_get(bpage->space, bpage->offset) == 0);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
/* Because this thread which does the unlocking is not the same that
did the locking, we use a pass value != 0 in unlock, which simply
removes the newest lock debug record, without checking the thread
id. */
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
buf_page_set_io_fix(bpage, BUF_IO_NONE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
switch (io_type) {
case BUF_IO_READ:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* NOTE that the call to ibuf may have moved the ownership of
the x-latch to this OS thread: do not let this confuse you in
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
debugging! */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(buf_pool->n_pend_reads > 0);
buf_pool->n_pend_reads--;
buf_pool->n_pages_read++;
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if (uncompressed) {
rw_lock_x_unlock_gen(&((buf_block_t*) bpage)->lock,
BUF_IO_READ);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
break;
case BUF_IO_WRITE:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/* Write means a flush operation: call the completion
routine in the flush system */
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
buf_flush_write_complete(bpage);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
if (uncompressed) {
rw_lock_s_unlock_gen(&((buf_block_t*) bpage)->lock,
BUF_IO_WRITE);
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->n_pages_written++;
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
break;
default:
ut_error;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
mutex_exit(buf_page_get_mutex(bpage));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
#ifdef UNIV_DEBUG
if (buf_debug_prints) {
branches/zip: buf_page_io_complete(): Change the parameter type from buf_block_t* to buf_page_t*. Handle compressed-only blocks.
2007-01-10 15:36:39 +01:00
fprintf(stderr, "Has %s page space %lu page no %lu\n",
io_type == BUF_IO_READ ? "read" : "written",
(ulong) buf_page_get_space(bpage),
(ulong) buf_page_get_page_no(bpage));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
#endif /* UNIV_DEBUG */
}
/*************************************************************************
Invalidates the file pages in the buffer pool when an archive recovery is
completed. All the file pages buffered must be in a replaceable state when
this function is called: not latched and not modified. */
void
buf_pool_invalidate(void)
/*=====================*/
{
ibool freed;
ut_ad(buf_all_freed());
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
freed = TRUE;
while (freed) {
freed = buf_LRU_search_and_free_block(100);
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_enter(&(buf_pool->mutex));
ut_ad(UT_LIST_GET_LEN(buf_pool->LRU) == 0);
mutex_exit(&(buf_pool->mutex));
}
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/*************************************************************************
Validates the buffer buf_pool data structure. */
ibool
buf_validate(void)
/*==============*/
{
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
buf_page_t* b;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_chunk_t* chunk;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint i;
ulint n_single_flush = 0;
ulint n_lru_flush = 0;
ulint n_list_flush = 0;
ulint n_lru = 0;
ulint n_flush = 0;
ulint n_free = 0;
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
ulint n_zip = 0;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(buf_pool);
mutex_enter(&(buf_pool->mutex));
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk = buf_pool->chunks;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
/* Check the uncompressed blocks. */
branches/zip: enum buf_page_state: Add BUF_BLOCK_ZIP_FREE, BUF_BLOCK_ZIP_DIRTY.
2006-12-05 13:10:30 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (i = buf_pool->n_chunks; i--; chunk++) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
ulint j;
buf_block_t* block = chunk->blocks;
for (j = chunk->size; j--; block++) {
mutex_enter(&block->mutex);
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
switch (buf_block_get_state(block)) {
branches/zip: enum buf_page_state: Add BUF_BLOCK_ZIP_FREE, BUF_BLOCK_ZIP_DIRTY.
2006-12-05 13:10:30 +01:00
case BUF_BLOCK_ZIP_FREE:
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
case BUF_BLOCK_ZIP_PAGE:
branches/zip: enum buf_page_state: Add BUF_BLOCK_ZIP_FREE, BUF_BLOCK_ZIP_DIRTY.
2006-12-05 13:10:30 +01:00
case BUF_BLOCK_ZIP_DIRTY:
/* These should only occur on
zip_clean, zip_free[], or flush_list. */
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
ut_error;
break;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: enum buf_page_state: Add BUF_BLOCK_ZIP_FREE, BUF_BLOCK_ZIP_DIRTY.
2006-12-05 13:10:30 +01:00
case BUF_BLOCK_FILE_PAGE:
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
ut_a(buf_page_hash_get(buf_block_get_space(
block),
buf_block_get_page_no(
block))
branches/zip: Make buf_pool->page_hash point to buf_page_t instead of buf_block_t. Move the fields "hash" and "file_page_was_freed" from buf_block_t to buf_page_t. buf_page_in_file(): New function, for checking block state in assertions.
2006-11-29 14:23:28 +01:00
== &block->page);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#ifdef UNIV_IBUF_DEBUG
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
ut_a(buf_page_get_io_fix(&block->page)
== BUF_IO_READ
branches/zip: Replace all references to block->space and block->offset, which should have been replaced in r1057.
2006-11-24 13:49:59 +01:00
|| !ibuf_count_get(buf_block_get_space(
block),
buf_block_get_page_no(
block)));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
#endif
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
switch (buf_page_get_io_fix(&block->page)) {
case BUF_IO_NONE:
break;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
case BUF_IO_WRITE:
branches/zip: Allow the code to compile with debugging enabled. This was broken in r1066.
2006-11-28 12:08:33 +01:00
switch (buf_page_get_flush_type(
&block->page)) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
case BUF_FLUSH_LRU:
n_lru_flush++;
ut_a(rw_lock_is_locked(
&block->lock,
RW_LOCK_SHARED));
break;
case BUF_FLUSH_LIST:
n_list_flush++;
break;
case BUF_FLUSH_SINGLE_PAGE:
n_single_flush++;
break;
default:
ut_error;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
break;
case BUF_IO_READ:
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
ut_a(rw_lock_is_locked(&block->lock,
RW_LOCK_EX));
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
break;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
n_lru++;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Allow the code to compile with debugging enabled. This was broken in r1066.
2006-11-28 12:08:33 +01:00
if (block->page.oldest_modification > 0) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
n_flush++;
}
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
break;
case BUF_BLOCK_NOT_USED:
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
n_free++;
branches/zip: Replace most references to block->state, block->offset, and block->space with buf_block_get_state(block), buf_block_get_page_no(block), and buf_block_get_space(block). enum buf_block_state: Replaces the #define'd buf_block_t.state values. buf_block_get_state(): New function. buf_block_get_frame(): Add __attribute__((const)).
2006-11-23 14:26:01 +01:00
break;
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
/* do nothing */
break;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
mutex_exit(&block->mutex);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
mutex_enter(&buf_pool->zip_mutex);
/* Check clean compressed-only blocks. */
for (b = UT_LIST_GET_FIRST(buf_pool->zip_clean); b;
b = UT_LIST_GET_NEXT(list, b)) {
ut_a(buf_page_get_state(b) == BUF_BLOCK_ZIP_PAGE);
branches/zip: Fix bugs. buf_LRU_free_block(): Do not allocate the control block when a compressed-only page is being freed. buf_validate(): Note that buf_pool->zip_clean may contain blocks with io_fix == BUF_IO_READ.
2007-01-18 21:29:35 +01:00
switch (buf_page_get_io_fix(b)) {
case BUF_IO_NONE:
/* All clean blocks should be I/O-unfixed. */
break;
case BUF_IO_READ:
/* In buf_LRU_free_block(), we temporarily set
b->io_fix = BUF_IO_READ for a newly allocated
control block in order to lock out
buf_page_init_for_read(). */
break;
default:
ut_error;
break;
}
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
ut_a(!b->oldest_modification);
ut_a(buf_page_hash_get(b->space, b->offset) == b);
n_lru++;
n_zip++;
}
/* Check dirty compressed-only blocks. */
for (b = UT_LIST_GET_FIRST(buf_pool->flush_list); b;
b = UT_LIST_GET_NEXT(list, b)) {
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
ut_ad(b->in_flush_list);
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
switch (buf_page_get_state(b)) {
case BUF_BLOCK_ZIP_DIRTY:
ut_a(b->oldest_modification);
n_lru++;
n_flush++;
n_zip++;
switch (buf_page_get_io_fix(b)) {
case BUF_IO_NONE:
case BUF_IO_READ:
break;
case BUF_IO_WRITE:
switch (buf_page_get_flush_type(b)) {
case BUF_FLUSH_LRU:
n_lru_flush++;
break;
case BUF_FLUSH_LIST:
n_list_flush++;
break;
case BUF_FLUSH_SINGLE_PAGE:
n_single_flush++;
break;
default:
ut_error;
}
break;
}
break;
case BUF_BLOCK_FILE_PAGE:
/* uncompressed page */
break;
case BUF_BLOCK_ZIP_FREE:
case BUF_BLOCK_ZIP_PAGE:
case BUF_BLOCK_NOT_USED:
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
ut_error;
break;
}
ut_a(buf_page_hash_get(b->space, b->offset) == b);
}
mutex_exit(&buf_pool->zip_mutex);
if (n_lru + n_free > buf_pool->curr_size + n_zip) {
fprintf(stderr, "n LRU %lu, n free %lu, pool %lu zip %lu\n",
(ulong) n_lru, (ulong) n_free,
(ulong) buf_pool->curr_size, (ulong) n_zip);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_error;
}
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
ut_a(UT_LIST_GET_LEN(buf_pool->LRU) == n_lru);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (UT_LIST_GET_LEN(buf_pool->free) != n_free) {
fprintf(stderr, "Free list len %lu, free blocks %lu\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
(ulong) UT_LIST_GET_LEN(buf_pool->free),
(ulong) n_free);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_error;
}
ut_a(UT_LIST_GET_LEN(buf_pool->flush_list) == n_flush);
branches/zip: buf_validate(): Validate also the compressed-only blocks.
2007-01-05 13:13:28 +01:00
ut_a(buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE] == n_single_flush);
ut_a(buf_pool->n_flush[BUF_FLUSH_LIST] == n_list_flush);
ut_a(buf_pool->n_flush[BUF_FLUSH_LRU] == n_lru_flush);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
ut_a(buf_LRU_validate());
ut_a(buf_flush_validate());
return(TRUE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#if defined UNIV_DEBUG_PRINT || defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/*************************************************************************
Prints info of the buffer buf_pool data structure. */
void
buf_print(void)
/*===========*/
{
dulint* index_ids;
ulint* counts;
ulint size;
ulint i;
ulint j;
dulint id;
ulint n_found;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_chunk_t* chunk;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
dict_index_t* index;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(buf_pool);
size = buf_pool->curr_size;
index_ids = mem_alloc(sizeof(dulint) * size);
counts = mem_alloc(sizeof(ulint) * size);
mutex_enter(&(buf_pool->mutex));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fprintf(stderr,
"buf_pool size %lu\n"
"database pages %lu\n"
"free pages %lu\n"
"modified database pages %lu\n"
"n pending reads %lu\n"
"n pending flush LRU %lu list %lu single page %lu\n"
"pages read %lu, created %lu, written %lu\n",
(ulong) size,
(ulong) UT_LIST_GET_LEN(buf_pool->LRU),
(ulong) UT_LIST_GET_LEN(buf_pool->free),
(ulong) UT_LIST_GET_LEN(buf_pool->flush_list),
(ulong) buf_pool->n_pend_reads,
(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);
/* Count the number of blocks belonging to each index in the buffer */
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
n_found = 0;
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk = buf_pool->chunks;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (i = buf_pool->n_chunks; i--; chunk++) {
buf_block_t* block = chunk->blocks;
ulint n_blocks = chunk->size;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (; n_blocks--; block++) {
const buf_frame_t* frame = block->frame;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
if (fil_page_get_type(frame) == FIL_PAGE_INDEX) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
id = btr_page_get_index_id(frame);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
/* Look for the id in the index_ids array */
j = 0;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
while (j < n_found) {
if (ut_dulint_cmp(index_ids[j],
id) == 0) {
counts[j]++;
break;
}
j++;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
if (j == n_found) {
n_found++;
index_ids[j] = id;
counts[j] = 1;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
}
}
mutex_exit(&(buf_pool->mutex));
for (i = 0; i < n_found; i++) {
index = dict_index_get_if_in_cache(index_ids[i]);
fprintf(stderr,
"Block count for index %lu in buffer is about %lu",
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
(ulong) ut_dulint_get_low(index_ids[i]),
(ulong) counts[i]);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
if (index) {
putc(' ', stderr);
dict_index_name_print(stderr, NULL, index);
}
putc('\n', stderr);
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mem_free(index_ids);
mem_free(counts);
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
ut_a(buf_validate());
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
branches/zip: Introduce a new preprocessor symbol, UNIV_BUF_DEBUG, and enclose buf_validate() and friends in it.
2006-11-16 10:00:30 +01:00
#endif /* UNIV_DEBUG_PRINT || UNIV_DEBUG || UNIV_BUF_DEBUG */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/*************************************************************************
Returns the number of latched pages in the buffer pool. */
ulint
buf_get_latched_pages_number(void)
branches/zip: buf_get_latched_pages_number(): Count compressed-only blocks too.
2007-01-05 22:00:07 +01:00
/*==============================*/
Import 5.0 code.
2005-10-27 09:29:40 +02:00
{
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_chunk_t* chunk;
branches/zip: buf_get_latched_pages_number(): Count compressed-only blocks too.
2007-01-05 22:00:07 +01:00
buf_page_t* b;
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
ulint i;
ulint fixed_pages_number = 0;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
mutex_enter(&(buf_pool->mutex));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk = buf_pool->chunks;
for (i = buf_pool->n_chunks; i--; chunk++) {
buf_block_t* block;
ulint j;
block = chunk->blocks;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (j = chunk->size; j--; block++) {
branches/zip: Remove most direct references to block->state. buf_block_set_state(): New function. Also checks the state transitions. buf_block_set_file_page(): New function.
2006-11-23 15:12:58 +01:00
if (buf_block_get_state(block)
!= BUF_BLOCK_FILE_PAGE) {
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
continue;
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
mutex_enter(&block->mutex);
branches/zip: Move the fields related to the least-recently-used algorithm of the buffer pool from buf_block_t to buf_page_t. Replace some buf_block_t* parameters with buf_page_t*. Add accessor functions.
2006-11-30 13:27:49 +01:00
if (block->page.buf_fix_count != 0
|| buf_page_get_io_fix(&block->page)
!= BUF_IO_NONE) {
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
fixed_pages_number++;
}
mutex_exit(&block->mutex);
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: buf_get_latched_pages_number(): Count compressed-only blocks too.
2007-01-05 22:00:07 +01:00
mutex_enter(&buf_pool->zip_mutex);
/* Traverse the lists of clean and dirty compressed-only blocks. */
for (b = UT_LIST_GET_FIRST(buf_pool->zip_clean); b;
b = UT_LIST_GET_NEXT(list, b)) {
ut_a(buf_page_get_state(b) == BUF_BLOCK_ZIP_PAGE);
ut_a(buf_page_get_io_fix(b) == BUF_IO_NONE);
if (b->buf_fix_count != 0
|| buf_page_get_io_fix(b) != BUF_IO_NONE) {
fixed_pages_number++;
}
}
for (b = UT_LIST_GET_FIRST(buf_pool->flush_list); b;
b = UT_LIST_GET_NEXT(list, b)) {
branches/zip: buf_page_t: Introduce the field in_flush_list #ifdef UNIV_DEBUG. buf_flush_batch(): Remove the local variable "found".
2007-01-09 13:35:42 +01:00
ut_ad(b->in_flush_list);
branches/zip: buf_get_latched_pages_number(): Count compressed-only blocks too.
2007-01-05 22:00:07 +01:00
switch (buf_page_get_state(b)) {
case BUF_BLOCK_ZIP_DIRTY:
if (b->buf_fix_count != 0
|| buf_page_get_io_fix(b) != BUF_IO_NONE) {
fixed_pages_number++;
}
break;
case BUF_BLOCK_FILE_PAGE:
/* uncompressed page */
break;
case BUF_BLOCK_ZIP_FREE:
case BUF_BLOCK_ZIP_PAGE:
case BUF_BLOCK_NOT_USED:
case BUF_BLOCK_READY_FOR_USE:
case BUF_BLOCK_MEMORY:
case BUF_BLOCK_REMOVE_HASH:
ut_error;
break;
}
}
mutex_exit(&buf_pool->zip_mutex);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
mutex_exit(&(buf_pool->mutex));
branches/zip: Merge revisions 465:532 from trunk.
2006-05-08 08:18:59 +02:00
return(fixed_pages_number);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/*************************************************************************
Returns the number of pending buf pool ios. */
ulint
buf_get_n_pending_ios(void)
/*=======================*/
{
return(buf_pool->n_pend_reads
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ buf_pool->n_flush[BUF_FLUSH_LRU]
+ buf_pool->n_flush[BUF_FLUSH_LIST]
+ buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE]);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
/*************************************************************************
Returns the ratio in percents of modified pages in the buffer pool /
database pages in the buffer pool. */
ulint
buf_get_modified_ratio_pct(void)
/*============================*/
{
ulint ratio;
mutex_enter(&(buf_pool->mutex));
ratio = (100 * UT_LIST_GET_LEN(buf_pool->flush_list))
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/ (1 + UT_LIST_GET_LEN(buf_pool->LRU)
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ UT_LIST_GET_LEN(buf_pool->free));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
/* 1 + is there to avoid division by zero */
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
return(ratio);
}
/*************************************************************************
Prints info of the buffer i/o. */
void
buf_print_io(
/*=========*/
FILE* file) /* in/out: buffer where to print */
{
time_t current_time;
double time_elapsed;
ulint size;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(buf_pool);
size = buf_pool->curr_size;
mutex_enter(&(buf_pool->mutex));
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
fprintf(file,
"Buffer pool size %lu\n"
"Free buffers %lu\n"
"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) UT_LIST_GET_LEN(buf_pool->free),
(ulong) UT_LIST_GET_LEN(buf_pool->LRU),
(ulong) UT_LIST_GET_LEN(buf_pool->flush_list),
(ulong) buf_pool->n_pend_reads,
(ulong) buf_pool->n_flush[BUF_FLUSH_LRU]
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ buf_pool->init_flush[BUF_FLUSH_LRU],
Import 5.0 code.
2005-10-27 09:29:40 +02:00
(ulong) buf_pool->n_flush[BUF_FLUSH_LIST]
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ buf_pool->init_flush[BUF_FLUSH_LIST],
Import 5.0 code.
2005-10-27 09:29:40 +02:00
(ulong) buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE]);
current_time = time(NULL);
time_elapsed = 0.001 + difftime(current_time,
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
buf_pool->last_printout_time);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
buf_pool->last_printout_time = current_time;
fprintf(file,
"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)
/ time_elapsed,
(buf_pool->n_pages_created - buf_pool->n_pages_created_old)
/ time_elapsed,
(buf_pool->n_pages_written - buf_pool->n_pages_written_old)
/ time_elapsed);
if (buf_pool->n_page_gets > buf_pool->n_page_gets_old) {
fprintf(file, "Buffer pool hit rate %lu / 1000\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
(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))));
Import 5.0 code.
2005-10-27 09:29:40 +02:00
} else {
fputs("No buffer pool page gets since the last printout\n",
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
file);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
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;
mutex_exit(&(buf_pool->mutex));
}
/**************************************************************************
Refreshes the statistics used to print per-second averages. */
void
buf_refresh_io_stats(void)
/*======================*/
{
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
buf_pool->last_printout_time = time(NULL);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
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;
}
/*************************************************************************
Checks that all file pages in the buffer are in a replaceable state. */
ibool
buf_all_freed(void)
/*===============*/
{
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
buf_chunk_t* chunk;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ulint i;
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ut_ad(buf_pool);
mutex_enter(&(buf_pool->mutex));
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
chunk = buf_pool->chunks;
branches/zip: Merge revisions 968:1009 from trunk.
2006-11-10 12:15:59 +01:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
for (i = buf_pool->n_chunks; i--; chunk++) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
const buf_block_t* block = buf_chunk_not_freed(chunk);
Import 5.0 code.
2005-10-27 09:29:40 +02:00
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
if (UNIV_LIKELY_NULL(block)) {
fprintf(stderr,
"Page %lu %lu still fixed or dirty\n",
branches/zip: Introduce buf_page_t, a common structure for compressed-only and uncompressed buffer pool pages. buf_block_t: Replace page_zip, space, and offset with buf_page_t page. Replace some integers with bit-fields. enum buf_block_state: Rename to buf_page_state. Add BUF_BLOCK_ZIP_PAGE. page_zip_des_t: Add the field "state". Make the integer fields bit-fields. page_zip_copy(): Document which fields are copied.
2006-11-24 09:32:18 +01:00
(ulong) block->page.space,
(ulong) block->page.offset);
branches/zip: Implement resizeable buffer pool. mysql.patch: Patch to change or add variables to MySQL innodb.patch: Patch to make the master thread poll requests to resize the buffer pool. Replace srv_pool_size and innobase_buffer_pool_size with srv_buf_pool_size, srv_buf_pool_old_size, and srv_buf_pool_curr_size. Add buf_chunk_t, a collection of buf_block_t.
2006-11-14 15:27:26 +01:00
ut_error;
Import 5.0 code.
2005-10-27 09:29:40 +02:00
}
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
mutex_exit(&(buf_pool->mutex));
return(TRUE);
branches/zip: Merge revisions 200:236 from trunk.
2006-02-23 20:25:29 +01:00
}
Import 5.0 code.
2005-10-27 09:29:40 +02:00
/*************************************************************************
Checks that there currently are no pending i/o-operations for the buffer
pool. */
ibool
buf_pool_check_no_pending_io(void)
/*==============================*/
/* out: TRUE if there is no pending i/o */
{
ibool ret;
mutex_enter(&(buf_pool->mutex));
if (buf_pool->n_pend_reads + buf_pool->n_flush[BUF_FLUSH_LRU]
branches/zip: Merge revisions 720:765 from trunk and reindent the code as per revisions r763:765.
2006-08-29 11:30:31 +02:00
+ buf_pool->n_flush[BUF_FLUSH_LIST]
+ buf_pool->n_flush[BUF_FLUSH_SINGLE_PAGE]) {
Import 5.0 code.
2005-10-27 09:29:40 +02:00
ret = FALSE;
} else {
ret = TRUE;
}
mutex_exit(&(buf_pool->mutex));
return(ret);
}
/*************************************************************************
Gets the current length of the free list of buffer blocks. */
ulint
buf_get_free_list_len(void)
/*=======================*/
{
ulint len;
mutex_enter(&(buf_pool->mutex));
len = UT_LIST_GET_LEN(buf_pool->free);
mutex_exit(&(buf_pool->mutex));
return(len);
}
Reference in a new issue Copy permalink