storage/innobase/include/buf0buf.ic

/*****************************************************************************

Copyright (c) 1995, 2023, Oracle and/or its affiliates.
Copyright (c) 2008, Google Inc.

Portions of this file contain modifications contributed and copyrighted by
Google, Inc. Those modifications are gratefully acknowledged and are described
briefly in the InnoDB documentation. The contributions by Google are
incorporated with their permission, and subject to the conditions contained in
the file COPYING.Google.

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License, version 2.0,
as published by the Free Software Foundation.

This program is also distributed with certain software (including
but not limited to OpenSSL) that is licensed under separate terms,
as designated in a particular file or component or in included license
documentation.  The authors of MySQL hereby grant you an additional
permission to link the program and your derivative works with the
separately licensed software that they have included with MySQL.

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, version 2.0, for more details.

You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA

*****************************************************************************/

/**************************************************//**
@file include/buf0buf.ic
The database buffer buf_pool

Created 11/5/1995 Heikki Tuuri
*******************************************************/

#include "mtr0mtr.h"
#ifndef UNIV_HOTBACKUP
#include "buf0flu.h"
#include "buf0lru.h"
#include "buf0rea.h"
#include "sync0debug.h"
#include "fsp0types.h"
#include "ut0new.h"

/** A chunk of buffers. The buffer pool is allocated in chunks. */
struct buf_chunk_t{
	ulint		size;		/*!< size of frames[] and blocks[] */
	unsigned char*	mem;		/*!< pointer to the memory area which
					was allocated for the frames */
	ut_new_pfx_t	mem_pfx;	/*!< Auxiliary structure, describing
					"mem". It is filled by the allocator's
					alloc method and later passed to the
					deallocate method. */
	buf_block_t*	blocks;		/*!< array of buffer control blocks */

	/** Get the size of 'mem' in bytes. */
	size_t	mem_size() const {
		return(mem_pfx.m_size);
	}
};

/*********************************************************************//**
Gets the current size of buffer buf_pool in bytes.
@return size in bytes */
UNIV_INLINE
ulint
buf_pool_get_curr_size(void)
/*========================*/
{
	return(srv_buf_pool_curr_size);
}

/********************************************************************//**
Calculates the index of a buffer pool to the buf_pool[] array.
@return the position of the buffer pool in buf_pool[] */
UNIV_INLINE
ulint
buf_pool_index(
/*===========*/
	const buf_pool_t*	buf_pool)	/*!< in: buffer pool */
{
	ulint	i = buf_pool - buf_pool_ptr;
	ut_ad(i < MAX_BUFFER_POOLS);
	ut_ad(i < srv_buf_pool_instances);
	return(i);
}

/******************************************************************//**
Returns the buffer pool instance given a page instance
@return buf_pool */
UNIV_INLINE
buf_pool_t*
buf_pool_from_bpage(
/*================*/
	const buf_page_t*	bpage) /*!< in: buffer pool page */
{
	ulint	i;
	i = bpage->buf_pool_index;
	ut_ad(i < srv_buf_pool_instances);
	return(&buf_pool_ptr[i]);
}

/******************************************************************//**
Returns the buffer pool instance given a block instance
@return buf_pool */
UNIV_INLINE
buf_pool_t*
buf_pool_from_block(
/*================*/
	const buf_block_t*	block) /*!< in: block */
{
	return(buf_pool_from_bpage(&block->page));
}

/*********************************************************************//**
Gets the current size of buffer buf_pool in pages.
@return size in pages*/
UNIV_INLINE
ulint
buf_pool_get_n_pages(void)
/*======================*/
{
	return(buf_pool_get_curr_size() / UNIV_PAGE_SIZE);
}

/********************************************************************//**
Reads the freed_page_clock of a buffer block.
@return freed_page_clock */
UNIV_INLINE
ulint
buf_page_get_freed_page_clock(
/*==========================*/
	const buf_page_t*	bpage)	/*!< in: block */
{
	/* This is sometimes read without holding buf_pool->mutex. */
	return(bpage->freed_page_clock);
}

/********************************************************************//**
Reads the freed_page_clock of a buffer block.
@return freed_page_clock */
UNIV_INLINE
ulint
buf_block_get_freed_page_clock(
/*===========================*/
	const buf_block_t*	block)	/*!< in: block */
{
	return(buf_page_get_freed_page_clock(&block->page));
}

/********************************************************************//**
Tells if a block is still close enough to the MRU end of the LRU list
meaning that it is not in danger of getting evicted and also implying
that it has been accessed recently.
Note that this is for heuristics only and does not reserve buffer pool
mutex.
@return TRUE if block is close to MRU end of LRU */
UNIV_INLINE
ibool
buf_page_peek_if_young(
/*===================*/
	const buf_page_t*	bpage)	/*!< in: block */
{
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);

	/* FIXME: bpage->freed_page_clock is 31 bits */
	return((buf_pool->freed_page_clock & ((1UL << 31) - 1))
	       < ((ulint) bpage->freed_page_clock
		  + (buf_pool->curr_size
		     * (BUF_LRU_OLD_RATIO_DIV - buf_pool->LRU_old_ratio)
		     / (BUF_LRU_OLD_RATIO_DIV * 4))));
}

/********************************************************************//**
Recommends a move of a block to the start of the LRU list if there is danger
of dropping from the buffer pool. NOTE: does not reserve the buffer pool
mutex.
@return TRUE if should be made younger */
UNIV_INLINE
ibool
buf_page_peek_if_too_old(
/*=====================*/
	const buf_page_t*	bpage)	/*!< in: block to make younger */
{
	buf_pool_t*		buf_pool = buf_pool_from_bpage(bpage);

	if (buf_pool->freed_page_clock == 0) {
		/* If eviction has not started yet, do not update the
		statistics or move blocks in the LRU list.  This is
		either the warm-up phase or an in-memory workload. */
		return(FALSE);
	} else if (buf_LRU_old_threshold_ms && bpage->old) {
		unsigned	access_time = buf_page_is_accessed(bpage);

		/* It is possible that the below comparison returns an
		unexpected result. 2^32 milliseconds pass in about 50 days,
		so if the difference between ut_time_monotonic_ms() and
		access_time is e.g. 50 days + 15 ms, then the below will behave
		as if it is 15 ms. This is known and fixing it would require to
		increase buf_page_t::access_time from 32 to 64 bits. */
		if (access_time > 0
		    && ((ib_uint32_t) (ut_time_monotonic_ms() - access_time))
		    >= buf_LRU_old_threshold_ms) {
			return(TRUE);
		}

		buf_pool->stat.n_pages_not_made_young++;
		return(FALSE);
	} else {
		return(!buf_page_peek_if_young(bpage));
	}
}
#endif /* !UNIV_HOTBACKUP */

/*********************************************************************//**
Gets the state of a block.
@return state */
UNIV_INLINE
enum buf_page_state
buf_page_get_state(
/*===============*/
	const buf_page_t*	bpage)	/*!< in: pointer to the control block */
{
	enum buf_page_state	state	= bpage->state;

#ifdef UNIV_DEBUG
	switch (state) {
	case BUF_BLOCK_POOL_WATCH:
	case BUF_BLOCK_ZIP_PAGE:
	case BUF_BLOCK_ZIP_DIRTY:
	case BUF_BLOCK_NOT_USED:
	case BUF_BLOCK_READY_FOR_USE:
	case BUF_BLOCK_FILE_PAGE:
	case BUF_BLOCK_MEMORY:
	case BUF_BLOCK_REMOVE_HASH:
		break;
	default:
		ut_error;
	}
#endif /* UNIV_DEBUG */

	return(state);
}
/*********************************************************************//**
Gets the state of a block.
@return state */
UNIV_INLINE
enum buf_page_state
buf_block_get_state(
/*================*/
	const buf_block_t*	block)	/*!< in: pointer to the control block */
{
	return(buf_page_get_state(&block->page));
}
/*********************************************************************//**
Sets the state of a block. */
UNIV_INLINE
void
buf_page_set_state(
/*===============*/
	buf_page_t*		bpage,	/*!< in/out: pointer to control block */
	enum buf_page_state	state)	/*!< in: state */
{
#ifdef UNIV_DEBUG
	enum buf_page_state	old_state	= buf_page_get_state(bpage);

	switch (old_state) {
	case BUF_BLOCK_POOL_WATCH:
		ut_error;
		break;
	case BUF_BLOCK_ZIP_PAGE:
		ut_a(state == BUF_BLOCK_ZIP_DIRTY);
		break;
	case BUF_BLOCK_ZIP_DIRTY:
		ut_a(state == BUF_BLOCK_ZIP_PAGE);
		break;
	case BUF_BLOCK_NOT_USED:
		ut_a(state == BUF_BLOCK_READY_FOR_USE);
		break;
	case BUF_BLOCK_READY_FOR_USE:
		ut_a(state == BUF_BLOCK_MEMORY
		     || state == BUF_BLOCK_FILE_PAGE
		     || state == BUF_BLOCK_NOT_USED);
		break;
	case BUF_BLOCK_MEMORY:
		ut_a(state == BUF_BLOCK_NOT_USED);
		break;
	case BUF_BLOCK_FILE_PAGE:
		ut_a(state == BUF_BLOCK_NOT_USED
		     || state == BUF_BLOCK_REMOVE_HASH);
		break;
	case BUF_BLOCK_REMOVE_HASH:
		ut_a(state == BUF_BLOCK_MEMORY);
		break;
	}
#endif /* UNIV_DEBUG */
	bpage->state = state;
	ut_ad(buf_page_get_state(bpage) == state);
}

/*********************************************************************//**
Sets the state of a block. */
UNIV_INLINE
void
buf_block_set_state(
/*================*/
	buf_block_t*		block,	/*!< in/out: pointer to control block */
	enum buf_page_state	state)	/*!< in: state */
{
	buf_page_set_state(&block->page, state);
}

/*********************************************************************//**
Determines if a block is mapped to a tablespace.
@return TRUE if mapped */
UNIV_INLINE
ibool
buf_page_in_file(
/*=============*/
	const buf_page_t*	bpage)	/*!< in: pointer to control block */
{
	switch (buf_page_get_state(bpage)) {
	case BUF_BLOCK_POOL_WATCH:
		ut_error;
		break;
	case BUF_BLOCK_ZIP_PAGE:
	case BUF_BLOCK_ZIP_DIRTY:
	case BUF_BLOCK_FILE_PAGE:
		return(TRUE);
	case BUF_BLOCK_NOT_USED:
	case BUF_BLOCK_READY_FOR_USE:
	case BUF_BLOCK_MEMORY:
	case BUF_BLOCK_REMOVE_HASH:
		break;
	}

	return(FALSE);
}

#ifndef UNIV_HOTBACKUP
/*********************************************************************//**
Determines if a block should be on unzip_LRU list.
@return TRUE if block belongs to unzip_LRU */
UNIV_INLINE
ibool
buf_page_belongs_to_unzip_LRU(
/*==========================*/
	const buf_page_t*	bpage)	/*!< in: pointer to control block */
{
	ut_ad(buf_page_in_file(bpage));

	return(bpage->zip.data
	       && buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE);
}

/*********************************************************************//**
Gets the mutex of a block.
@return pointer to mutex protecting bpage */
UNIV_INLINE
BPageMutex*
buf_page_get_mutex(
/*===============*/
	const buf_page_t*	bpage)	/*!< in: pointer to control block */
{
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);

	switch (buf_page_get_state(bpage)) {
	case BUF_BLOCK_POOL_WATCH:
		ut_error;
		return(NULL);
	case BUF_BLOCK_ZIP_PAGE:
	case BUF_BLOCK_ZIP_DIRTY:
		return(&buf_pool->zip_mutex);
	default:
		return(&((buf_block_t*) bpage)->mutex);
	}
}

/*********************************************************************//**
Get the flush type of a page.
@return flush type */
UNIV_INLINE
buf_flush_t
buf_page_get_flush_type(
/*====================*/
	const buf_page_t*	bpage)	/*!< in: buffer page */
{
	buf_flush_t	flush_type = (buf_flush_t) bpage->flush_type;

#ifdef UNIV_DEBUG
	switch (flush_type) {
	case BUF_FLUSH_LRU:
	case BUF_FLUSH_LIST:
	case BUF_FLUSH_SINGLE_PAGE:
		return(flush_type);
	case BUF_FLUSH_N_TYPES:
		ut_error;
	}
	ut_error;
#endif /* UNIV_DEBUG */
	return(flush_type);
}
/*********************************************************************//**
Set the flush type of a page. */
UNIV_INLINE
void
buf_page_set_flush_type(
/*====================*/
	buf_page_t*	bpage,		/*!< in: buffer page */
	buf_flush_t	flush_type)	/*!< in: flush type */
{
	bpage->flush_type = flush_type;
	ut_ad(buf_page_get_flush_type(bpage) == flush_type);
}

/** Map a block to a file page.
@param[in,out]	block	pointer to control block
@param[in]	page_id	page id */
UNIV_INLINE
void
buf_block_set_file_page(
	buf_block_t*		block,
	const page_id_t&	page_id)
{
	buf_block_set_state(block, BUF_BLOCK_FILE_PAGE);
	block->page.id.copy_from(page_id);
}

/*********************************************************************//**
Gets the io_fix state of a block.
@return io_fix state */
UNIV_INLINE
enum buf_io_fix
buf_page_get_io_fix(
/*================*/
	const buf_page_t*	bpage)	/*!< in: pointer to the control block */
{
	ut_ad(bpage != NULL);

	enum buf_io_fix	io_fix	= bpage->io_fix;

#ifdef UNIV_DEBUG
	switch (io_fix) {
	case BUF_IO_NONE:
	case BUF_IO_READ:
	case BUF_IO_WRITE:
	case BUF_IO_PIN:
		return(io_fix);
	}
	ut_error;
#endif /* UNIV_DEBUG */
	return(io_fix);
}

/*********************************************************************//**
Gets the io_fix state of a block.
@return io_fix state */
UNIV_INLINE
enum buf_io_fix
buf_block_get_io_fix(
/*=================*/
	const buf_block_t*	block)	/*!< in: pointer to the control block */
{
	return(buf_page_get_io_fix(&block->page));
}

/*********************************************************************//**
Sets the io_fix state of a block. */
UNIV_INLINE
void
buf_page_set_io_fix(
/*================*/
	buf_page_t*	bpage,	/*!< in/out: control block */
	enum buf_io_fix	io_fix)	/*!< in: io_fix state */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(buf_pool_mutex_own(buf_pool));
#endif /* UNIV_DEBUG */
	ut_ad(mutex_own(buf_page_get_mutex(bpage)));

	bpage->io_fix = io_fix;
	ut_ad(buf_page_get_io_fix(bpage) == io_fix);
}

/*********************************************************************//**
Sets the io_fix state of a block. */
UNIV_INLINE
void
buf_block_set_io_fix(
/*=================*/
	buf_block_t*	block,	/*!< in/out: control block */
	enum buf_io_fix	io_fix)	/*!< in: io_fix state */
{
	buf_page_set_io_fix(&block->page, io_fix);
}

/*********************************************************************//**
Makes a block sticky. A sticky block implies that even after we release
the buf_pool->mutex and the block->mutex:
* it cannot be removed from the flush_list
* the block descriptor cannot be relocated
* it cannot be removed from the LRU list
Note that:
* the block can still change its position in the LRU list
* the next and previous pointers can change. */
UNIV_INLINE
void
buf_page_set_sticky(
/*================*/
	buf_page_t*	bpage)	/*!< in/out: control block */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(buf_pool_mutex_own(buf_pool));
#endif /* UNIV_DEBUG */
	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
	ut_ad(buf_page_get_io_fix(bpage) == BUF_IO_NONE);

	bpage->io_fix = BUF_IO_PIN;
}

/*********************************************************************//**
Removes stickiness of a block. */
UNIV_INLINE
void
buf_page_unset_sticky(
/*==================*/
	buf_page_t*	bpage)	/*!< in/out: control block */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(buf_pool_mutex_own(buf_pool));
#endif /* UNIV_DEBUG */
	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
	ut_ad(buf_page_get_io_fix(bpage) == BUF_IO_PIN);

	bpage->io_fix = BUF_IO_NONE;
}

/********************************************************************//**
Determine if a buffer block can be relocated in memory.  The block
can be dirty, but it must not be I/O-fixed or bufferfixed. */
UNIV_INLINE
ibool
buf_page_can_relocate(
/*==================*/
	const buf_page_t*	bpage)	/*!< control block being relocated */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(buf_pool_mutex_own(buf_pool));
#endif /* UNIV_DEBUG */
	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
	ut_ad(buf_page_in_file(bpage));
	ut_ad(bpage->in_LRU_list);

	return(buf_page_get_io_fix(bpage) == BUF_IO_NONE
	       && bpage->buf_fix_count == 0);
}

/*********************************************************************//**
Determine if a block has been flagged old.
@return TRUE if old */
UNIV_INLINE
ibool
buf_page_is_old(
/*============*/
	const buf_page_t*	bpage)	/*!< in: control block */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(buf_pool_mutex_own(buf_pool));
#endif /* UNIV_DEBUG */
	ut_ad(buf_page_in_file(bpage));

	return(bpage->old);
}

/*********************************************************************//**
Flag a block old. */
UNIV_INLINE
void
buf_page_set_old(
/*=============*/
	buf_page_t*	bpage,	/*!< in/out: control block */
	ibool		old)	/*!< in: old */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
#endif /* UNIV_DEBUG */
	ut_a(buf_page_in_file(bpage));
	ut_ad(buf_pool_mutex_own(buf_pool));
	ut_ad(bpage->in_LRU_list);

#ifdef UNIV_LRU_DEBUG
	ut_a((buf_pool->LRU_old_len == 0) == (buf_pool->LRU_old == NULL));
	/* If a block is flagged "old", the LRU_old list must exist. */
	ut_a(!old || buf_pool->LRU_old);

	if (UT_LIST_GET_PREV(LRU, bpage) && UT_LIST_GET_NEXT(LRU, bpage)) {
		const buf_page_t*	prev = UT_LIST_GET_PREV(LRU, bpage);
		const buf_page_t*	next = UT_LIST_GET_NEXT(LRU, bpage);
		if (prev->old == next->old) {
			ut_a(prev->old == old);
		} else {
			ut_a(!prev->old);
			ut_a(buf_pool->LRU_old == (old ? bpage : next));
		}
	}
#endif /* UNIV_LRU_DEBUG */

	bpage->old = old;
}

/*********************************************************************//**
Determine the time of first access of a block in the buffer pool.
@return ut_time_monotonic_ms() at the time of first access, 0 if not accessed
*/
UNIV_INLINE
unsigned
buf_page_is_accessed(
/*=================*/
	const buf_page_t*	bpage)	/*!< in: control block */
{
	ut_ad(buf_page_in_file(bpage));

	return(bpage->access_time);
}

/*********************************************************************//**
Flag a block accessed. */
UNIV_INLINE
void
buf_page_set_accessed(
/*==================*/
	buf_page_t*	bpage)		/*!< in/out: control block */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage(bpage);
	ut_ad(!buf_pool_mutex_own(buf_pool));
	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
#endif /* UNIV_DEBUG */

	ut_a(buf_page_in_file(bpage));

	if (bpage->access_time == 0) {
		/* Make this the time of the first access. */
		bpage->access_time = static_cast<uint>(ut_time_monotonic_ms());
	}
}

/*********************************************************************//**
Gets the buf_block_t handle of a buffered file block if an uncompressed
page frame exists, or NULL.
@return control block, or NULL */
UNIV_INLINE
buf_block_t*
buf_page_get_block(
/*===============*/
	buf_page_t*	bpage)	/*!< in: control block, or NULL */
{
	if (bpage != NULL) {
		ut_ad(buf_page_in_file(bpage));

		if (buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE) {
			return((buf_block_t*) bpage);
		}
	}

	return(NULL);
}
#endif /* !UNIV_HOTBACKUP */

#ifdef UNIV_DEBUG
/*********************************************************************//**
Gets a pointer to the memory frame of a block.
@return pointer to the frame */
UNIV_INLINE
buf_frame_t*
buf_block_get_frame(
/*================*/
	const buf_block_t*	block)	/*!< in: pointer to the control block */
{
	ut_ad(block);

	switch (buf_block_get_state(block)) {
	case BUF_BLOCK_POOL_WATCH:
	case BUF_BLOCK_ZIP_PAGE:
	case BUF_BLOCK_ZIP_DIRTY:
	case BUF_BLOCK_NOT_USED:
		ut_error;
		break;
	case BUF_BLOCK_FILE_PAGE:
# ifndef UNIV_HOTBACKUP
		ut_a(block->page.buf_fix_count > 0);
# endif /* !UNIV_HOTBACKUP */
		/* fall through */
	case BUF_BLOCK_READY_FOR_USE:
	case BUF_BLOCK_MEMORY:
	case BUF_BLOCK_REMOVE_HASH:
		goto ok;
	}
	ut_error;
ok:
	return((buf_frame_t*) block->frame);
}
#endif /* UNIV_DEBUG */

/***********************************************************************
FIXME_FTS Gets the frame the pointer is pointing to. */
UNIV_INLINE
buf_frame_t*
buf_frame_align(
/*============*/
                        /* out: pointer to frame */
        byte*   ptr)    /* in: pointer to a frame */
{
        buf_frame_t*    frame;

        ut_ad(ptr);

        frame = (buf_frame_t*) ut_align_down(ptr, UNIV_PAGE_SIZE);

        return(frame);
}

/**********************************************************************//**
Gets the space id, page offset, and byte offset within page of a
pointer pointing to a buffer frame containing a file page. */
UNIV_INLINE
void
buf_ptr_get_fsp_addr(
/*=================*/
	const void*	ptr,	/*!< in: pointer to a buffer frame */
	ulint*		space,	/*!< out: space id */
	fil_addr_t*	addr)	/*!< out: page offset and byte offset */
{
	const page_t*	page = (const page_t*) ut_align_down(ptr,
							     UNIV_PAGE_SIZE);

	*space = mach_read_from_4(page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID);
	addr->page = mach_read_from_4(page + FIL_PAGE_OFFSET);
	addr->boffset = ut_align_offset(ptr, UNIV_PAGE_SIZE);
}

#ifndef UNIV_HOTBACKUP
/**********************************************************************//**
Gets the hash value of the page the pointer is pointing to. This can be used
in searches in the lock hash table.
@return lock hash value */
UNIV_INLINE
ulint
buf_block_get_lock_hash_val(
/*========================*/
	const buf_block_t*	block)	/*!< in: block */
{
	ut_ad(block);
	ut_ad(buf_page_in_file(&block->page));
	ut_ad(rw_lock_own(&(((buf_block_t*) block)->lock), RW_LOCK_X)
	      || rw_lock_own(&(((buf_block_t*) block)->lock), RW_LOCK_S));

	return(block->lock_hash_val);
}

/********************************************************************//**
Allocates a buf_page_t descriptor. This function must succeed. In case
of failure we assert in this function.
@return: the allocated descriptor. */
UNIV_INLINE
buf_page_t*
buf_page_alloc_descriptor(void)
/*===========================*/
{
	buf_page_t*	bpage;

	bpage = (buf_page_t*) ut_zalloc_nokey(sizeof *bpage);
	ut_ad(bpage);
	UNIV_MEM_ALLOC(bpage, sizeof *bpage);

	return(bpage);
}

/********************************************************************//**
Free a buf_page_t descriptor. */
UNIV_INLINE
void
buf_page_free_descriptor(
/*=====================*/
	buf_page_t*	bpage)	/*!< in: bpage descriptor to free. */
{
	ut_free(bpage);
}

/********************************************************************//**
Frees a buffer block which does not contain a file page. */
UNIV_INLINE
void
buf_block_free(
/*===========*/
	buf_block_t*	block)	/*!< in, own: block to be freed */
{
	buf_pool_t*	buf_pool = buf_pool_from_bpage((buf_page_t*) block);

	buf_pool_mutex_enter(buf_pool);

	buf_page_mutex_enter(block);

	ut_a(buf_block_get_state(block) != BUF_BLOCK_FILE_PAGE);

	buf_LRU_block_free_non_file_page(block);

	buf_page_mutex_exit(block);

	buf_pool_mutex_exit(buf_pool);
}
#endif /* !UNIV_HOTBACKUP */

/*********************************************************************//**
Copies contents of a buffer frame to a given buffer.
@return buf */
UNIV_INLINE
byte*
buf_frame_copy(
/*===========*/
	byte*			buf,	/*!< in: buffer to copy to */
	const buf_frame_t*	frame)	/*!< in: buffer frame */
{
	ut_ad(buf && frame);

	ut_memcpy(buf, frame, UNIV_PAGE_SIZE);

	return(buf);
}

#ifndef UNIV_HOTBACKUP
/********************************************************************//**
Gets the youngest modification log sequence number for a frame.
Returns zero if not file page or no modification occurred yet.
@return newest modification to page */
UNIV_INLINE
lsn_t
buf_page_get_newest_modification(
/*=============================*/
	const buf_page_t*	bpage)	/*!< in: block containing the
					page frame */
{
	lsn_t		lsn;
	BPageMutex*	block_mutex = buf_page_get_mutex(bpage);

	mutex_enter(block_mutex);

	if (buf_page_in_file(bpage)) {
		lsn = bpage->newest_modification;
	} else {
		lsn = 0;
	}

	mutex_exit(block_mutex);

	return(lsn);
}

/********************************************************************//**
Increments the modify clock of a frame by 1. The caller must (1) own the
buf_pool mutex and block bufferfix count has to be zero, (2) or own an x-lock
on the block. */
UNIV_INLINE
void
buf_block_modify_clock_inc(
/*=======================*/
	buf_block_t*	block)	/*!< in: block */
{
#ifdef UNIV_DEBUG
	buf_pool_t*	buf_pool = buf_pool_from_bpage((buf_page_t*) block);

	/* No latch is acquired if block belongs to intrinsic table. */
	if (!fsp_is_system_temporary(block->page.id.space())) {
		ut_ad((buf_pool_mutex_own(buf_pool)
		       && (block->page.buf_fix_count == 0))
		      || rw_lock_own_flagged(&block->lock,
					     RW_LOCK_FLAG_X | RW_LOCK_FLAG_SX));
	}
#endif /* UNIV_DEBUG */
	assert_block_ahi_valid(block);

	block->modify_clock++;
}

/********************************************************************//**
Returns the value of the modify clock. The caller must have an s-lock
or x-lock on the block.
@return value */
UNIV_INLINE
ib_uint64_t
buf_block_get_modify_clock(
/*=======================*/
	buf_block_t*	block)	/*!< in: block */
{
#ifdef UNIV_DEBUG
	/* No latch is acquired if block belongs to intrinsic table. */
	if (!fsp_is_system_temporary(block->page.id.space())) {
		ut_ad(rw_lock_own(&(block->lock), RW_LOCK_S)
		      || rw_lock_own(&(block->lock), RW_LOCK_X)
		      || rw_lock_own(&(block->lock), RW_LOCK_SX));
	}
#endif /* UNIV_DEBUG */

	return(block->modify_clock);
}

/** Increments the bufferfix count.
@param[in,out]	bpage	block to bufferfix
@return the count */
UNIV_INLINE
ulint
buf_block_fix(
	buf_page_t*	bpage)
{
	return(os_atomic_increment_uint32(&bpage->buf_fix_count, 1));
}

/** Increments the bufferfix count.
@param[in,out]	block	block to bufferfix
@return the count */
UNIV_INLINE
ulint
buf_block_fix(
	buf_block_t*	block)
{
	return(buf_block_fix(&block->page));
}

/*******************************************************************//**
Increments the bufferfix count. */
UNIV_INLINE
void
buf_block_buf_fix_inc_func(
/*=======================*/
#ifdef UNIV_DEBUG
	const char*	file,	/*!< in: file name */
	ulint		line,	/*!< in: line */
#endif /* UNIV_DEBUG */
	buf_block_t*	block)	/*!< in/out: block to bufferfix */
{
#ifdef UNIV_DEBUG
	/* No debug latch is acquired if block belongs to system temporary.
	Debug latch is not of much help if access to block is single
	threaded. */
	if (!fsp_is_system_temporary(block->page.id.space())) {
		ibool   ret;
		ret = rw_lock_s_lock_nowait(&block->debug_latch, file, line);
		ut_a(ret);
	}
#endif /* UNIV_DEBUG */

	buf_block_fix(block);
}

/** Decrements the bufferfix count.
@param[in,out]	bpage	block to bufferunfix
@return	the remaining buffer-fix count */
UNIV_INLINE
ulint
buf_block_unfix(
	buf_page_t*	bpage)
{
	ulint	count = os_atomic_decrement_uint32(&bpage->buf_fix_count, 1);
	ut_ad(count + 1 != 0);
	return(count);
}

/** Decrements the bufferfix count.
@param[in,out]	block	block to bufferunfix
@return the remaining buffer-fix count */
UNIV_INLINE
ulint
buf_block_unfix(
	buf_block_t*	block)
{
	return(buf_block_unfix(&block->page));
}

/*******************************************************************//**
Decrements the bufferfix count. */
UNIV_INLINE
void
buf_block_buf_fix_dec(
/*==================*/
	buf_block_t*	block)	/*!< in/out: block to bufferunfix */
{
#ifdef UNIV_DEBUG
	/* No debug latch is acquired if block belongs to system temporary.
	Debug latch is not of much help if access to block is single
	threaded. */
	if (!fsp_is_system_temporary(block->page.id.space())) {
		rw_lock_s_unlock(&block->debug_latch);
	}
#endif /* UNIV_DEBUG */

	buf_block_unfix(block);
}

/** Returns the buffer pool instance given a page id.
@param[in]	page_id	page id
@return buffer pool */
UNIV_INLINE
buf_pool_t*
buf_pool_get(
	const page_id_t&	page_id)
{
        /* 2log of BUF_READ_AHEAD_AREA (64) */
        ulint		ignored_page_no = page_id.page_no() >> 6;

        page_id_t	id(page_id.space(), ignored_page_no);

        ulint		i = id.fold() % srv_buf_pool_instances;

        return(&buf_pool_ptr[i]);
}

/******************************************************************//**
Returns the buffer pool instance given its array index
@return buffer pool */
UNIV_INLINE
buf_pool_t*
buf_pool_from_array(
/*================*/
	ulint	index)		/*!< in: array index to get
				buffer pool instance from */
{
	ut_ad(index < MAX_BUFFER_POOLS);
	ut_ad(index < srv_buf_pool_instances);
	return(&buf_pool_ptr[index]);
}

/** Returns the control block of a file page, NULL if not found.
@param[in]	buf_pool	buffer pool instance
@param[in]	page_id		page id
@return block, NULL if not found */
UNIV_INLINE
buf_page_t*
buf_page_hash_get_low(
	buf_pool_t*		buf_pool,
	const page_id_t&	page_id)
{
	buf_page_t*	bpage;

#ifdef UNIV_DEBUG
	rw_lock_t*	hash_lock;

	hash_lock = hash_get_lock(buf_pool->page_hash, page_id.fold());
	ut_ad(rw_lock_own(hash_lock, RW_LOCK_X)
	      || rw_lock_own(hash_lock, RW_LOCK_S));
#endif /* UNIV_DEBUG */

	/* Look for the page in the hash table */

	HASH_SEARCH(hash, buf_pool->page_hash, page_id.fold(), buf_page_t*,
		    bpage,
		    ut_ad(bpage->in_page_hash && !bpage->in_zip_hash
			  && buf_page_in_file(bpage)),
		    page_id.equals_to(bpage->id));
	if (bpage) {
		ut_a(buf_page_in_file(bpage));
		ut_ad(bpage->in_page_hash);
		ut_ad(!bpage->in_zip_hash);
		ut_ad(buf_pool_from_bpage(bpage) == buf_pool);
	}

	return(bpage);
}

/** Returns the control block of a file page, NULL if not found.
If the block is found and lock is not NULL then the appropriate
page_hash lock is acquired in the specified lock mode. Otherwise,
mode value is ignored. It is up to the caller to release the
lock. If the block is found and the lock is NULL then the page_hash
lock is released by this function.
@param[in]	buf_pool	buffer pool instance
@param[in]	page_id		page id
@param[in,out]	lock		lock of the page hash acquired if bpage is
found, NULL otherwise. If NULL is passed then the hash_lock is released by
this function.
@param[in]	lock_mode	RW_LOCK_X or RW_LOCK_S. Ignored if
lock == NULL
@param[in]	watch		if true, return watch sentinel also.
@return pointer to the bpage or NULL; if NULL, lock is also NULL or
a watch sentinel. */
UNIV_INLINE
buf_page_t*
buf_page_hash_get_locked(
	buf_pool_t*		buf_pool,
	const page_id_t&	page_id,
	rw_lock_t**		lock,
	ulint			lock_mode,
	bool			watch)
{
	buf_page_t*	bpage = NULL;
	rw_lock_t*	hash_lock;
	ulint		mode = RW_LOCK_S;

	if (lock != NULL) {
		*lock = NULL;
		ut_ad(lock_mode == RW_LOCK_X
		      || lock_mode == RW_LOCK_S);
		mode = lock_mode;
	}

	hash_lock = hash_get_lock(buf_pool->page_hash, page_id.fold());

	ut_ad(!rw_lock_own(hash_lock, RW_LOCK_X)
	      && !rw_lock_own(hash_lock, RW_LOCK_S));

	if (mode == RW_LOCK_S) {
		rw_lock_s_lock(hash_lock);

		/* If not own buf_pool_mutex, page_hash can be changed. */
		hash_lock = hash_lock_s_confirm(
			hash_lock, buf_pool->page_hash, page_id.fold());
	} else {
		rw_lock_x_lock(hash_lock);
		/* If not own buf_pool_mutex, page_hash can be changed. */
		hash_lock = hash_lock_x_confirm(
			hash_lock, buf_pool->page_hash, page_id.fold());
	}

	bpage = buf_page_hash_get_low(buf_pool, page_id);

	if (!bpage || buf_pool_watch_is_sentinel(buf_pool, bpage)) {
		if (!watch) {
			bpage = NULL;
		}
		goto unlock_and_exit;
	}

	ut_ad(buf_page_in_file(bpage));
	ut_ad(page_id.equals_to(bpage->id));

	if (lock == NULL) {
		/* The caller wants us to release the page_hash lock */
		goto unlock_and_exit;
	} else {
		/* To be released by the caller */
		*lock = hash_lock;
		goto exit;
	}

unlock_and_exit:
	if (mode == RW_LOCK_S) {
		rw_lock_s_unlock(hash_lock);
	} else {
		rw_lock_x_unlock(hash_lock);
	}
exit:
	return(bpage);
}

/** Returns the control block of a file page, NULL if not found.
If the block is found and lock is not NULL then the appropriate
page_hash lock is acquired in the specified lock mode. Otherwise,
mode value is ignored. It is up to the caller to release the
lock. If the block is found and the lock is NULL then the page_hash
lock is released by this function.
@param[in]	buf_pool	buffer pool instance
@param[in]	page_id		page id
@param[in,out]	lock		lock of the page hash acquired if bpage is
found, NULL otherwise. If NULL is passed then the hash_lock is released by
this function.
@param[in]	lock_mode	RW_LOCK_X or RW_LOCK_S. Ignored if
lock == NULL
@return pointer to the block or NULL; if NULL, lock is also NULL. */
UNIV_INLINE
buf_block_t*
buf_block_hash_get_locked(
	buf_pool_t*		buf_pool,
	const page_id_t&	page_id,
	rw_lock_t**		lock,
	ulint			lock_mode)
{
	buf_page_t*	bpage = buf_page_hash_get_locked(buf_pool,
							 page_id,
							 lock,
							 lock_mode);
	buf_block_t*	block = buf_page_get_block(bpage);

	if (block != NULL) {

		ut_ad(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
		ut_ad(!lock || rw_lock_own(*lock, lock_mode));

		return(block);
	} else if (bpage) {
		/* It is not a block. Just a bpage */
		ut_ad(buf_page_in_file(bpage));

		if (lock) {
			if (lock_mode == RW_LOCK_S) {
				rw_lock_s_unlock(*lock);
			} else {
				rw_lock_x_unlock(*lock);
			}
		}
		*lock = NULL;
		return(NULL);
	}

	ut_ad(!bpage);
	ut_ad(lock == NULL ||*lock == NULL);
	return(NULL);
}

/** Returns TRUE if the page can be found in the buffer pool hash table.
NOTE that it is possible that the page is not yet read from disk,
though.
@param[in]	page_id	page id
@return TRUE if found in the page hash table */
UNIV_INLINE
ibool
buf_page_peek(
	const page_id_t&	page_id)
{
	buf_pool_t*	buf_pool = buf_pool_get(page_id);

	return(buf_page_hash_get(buf_pool, page_id) != NULL);
}

/********************************************************************//**
Releases a compressed-only page acquired with buf_page_get_zip(). */
UNIV_INLINE
void
buf_page_release_zip(
/*=================*/
	buf_page_t*	bpage)		/*!< in: buffer block */
{
	ut_ad(bpage);
	ut_a(bpage->buf_fix_count > 0);

	switch (buf_page_get_state(bpage)) {
	case BUF_BLOCK_FILE_PAGE:
#ifdef UNIV_DEBUG
	{
		/* No debug latch is acquired if block belongs to system
		temporary. Debug latch is not of much help if access to block
		is single threaded. */
		buf_block_t*	block = reinterpret_cast<buf_block_t*>(bpage);
		if (!fsp_is_system_temporary(block->page.id.space())) {
			rw_lock_s_unlock(&block->debug_latch);
		}
	}
#endif /* UNIV_DEBUG */
		/* Fall through */

	case BUF_BLOCK_ZIP_PAGE:
	case BUF_BLOCK_ZIP_DIRTY:
		buf_block_unfix(reinterpret_cast<buf_block_t*>(bpage));
		return;

	case BUF_BLOCK_POOL_WATCH:
	case BUF_BLOCK_NOT_USED:
	case BUF_BLOCK_READY_FOR_USE:
	case BUF_BLOCK_MEMORY:
	case BUF_BLOCK_REMOVE_HASH:
		break;
	}

	ut_error;
}

/********************************************************************//**
Releases a latch, if specified. */
UNIV_INLINE
void
buf_page_release_latch(
/*===================*/
	buf_block_t*	block,		/*!< in: buffer block */
	ulint		rw_latch)	/*!< in: RW_S_LATCH, RW_X_LATCH,
					RW_NO_LATCH */
{
#ifdef UNIV_DEBUG
	/* No debug latch is acquired if block belongs to system
	temporary. Debug latch is not of much help if access to block
	is single threaded. */
	if (!fsp_is_system_temporary(block->page.id.space())) {
		rw_lock_s_unlock(&block->debug_latch);
	}
#endif /* UNIV_DEBUG */

	if (rw_latch == RW_S_LATCH) {
		rw_lock_s_unlock(&block->lock);
	} else if (rw_latch == RW_SX_LATCH) {
		rw_lock_sx_unlock(&block->lock);
	} else if (rw_latch == RW_X_LATCH) {
		rw_lock_x_unlock(&block->lock);
	}
}

#ifdef UNIV_DEBUG
/*********************************************************************//**
Adds latch level info for the rw-lock protecting the buffer frame. This
should be called in the debug version after a successful latching of a
page if we know the latching order level of the acquired latch. */
UNIV_INLINE
void
buf_block_dbg_add_level(
/*====================*/
	buf_block_t*	block,	/*!< in: buffer page
				where we have acquired latch */
	latch_level_t	level)	/*!< in: latching order level */
{
	sync_check_lock(&block->lock, level);
}

#endif /* UNIV_DEBUG */
/********************************************************************//**
Acquire mutex on all buffer pool instances. */
UNIV_INLINE
void
buf_pool_mutex_enter_all(void)
/*==========================*/
{
	for (ulint i = 0; i < srv_buf_pool_instances; ++i) {
		buf_pool_t*	buf_pool = buf_pool_from_array(i);

		buf_pool_mutex_enter(buf_pool);
	}
}

/********************************************************************//**
Release mutex on all buffer pool instances. */
UNIV_INLINE
void
buf_pool_mutex_exit_all(void)
/*=========================*/
{
	ulint   i;

	for (i = 0; i < srv_buf_pool_instances; i++) {
		buf_pool_t*	buf_pool;

		buf_pool = buf_pool_from_array(i);
		buf_pool_mutex_exit(buf_pool);
	}
}
/*********************************************************************//**
Get the nth chunk's buffer block in the specified buffer pool.
@return the nth chunk's buffer block. */
UNIV_INLINE
buf_block_t*
buf_get_nth_chunk_block(
/*====================*/
	const buf_pool_t* buf_pool,	/*!< in: buffer pool instance */
	ulint		n,		/*!< in: nth chunk in the buffer pool */
	ulint*		chunk_size)	/*!< in: chunk size */
{
	const buf_chunk_t*	chunk;

	chunk = buf_pool->chunks + n;
	*chunk_size = chunk->size;
	return(chunk->blocks);
}

/** Calculate aligned buffer pool size based on srv_buf_pool_chunk_unit,
if needed.
@param[in]	size	size in bytes
@return	aligned size */
UNIV_INLINE
ulint
buf_pool_size_align(
	ulint	size)
{
	const ulint	m = srv_buf_pool_instances * srv_buf_pool_chunk_unit;
	size = ut_max(size, srv_buf_pool_min_size);

	if (size % m == 0) {
		return(size);
	} else {
		return((size / m + 1) * m);
	}
}

#endif /* !UNIV_HOTBACKUP */