mariadb/innobase/buf/buf0lru.c
heikki@hundin.mysql.fi b1b47e93b1 buf0buf.c, buf0buf.ic, buf0buf.h:
Reduce memory usage of the buffer headers
Many files:
  Merge InnoDB-4.1 with AWE support
2003-01-06 22:07:25 +02:00

853 lines
21 KiB
C

/******************************************************
The database buffer replacement algorithm
(c) 1995 Innobase Oy
Created 11/5/1995 Heikki Tuuri
*******************************************************/
#include "buf0lru.h"
#ifdef UNIV_NONINL
#include "buf0lru.ic"
#include "srv0srv.h" /* Needed to getsrv_print_innodb_monitor */
#endif
#include "ut0byte.h"
#include "ut0lst.h"
#include "ut0rnd.h"
#include "sync0sync.h"
#include "sync0rw.h"
#include "hash0hash.h"
#include "os0sync.h"
#include "fil0fil.h"
#include "btr0btr.h"
#include "buf0buf.h"
#include "buf0flu.h"
#include "buf0rea.h"
#include "btr0sea.h"
#include "os0file.h"
#include "log0recv.h"
/* The number of blocks from the LRU_old pointer onward, including the block
pointed to, must be 3/8 of the whole LRU list length, except that the
tolerance defined below is allowed. Note that the tolerance must be small
enough such that for even the BUF_LRU_OLD_MIN_LEN long LRU list, the
LRU_old pointer is not allowed to point to either end of the LRU list. */
#define BUF_LRU_OLD_TOLERANCE 20
/* The whole LRU list length is divided by this number to determine an
initial segment in buf_LRU_get_recent_limit */
#define BUF_LRU_INITIAL_RATIO 8
/**********************************************************************
Takes a block out of the LRU list and page hash table and sets the block
state to BUF_BLOCK_REMOVE_HASH. */
static
void
buf_LRU_block_remove_hashed_page(
/*=============================*/
buf_block_t* block); /* in: block, must contain a file page and
be in a state where it can be freed; there
may or may not be a hash index to the page */
/**********************************************************************
Puts a file page whose has no hash index to the free list. */
static
void
buf_LRU_block_free_hashed_page(
/*===========================*/
buf_block_t* block); /* in: block, must contain a file page and
be in a state where it can be freed */
/**********************************************************************
Gets the minimum LRU_position field for the blocks in an initial segment
(determined by BUF_LRU_INITIAL_RATIO) of the LRU list. The limit is not
guaranteed to be precise, because the ulint_clock may wrap around. */
ulint
buf_LRU_get_recent_limit(void)
/*==========================*/
/* out: the limit; zero if could not determine it */
{
buf_block_t* block;
ulint len;
ulint limit;
mutex_enter(&(buf_pool->mutex));
len = UT_LIST_GET_LEN(buf_pool->LRU);
if (len < BUF_LRU_OLD_MIN_LEN) {
/* The LRU list is too short to do read-ahead */
mutex_exit(&(buf_pool->mutex));
return(0);
}
block = UT_LIST_GET_FIRST(buf_pool->LRU);
limit = block->LRU_position - len / BUF_LRU_INITIAL_RATIO;
mutex_exit(&(buf_pool->mutex));
return(limit);
}
/**********************************************************************
Look for a replaceable block from the end of the LRU list and put it to
the free list if found. */
ibool
buf_LRU_search_and_free_block(
/*==========================*/
/* out: TRUE if freed */
ulint n_iterations __attribute__((unused))) /* in: how many times
this has been called repeatedly without
result: a high value means that we should
search farther */
{
buf_block_t* block;
ibool freed;
mutex_enter(&(buf_pool->mutex));
freed = FALSE;
block = UT_LIST_GET_LAST(buf_pool->LRU);
while (block != NULL) {
if (buf_flush_ready_for_replace(block)) {
if (buf_debug_prints) {
printf(
"Putting space %lu page %lu to free list\n",
block->space, block->offset);
}
buf_LRU_block_remove_hashed_page(block);
mutex_exit(&(buf_pool->mutex));
/* Remove possible adaptive hash index built on the
page; in the case of AWE the block may not have a
frame at all */
if (block->frame) {
btr_search_drop_page_hash_index(block->frame);
}
mutex_enter(&(buf_pool->mutex));
ut_a(block->buf_fix_count == 0);
buf_LRU_block_free_hashed_page(block);
freed = TRUE;
break;
}
block = UT_LIST_GET_PREV(LRU, block);
}
if (buf_pool->LRU_flush_ended > 0) {
buf_pool->LRU_flush_ended--;
}
if (!freed) {
buf_pool->LRU_flush_ended = 0;
}
mutex_exit(&(buf_pool->mutex));
return(freed);
}
/**********************************************************************
Tries to remove LRU flushed blocks from the end of the LRU list and put them
to the free list. This is beneficial for the efficiency of the insert buffer
operation, as flushed pages from non-unique non-clustered indexes are here
taken out of the buffer pool, and their inserts redirected to the insert
buffer. Otherwise, the flushed blocks could get modified again before read
operations need new buffer blocks, and the i/o work done in flushing would be
wasted. */
void
buf_LRU_try_free_flushed_blocks(void)
/*=================================*/
{
mutex_enter(&(buf_pool->mutex));
while (buf_pool->LRU_flush_ended > 0) {
mutex_exit(&(buf_pool->mutex));
buf_LRU_search_and_free_block(0);
mutex_enter(&(buf_pool->mutex));
}
mutex_exit(&(buf_pool->mutex));
}
/**********************************************************************
Returns a free block from buf_pool. The block is taken off the free list.
If it is empty, blocks are moved from the end of the LRU list to the free
list. */
buf_block_t*
buf_LRU_get_free_block(void)
/*========================*/
/* out: the free control block; also if AWE is
used, it is guaranteed that the block has its
page mapped to a frame when we return */
{
buf_block_t* block = NULL;
ibool freed;
ulint n_iterations = 0;
ibool mon_value_was = 0; /* remove bug */
ibool started_monitor = FALSE;
loop:
mutex_enter(&(buf_pool->mutex));
if (!recv_recovery_on && UT_LIST_GET_LEN(buf_pool->free)
+ UT_LIST_GET_LEN(buf_pool->LRU) < buf_pool->max_size / 10) {
ut_print_timestamp(stderr);
fprintf(stderr,
" InnoDB: ERROR: over 9 / 10 of the buffer pool is occupied by\n"
"InnoDB: lock heaps or the adaptive hash index!\n"
"InnoDB: We intentionally generate a seg fault to print a stack trace\n"
"InnoDB: on Linux!\n");
ut_a(0);
} else if (!recv_recovery_on && UT_LIST_GET_LEN(buf_pool->free)
+ UT_LIST_GET_LEN(buf_pool->LRU) < buf_pool->max_size / 5) {
/* Over 80 % of the buffer pool is occupied by lock heaps
or the adaptive hash index. This may be a memory leak! */
ut_print_timestamp(stderr);
fprintf(stderr,
" InnoDB: WARNING: over 4 / 5 of the buffer pool is occupied by\n"
"InnoDB: lock heaps or the adaptive hash index! Check that your\n"
"InnoDB: transactions do not set too many row locks. Starting InnoDB\n"
"InnoDB: Monitor to print diagnostics, including lock heap and hash index\n"
"InnoDB: sizes.\n");
srv_print_innodb_monitor = TRUE;
} else if (!recv_recovery_on && UT_LIST_GET_LEN(buf_pool->free)
+ UT_LIST_GET_LEN(buf_pool->LRU) < buf_pool->max_size / 4) {
/* Switch off the InnoDB Monitor; this is a simple way
to stop the monitor if the situation becomes less urgent,
but may also surprise users! */
srv_print_innodb_monitor = FALSE;
}
if (buf_pool->LRU_flush_ended > 0) {
mutex_exit(&(buf_pool->mutex));
buf_LRU_try_free_flushed_blocks();
mutex_enter(&(buf_pool->mutex));
}
/* If there is a block in the free list, take it */
if (UT_LIST_GET_LEN(buf_pool->free) > 0) {
block = UT_LIST_GET_FIRST(buf_pool->free);
UT_LIST_REMOVE(free, buf_pool->free, block);
if (srv_use_awe) {
if (block->frame) {
/* Remove from the list of mapped pages */
UT_LIST_REMOVE(awe_LRU_free_mapped,
buf_pool->awe_LRU_free_mapped, block);
} else {
/* We map the page to a frame; second param
FALSE below because we do not want it to be
added to the awe_LRU_free_mapped list */
buf_awe_map_page_to_frame(block, FALSE);
}
}
block->state = BUF_BLOCK_READY_FOR_USE;
mutex_exit(&(buf_pool->mutex));
if (started_monitor) {
srv_print_innodb_monitor = mon_value_was;
}
return(block);
}
/* If no block was in the free list, search from the end of the LRU
list and try to free a block there */
mutex_exit(&(buf_pool->mutex));
freed = buf_LRU_search_and_free_block(n_iterations);
if (freed > 0) {
goto loop;
}
if (n_iterations > 30) {
ut_print_timestamp(stderr);
fprintf(stderr,
"InnoDB: Warning: difficult to find free blocks from\n"
"InnoDB: the buffer pool (%lu search iterations)! Consider\n"
"InnoDB: increasing the buffer pool size.\n",
n_iterations);
fprintf(stderr,
"InnoDB: It is also possible that in your Unix version\n"
"InnoDB: fsync is very slow, or completely frozen inside\n"
"InnoDB: the OS kernel. Then upgrading to a newer version\n"
"InnoDB: of your operating system may help. Look at the\n"
"InnoDB: number of fsyncs in diagnostic info below.\n");
fprintf(stderr,
"InnoDB: Pending flushes (fsync) log: %lu; buffer pool: %lu\n",
fil_n_pending_log_flushes,
fil_n_pending_tablespace_flushes);
fprintf(stderr,
"InnoDB: %lu OS file reads, %lu OS file writes, %lu OS fsyncs\n",
os_n_file_reads, os_n_file_writes, os_n_fsyncs);
fprintf(stderr,
"InnoDB: Starting InnoDB Monitor to print further\n"
"InnoDB: diagnostics to the standard output.\n");
mon_value_was = srv_print_innodb_monitor;
started_monitor = TRUE;
srv_print_innodb_monitor = TRUE;
}
/* No free block was found: try to flush the LRU list */
buf_flush_free_margin();
os_aio_simulated_wake_handler_threads();
if (n_iterations > 10) {
os_thread_sleep(500000);
}
n_iterations++;
goto loop;
}
/***********************************************************************
Moves the LRU_old pointer so that the length of the old blocks list
is inside the allowed limits. */
UNIV_INLINE
void
buf_LRU_old_adjust_len(void)
/*========================*/
{
ulint old_len;
ulint new_len;
ut_ad(buf_pool->LRU_old);
ut_ad(mutex_own(&(buf_pool->mutex)));
ut_ad(3 * (BUF_LRU_OLD_MIN_LEN / 8) > BUF_LRU_OLD_TOLERANCE + 5);
for (;;) {
old_len = buf_pool->LRU_old_len;
new_len = 3 * (UT_LIST_GET_LEN(buf_pool->LRU) / 8);
/* Update the LRU_old pointer if necessary */
if (old_len < new_len - BUF_LRU_OLD_TOLERANCE) {
buf_pool->LRU_old = UT_LIST_GET_PREV(LRU,
buf_pool->LRU_old);
(buf_pool->LRU_old)->old = TRUE;
buf_pool->LRU_old_len++;
} else if (old_len > new_len + BUF_LRU_OLD_TOLERANCE) {
(buf_pool->LRU_old)->old = FALSE;
buf_pool->LRU_old = UT_LIST_GET_NEXT(LRU,
buf_pool->LRU_old);
buf_pool->LRU_old_len--;
} else {
ut_ad(buf_pool->LRU_old); /* Check that we did not
fall out of the LRU list */
return;
}
}
}
/***********************************************************************
Initializes the old blocks pointer in the LRU list.
This function should be called when the LRU list grows to
BUF_LRU_OLD_MIN_LEN length. */
static
void
buf_LRU_old_init(void)
/*==================*/
{
buf_block_t* block;
ut_ad(UT_LIST_GET_LEN(buf_pool->LRU) == BUF_LRU_OLD_MIN_LEN);
/* We first initialize all blocks in the LRU list as old and then use
the adjust function to move the LRU_old pointer to the right
position */
block = UT_LIST_GET_FIRST(buf_pool->LRU);
while (block != NULL) {
block->old = TRUE;
block = UT_LIST_GET_NEXT(LRU, block);
}
buf_pool->LRU_old = UT_LIST_GET_FIRST(buf_pool->LRU);
buf_pool->LRU_old_len = UT_LIST_GET_LEN(buf_pool->LRU);
buf_LRU_old_adjust_len();
}
/**********************************************************************
Removes a block from the LRU list. */
UNIV_INLINE
void
buf_LRU_remove_block(
/*=================*/
buf_block_t* block) /* in: control block */
{
ut_ad(buf_pool);
ut_ad(block);
ut_ad(mutex_own(&(buf_pool->mutex)));
/* If the LRU_old pointer is defined and points to just this block,
move it backward one step */
if (block == buf_pool->LRU_old) {
/* Below: the previous block is guaranteed to exist, because
the LRU_old pointer is only allowed to differ by the
tolerance value from strict 3/8 of the LRU list length. */
buf_pool->LRU_old = UT_LIST_GET_PREV(LRU, block);
(buf_pool->LRU_old)->old = TRUE;
buf_pool->LRU_old_len++;
ut_ad(buf_pool->LRU_old);
}
/* Remove the block from the LRU list */
UT_LIST_REMOVE(LRU, buf_pool->LRU, block);
if (srv_use_awe && block->frame) {
/* Remove from the list of mapped pages */
UT_LIST_REMOVE(awe_LRU_free_mapped,
buf_pool->awe_LRU_free_mapped, block);
}
/* If the LRU list is so short that LRU_old not defined, return */
if (UT_LIST_GET_LEN(buf_pool->LRU) < BUF_LRU_OLD_MIN_LEN) {
buf_pool->LRU_old = NULL;
return;
}
ut_ad(buf_pool->LRU_old);
/* Update the LRU_old_len field if necessary */
if (block->old) {
buf_pool->LRU_old_len--;
}
/* Adjust the length of the old block list if necessary */
buf_LRU_old_adjust_len();
}
/**********************************************************************
Adds a block to the LRU list end. */
UNIV_INLINE
void
buf_LRU_add_block_to_end_low(
/*=========================*/
buf_block_t* block) /* in: control block */
{
buf_block_t* last_block;
ut_ad(buf_pool);
ut_ad(block);
ut_ad(mutex_own(&(buf_pool->mutex)));
block->old = TRUE;
last_block = UT_LIST_GET_LAST(buf_pool->LRU);
if (last_block) {
block->LRU_position = last_block->LRU_position;
} else {
block->LRU_position = buf_pool_clock_tic();
}
UT_LIST_ADD_LAST(LRU, buf_pool->LRU, block);
if (srv_use_awe && block->frame) {
/* Add to the list of mapped pages */
UT_LIST_ADD_LAST(awe_LRU_free_mapped,
buf_pool->awe_LRU_free_mapped, block);
}
if (UT_LIST_GET_LEN(buf_pool->LRU) >= BUF_LRU_OLD_MIN_LEN) {
buf_pool->LRU_old_len++;
}
if (UT_LIST_GET_LEN(buf_pool->LRU) > BUF_LRU_OLD_MIN_LEN) {
ut_ad(buf_pool->LRU_old);
/* Adjust the length of the old block list if necessary */
buf_LRU_old_adjust_len();
} else if (UT_LIST_GET_LEN(buf_pool->LRU) == BUF_LRU_OLD_MIN_LEN) {
/* The LRU list is now long enough for LRU_old to become
defined: init it */
buf_LRU_old_init();
}
}
/**********************************************************************
Adds a block to the LRU list. */
UNIV_INLINE
void
buf_LRU_add_block_low(
/*==================*/
buf_block_t* block, /* in: control block */
ibool old) /* in: TRUE if should be put to the old blocks
in the LRU list, else put to the start; if the
LRU list is very short, the block is added to
the start, regardless of this parameter */
{
ulint cl;
ut_ad(buf_pool);
ut_ad(block);
ut_ad(mutex_own(&(buf_pool->mutex)));
block->old = old;
cl = buf_pool_clock_tic();
if (srv_use_awe && block->frame) {
/* Add to the list of mapped pages; for simplicity we always
add to the start, even if the user would have set 'old'
TRUE */
UT_LIST_ADD_FIRST(awe_LRU_free_mapped,
buf_pool->awe_LRU_free_mapped, block);
}
if (!old || (UT_LIST_GET_LEN(buf_pool->LRU) < BUF_LRU_OLD_MIN_LEN)) {
UT_LIST_ADD_FIRST(LRU, buf_pool->LRU, block);
block->LRU_position = cl;
block->freed_page_clock = buf_pool->freed_page_clock;
} else {
UT_LIST_INSERT_AFTER(LRU, buf_pool->LRU, buf_pool->LRU_old,
block);
buf_pool->LRU_old_len++;
/* We copy the LRU position field of the previous block
to the new block */
block->LRU_position = (buf_pool->LRU_old)->LRU_position;
}
if (UT_LIST_GET_LEN(buf_pool->LRU) > BUF_LRU_OLD_MIN_LEN) {
ut_ad(buf_pool->LRU_old);
/* Adjust the length of the old block list if necessary */
buf_LRU_old_adjust_len();
} else if (UT_LIST_GET_LEN(buf_pool->LRU) == BUF_LRU_OLD_MIN_LEN) {
/* The LRU list is now long enough for LRU_old to become
defined: init it */
buf_LRU_old_init();
}
}
/**********************************************************************
Adds a block to the LRU list. */
void
buf_LRU_add_block(
/*==============*/
buf_block_t* block, /* in: control block */
ibool old) /* in: TRUE if should be put to the old
blocks in the LRU list, else put to the start;
if the LRU list is very short, the block is
added to the start, regardless of this
parameter */
{
buf_LRU_add_block_low(block, old);
}
/**********************************************************************
Moves a block to the start of the LRU list. */
void
buf_LRU_make_block_young(
/*=====================*/
buf_block_t* block) /* in: control block */
{
buf_LRU_remove_block(block);
buf_LRU_add_block_low(block, FALSE);
}
/**********************************************************************
Moves a block to the end of the LRU list. */
void
buf_LRU_make_block_old(
/*===================*/
buf_block_t* block) /* in: control block */
{
buf_LRU_remove_block(block);
buf_LRU_add_block_to_end_low(block);
}
/**********************************************************************
Puts a block back to the free list. */
void
buf_LRU_block_free_non_file_page(
/*=============================*/
buf_block_t* block) /* in: block, must not contain a file page */
{
ut_ad(mutex_own(&(buf_pool->mutex)));
ut_ad(block);
ut_ad((block->state == BUF_BLOCK_MEMORY)
|| (block->state == BUF_BLOCK_READY_FOR_USE));
block->state = BUF_BLOCK_NOT_USED;
#ifdef UNIV_DEBUG
/* Wipe contents of page to reveal possible stale pointers to it */
memset(block->frame, '\0', UNIV_PAGE_SIZE);
#endif
UT_LIST_ADD_FIRST(free, buf_pool->free, block);
if (srv_use_awe && block->frame) {
/* Add to the list of mapped pages */
UT_LIST_ADD_FIRST(awe_LRU_free_mapped,
buf_pool->awe_LRU_free_mapped, block);
}
}
/**********************************************************************
Takes a block out of the LRU list and page hash table and sets the block
state to BUF_BLOCK_REMOVE_HASH. */
static
void
buf_LRU_block_remove_hashed_page(
/*=============================*/
buf_block_t* block) /* in: block, must contain a file page and
be in a state where it can be freed; there
may or may not be a hash index to the page */
{
ut_ad(mutex_own(&(buf_pool->mutex)));
ut_ad(block);
ut_ad(block->state == BUF_BLOCK_FILE_PAGE);
ut_a(block->io_fix == 0);
ut_a(block->buf_fix_count == 0);
ut_a(ut_dulint_cmp(block->oldest_modification, ut_dulint_zero) == 0);
buf_LRU_remove_block(block);
buf_pool->freed_page_clock += 1;
/* Note that if AWE is enabled the block may not have a frame at all */
buf_block_modify_clock_inc(block);
HASH_DELETE(buf_block_t, hash, buf_pool->page_hash,
buf_page_address_fold(block->space, block->offset),
block);
block->state = BUF_BLOCK_REMOVE_HASH;
}
/**********************************************************************
Puts a file page whose has no hash index to the free list. */
static
void
buf_LRU_block_free_hashed_page(
/*===========================*/
buf_block_t* block) /* in: block, must contain a file page and
be in a state where it can be freed */
{
ut_ad(mutex_own(&(buf_pool->mutex)));
ut_ad(block->state == BUF_BLOCK_REMOVE_HASH);
block->state = BUF_BLOCK_MEMORY;
buf_LRU_block_free_non_file_page(block);
}
/**************************************************************************
Validates the LRU list. */
ibool
buf_LRU_validate(void)
/*==================*/
{
buf_block_t* block;
ulint old_len;
ulint new_len;
ulint LRU_pos;
ut_ad(buf_pool);
mutex_enter(&(buf_pool->mutex));
if (UT_LIST_GET_LEN(buf_pool->LRU) >= BUF_LRU_OLD_MIN_LEN) {
ut_a(buf_pool->LRU_old);
old_len = buf_pool->LRU_old_len;
new_len = 3 * (UT_LIST_GET_LEN(buf_pool->LRU) / 8);
ut_a(old_len >= new_len - BUF_LRU_OLD_TOLERANCE);
ut_a(old_len <= new_len + BUF_LRU_OLD_TOLERANCE);
}
UT_LIST_VALIDATE(LRU, buf_block_t, buf_pool->LRU);
block = UT_LIST_GET_FIRST(buf_pool->LRU);
old_len = 0;
while (block != NULL) {
ut_a(block->state == BUF_BLOCK_FILE_PAGE);
if (block->old) {
old_len++;
}
if (buf_pool->LRU_old && (old_len == 1)) {
ut_a(buf_pool->LRU_old == block);
}
LRU_pos = block->LRU_position;
block = UT_LIST_GET_NEXT(LRU, block);
if (block) {
/* If the following assert fails, it may
not be an error: just the buf_pool clock
has wrapped around */
ut_a(LRU_pos >= block->LRU_position);
}
}
if (buf_pool->LRU_old) {
ut_a(buf_pool->LRU_old_len == old_len);
}
UT_LIST_VALIDATE(free, buf_block_t, buf_pool->free);
block = UT_LIST_GET_FIRST(buf_pool->free);
while (block != NULL) {
ut_a(block->state == BUF_BLOCK_NOT_USED);
block = UT_LIST_GET_NEXT(free, block);
}
mutex_exit(&(buf_pool->mutex));
return(TRUE);
}
/**************************************************************************
Prints the LRU list. */
void
buf_LRU_print(void)
/*===============*/
{
buf_block_t* block;
buf_frame_t* frame;
ulint len;
ut_ad(buf_pool);
mutex_enter(&(buf_pool->mutex));
printf("Pool ulint clock %lu\n", buf_pool->ulint_clock);
block = UT_LIST_GET_FIRST(buf_pool->LRU);
len = 0;
while (block != NULL) {
printf("BLOCK %lu ", block->offset);
if (block->old) {
printf("old ");
}
if (block->buf_fix_count) {
printf("buffix count %lu ", block->buf_fix_count);
}
if (block->io_fix) {
printf("io_fix %lu ", block->io_fix);
}
if (ut_dulint_cmp(block->oldest_modification,
ut_dulint_zero) > 0) {
printf("modif. ");
}
printf("LRU pos %lu ", block->LRU_position);
frame = buf_block_get_frame(block);
printf("type %lu ", fil_page_get_type(frame));
printf("index id %lu ", ut_dulint_get_low(
btr_page_get_index_id(frame)));
block = UT_LIST_GET_NEXT(LRU, block);
len++;
if (len % 10 == 0) {
printf("\n");
}
}
mutex_exit(&(buf_pool->mutex));
}