page0page.c

这是linux下运行的mysql软件包,可用于linux 下安装 php + mysql + apach 的网络配置

/******************************************************
Index page routines

(c) 1994-1996 Innobase Oy

Created 2/2/1994 Heikki Tuuri
*******************************************************/

#define THIS_MODULE
#include "page0page.h"
#ifdef UNIV_NONINL
#include "page0page.ic"
#endif
#undef THIS_MODULE

#include "page0cur.h"
#include "lock0lock.h"
#include "fut0lst.h"
#include "btr0sea.h"
#include "buf0buf.h"
#include "srv0srv.h"
#include "btr0btr.h"

/*			THE INDEX PAGE
			==============
		
The index page consists of a page header which contains the page's
id and other information. On top of it are the the index records
in a heap linked into a one way linear list according to alphabetic order.

Just below page end is an array of pointers which we call page directory,
to about every sixth record in the list. The pointers are placed in
the directory in the alphabetical order of the records pointed to,
enabling us to make binary search using the array. Each slot n:o I
in the directory points to a record, where a 4-bit field contains a count
of those records which are in the linear list between pointer I and 
the pointer I - 1 in the directory, including the record
pointed to by pointer I and not including the record pointed to by I - 1.
We say that the record pointed to by slot I, or that slot I, owns
these records. The count is always kept in the range 4 to 8, with
the exception that it is 1 for the first slot, and 1--8 for the second slot.  
		
An essentially binary search can be performed in the list of index
records, like we could do if we had pointer to every record in the
page directory. The data structure is, however, more efficient when
we are doing inserts, because most inserts are just pushed on a heap.
Only every 8th insert requires block move in the directory pointer
table, which itself is quite small. A record is deleted from the page
by just taking it off the linear list and updating the number of owned
records-field of the record which owns it, and updating the page directory,
if necessary. A special case is the one when the record owns itself.
Because the overhead of inserts is so small, we may also increase the
page size from the projected default of 8 kB to 64 kB without too
much loss of efficiency in inserts. Bigger page becomes actual
when the disk transfer rate compared to seek and latency time rises.
On the present system, the page size is set so that the page transfer
time (3 ms) is 20 % of the disk random access time (15 ms).

When the page is split, merged, or becomes full but contains deleted
records, we have to reorganize the page.

Assuming a page size of 8 kB, a typical index page of a secondary
index contains 300 index entries, and the size of the page directory
is 50 x 4 bytes = 200 bytes. */

/*******************************************************************
Looks for the directory slot which owns the given record. */

ulint
page_dir_find_owner_slot(
/*=====================*/
			/* out: the directory slot number */
	rec_t*	rec)	/* in: the physical record */
{
	page_t*				page;
	register uint16			rec_offs_bytes;
	register page_dir_slot_t*	slot;
	register const page_dir_slot_t*	first_slot;
	register rec_t*			r = rec;

	ut_ad(page_rec_check(rec));

	page = buf_frame_align(rec);
	first_slot = page_dir_get_nth_slot(page, 0);
	slot = page_dir_get_nth_slot(page, page_dir_get_n_slots(page) - 1);

	if (page_is_comp(page)) {
		while (rec_get_n_owned(r, TRUE) == 0) {
			r = page + rec_get_next_offs(r, TRUE);
			ut_ad(r >= page + PAGE_NEW_SUPREMUM);
			ut_ad(r < page + (UNIV_PAGE_SIZE - PAGE_DIR));
		}
	} else {
		while (rec_get_n_owned(r, FALSE) == 0) {
			r = page + rec_get_next_offs(r, FALSE);
			ut_ad(r >= page + PAGE_OLD_SUPREMUM);
			ut_ad(r < page + (UNIV_PAGE_SIZE - PAGE_DIR));
		}
	}

	rec_offs_bytes = mach_encode_2(r - page);

	while (UNIV_LIKELY(*(uint16*) slot != rec_offs_bytes)) {

		if (UNIV_UNLIKELY(slot == first_slot)) {
			fprintf(stderr,
			"InnoDB: Probable data corruption on page %lu\n"
			"InnoDB: Original record ",
			(ulong) buf_frame_get_page_no(page));

			if (page_is_comp(page)) {
				fputs("(compact record)", stderr);
			} else {
				rec_print_old(stderr, rec);
			}

			fputs("\n"
			"InnoDB: on that page.\n"
			"InnoDB: Cannot find the dir slot for record ",
				stderr);
			if (page_is_comp(page)) {
				fputs("(compact record)", stderr);
			} else {
				rec_print_old(stderr, page
					+ mach_decode_2(rec_offs_bytes));
			}
			fputs("\n"
			"InnoDB: on that page!\n", stderr);

			buf_page_print(page);

			ut_error;
		}

		slot += PAGE_DIR_SLOT_SIZE;
	}

	return(((ulint) (first_slot - slot)) / PAGE_DIR_SLOT_SIZE);
}

/******************************************************************
Used to check the consistency of a directory slot. */
static
ibool
page_dir_slot_check(
/*================*/
					/* out: TRUE if succeed */
	page_dir_slot_t*	slot)	/* in: slot */
{
	page_t*	page;
	ulint	n_slots;
	ulint	n_owned;

	ut_a(slot);

	page = buf_frame_align(slot);

	n_slots = page_dir_get_n_slots(page);

	ut_a(slot <= page_dir_get_nth_slot(page, 0));
	ut_a(slot >= page_dir_get_nth_slot(page, n_slots - 1));

	ut_a(page_rec_check(page_dir_slot_get_rec(slot)));

	n_owned = rec_get_n_owned(page_dir_slot_get_rec(slot),
					page_is_comp(page));

	if (slot == page_dir_get_nth_slot(page, 0)) {
		ut_a(n_owned == 1);
	} else if (slot == page_dir_get_nth_slot(page, n_slots - 1)) {
		ut_a(n_owned >= 1);
		ut_a(n_owned <= PAGE_DIR_SLOT_MAX_N_OWNED);
	} else {
		ut_a(n_owned >= PAGE_DIR_SLOT_MIN_N_OWNED);
		ut_a(n_owned <= PAGE_DIR_SLOT_MAX_N_OWNED);
	}

	return(TRUE);
}

/*****************************************************************
Sets the max trx id field value. */

void
page_set_max_trx_id(
/*================*/
	page_t*	page,	/* in: page */
	dulint	trx_id)	/* in: transaction id */
{
	buf_block_t*	block;

	ut_ad(page);

	block = buf_block_align(page);

	if (block->is_hashed) {
		rw_lock_x_lock(&btr_search_latch);
	}

	/* It is not necessary to write this change to the redo log, as
	during a database recovery we assume that the max trx id of every
	page is the maximum trx id assigned before the crash. */
	
	mach_write_to_8(page + PAGE_HEADER + PAGE_MAX_TRX_ID, trx_id);

	if (block->is_hashed) {
		rw_lock_x_unlock(&btr_search_latch);
	}
}

/****************************************************************
Allocates a block of memory from an index page. */

byte*
page_mem_alloc(
/*===========*/
				/* out: pointer to start of allocated
				buffer, or NULL if allocation fails */
	page_t*		page,	/* in: index page */
	ulint		need,	/* in: number of bytes needed */
	dict_index_t*	index,	/* in: record descriptor */
	ulint*		heap_no)/* out: this contains the heap number
				of the allocated record
				if allocation succeeds */
{
	rec_t*	rec;
	byte*	block;
	ulint	avl_space;
	ulint	garbage;
	
	ut_ad(page && heap_no);

	/* If there are records in the free list, look if the first is
	big enough */

	rec = page_header_get_ptr(page, PAGE_FREE);

	if (rec) {
		mem_heap_t*	heap		= NULL;
		ulint		offsets_[REC_OFFS_NORMAL_SIZE];
		ulint*		offsets		= offsets_;
		*offsets_ = (sizeof offsets_) / sizeof *offsets_;

		offsets = rec_get_offsets(rec, index, offsets,
						ULINT_UNDEFINED, &heap);

		if (rec_offs_size(offsets) >= need) {
			page_header_set_ptr(page, PAGE_FREE,
				page_rec_get_next(rec));

			garbage = page_header_get_field(page, PAGE_GARBAGE);
			ut_ad(garbage >= need);

			page_header_set_field(page, PAGE_GARBAGE,
							garbage - need);

			*heap_no = rec_get_heap_no(rec, page_is_comp(page));

			block = rec_get_start(rec, offsets);
			if (UNIV_LIKELY_NULL(heap)) {
				mem_heap_free(heap);
			}
			return(block);
		}

		if (UNIV_LIKELY_NULL(heap)) {
			mem_heap_free(heap);
		}
	}

	/* Could not find space from the free list, try top of heap */
	
	avl_space = page_get_max_insert_size(page, 1);
	
	if (avl_space >= need) {
		block = page_header_get_ptr(page, PAGE_HEAP_TOP);

		page_header_set_ptr(page, PAGE_HEAP_TOP, block + need);
		*heap_no = page_dir_get_n_heap(page);

		page_dir_set_n_heap(page, 1 + *heap_no);

		return(block);
	}

	return(NULL);
}

/**************************************************************
Writes a log record of page creation. */
UNIV_INLINE
void
page_create_write_log(
/*==================*/
	buf_frame_t*	frame,	/* in: a buffer frame where the page is
				created */
	mtr_t*		mtr,	/* in: mini-transaction handle */
	ulint		comp)	/* in: nonzero=compact page format */
{
	mlog_write_initial_log_record(frame,
			comp ? MLOG_COMP_PAGE_CREATE : MLOG_PAGE_CREATE, mtr);
}

/***************************************************************
Parses a redo log record of creating a page. */

byte*
page_parse_create(
/*==============*/
			/* out: end of log record or NULL */
	byte*	ptr,	/* in: buffer */
	byte*	end_ptr __attribute__((unused)), /* in: buffer end */
	ulint	comp,	/* in: nonzero=compact page format */
	page_t*	page,	/* in: page or NULL */
	mtr_t*	mtr)	/* in: mtr or NULL */
{
	ut_ad(ptr && end_ptr);

	/* The record is empty, except for the record initial part */

	if (page) {
		page_create(page, mtr, comp);
	}

	return(ptr);
}

/**************************************************************
The index page creation function. */

page_t* 
page_create(
/*========*/
				/* out: pointer to the page */
	buf_frame_t*	frame,	/* in: a buffer frame where the page is
				created */
	mtr_t*		mtr,	/* in: mini-transaction handle */
	ulint		comp)	/* in: nonzero=compact page format */
{
	page_dir_slot_t* slot;
	mem_heap_t*	heap;
	dtuple_t*	tuple;	
	dfield_t*	field;
	byte*		heap_top;
	rec_t*		infimum_rec;
	rec_t*		supremum_rec;
	page_t*		page;
	dict_index_t*	index;
	ulint*		offsets;

	index = comp ? srv_sys->dummy_ind2 : srv_sys->dummy_ind1;
	
	ut_ad(frame && mtr);
	ut_ad(PAGE_BTR_IBUF_FREE_LIST + FLST_BASE_NODE_SIZE
	      <= PAGE_DATA);
	ut_ad(PAGE_BTR_IBUF_FREE_LIST_NODE + FLST_NODE_SIZE
	      <= PAGE_DATA);

	/* 1. INCREMENT MODIFY CLOCK */
	buf_frame_modify_clock_inc(frame);

	/* 2. WRITE LOG INFORMATION */
	page_create_write_log(frame, mtr, comp);
	
	page = frame;

	fil_page_set_type(page, FIL_PAGE_INDEX);

	heap = mem_heap_create(200);
		
	/* 3. CREATE THE INFIMUM AND SUPREMUM RECORDS */

	/* Create first a data tuple for infimum record */
	tuple = dtuple_create(heap, 1);
	dtuple_set_info_bits(tuple, REC_STATUS_INFIMUM);
	field = dtuple_get_nth_field(tuple, 0);

	dfield_set_data(field, "infimum", 8);
	dtype_set(dfield_get_type(field),
			DATA_VARCHAR, DATA_ENGLISH | DATA_NOT_NULL, 8, 0);
	/* Set the corresponding physical record to its place in the page
	record heap */

	heap_top = page + PAGE_DATA;
	
	infimum_rec = rec_convert_dtuple_to_rec(heap_top, index, tuple);

	ut_a(infimum_rec ==
		page + (comp ? PAGE_NEW_INFIMUM : PAGE_OLD_INFIMUM));

	rec_set_n_owned(infimum_rec, comp, 1);
	rec_set_heap_no(infimum_rec, comp, 0);
	offsets = rec_get_offsets(infimum_rec, index, NULL,
						ULINT_UNDEFINED, &heap);

	heap_top = rec_get_end(infimum_rec, offsets);

	/* Create then a tuple for supremum */

	tuple = dtuple_create(heap, 1);
	dtuple_set_info_bits(tuple, REC_STATUS_SUPREMUM);
	field = dtuple_get_nth_field(tuple, 0);

	dfield_set_data(field, "supremum", comp ? 8 : 9);
	dtype_set(dfield_get_type(field),
		DATA_VARCHAR, DATA_ENGLISH | DATA_NOT_NULL, comp ? 8 : 9, 0);

	supremum_rec = rec_convert_dtuple_to_rec(heap_top, index, tuple);

	ut_a(supremum_rec ==
		page + (comp ? PAGE_NEW_SUPREMUM : PAGE_OLD_SUPREMUM));

	rec_set_n_owned(supremum_rec, comp, 1);
	rec_set_heap_no(supremum_rec, comp, 1);

	offsets = rec_get_offsets(supremum_rec, index, offsets,
						ULINT_UNDEFINED, &heap);
	heap_top = rec_get_end(supremum_rec, offsets);

	ut_ad(heap_top ==
		page + (comp ? PAGE_NEW_SUPREMUM_END : PAGE_OLD_SUPREMUM_END));

	mem_heap_free(heap);

	/* 4. INITIALIZE THE PAGE */

	page_header_set_field(page, PAGE_N_DIR_SLOTS, 2);
	page_header_set_ptr(page, PAGE_HEAP_TOP, heap_top);
	page_header_set_field(page, PAGE_N_HEAP, comp ? 0x8002 : 2);
	page_header_set_ptr(page, PAGE_FREE, NULL);
	page_header_set_field(page, PAGE_GARBAGE, 0);
	page_header_set_ptr(page, PAGE_LAST_INSERT, NULL);
	page_header_set_field(page, PAGE_DIRECTION, PAGE_NO_DIRECTION);
	page_header_set_field(page, PAGE_N_DIRECTION, 0);
	page_header_set_field(page, PAGE_N_RECS, 0);
	page_set_max_trx_id(page, ut_dulint_zero);
	memset(heap_top, 0, UNIV_PAGE_SIZE - PAGE_EMPTY_DIR_START
						- (heap_top - page));

	/* 5. SET POINTERS IN RECORDS AND DIR SLOTS */

	/* Set the slots to point to infimum and supremum. */

	slot = page_dir_get_nth_slot(page, 0);
	page_dir_slot_set_rec(slot, infimum_rec);

	slot = page_dir_get_nth_slot(page, 1);
	page_dir_slot_set_rec(slot, supremum_rec);

	/* Set the next pointers in infimum and supremum */
	
	rec_set_next_offs(infimum_rec, comp, (ulint)(supremum_rec - page));
	rec_set_next_offs(supremum_rec, comp, 0);

	return(page);
}

/*****************************************************************
Differs from page_copy_rec_list_end, because this function does not
touch the lock table and max trx id on page. */

void
page_copy_rec_list_end_no_locks(
/*============================*/
	page_t*		new_page,	/* in: index page to copy to */
	page_t*		page,		/* in: index page */
	rec_t*		rec,		/* in: record on page */
	dict_index_t*	index,		/* in: record descriptor */
	mtr_t*		mtr)		/* in: mtr */
{
	page_cur_t	cur1;
	page_cur_t	cur2;
	rec_t*		sup;
	mem_heap_t*	heap		= NULL;
	ulint		offsets_[REC_OFFS_NORMAL_SIZE];
	ulint*		offsets		= offsets_;
	*offsets_ = (sizeof offsets_) / sizeof *offsets_;

	page_cur_position(rec, &cur1);

	if (page_cur_is_before_first(&cur1)) {

		page_cur_move_to_next(&cur1);
	}

	ut_a((ibool)!!page_is_comp(new_page) == index->table->comp);
	ut_a(page_is_comp(new_page) == page_is_comp(page));
	ut_a(mach_read_from_2(new_page + UNIV_PAGE_SIZE - 10) == (ulint)
		(page_is_comp(new_page)
		? PAGE_NEW_INFIMUM : PAGE_OLD_INFIMUM));

	page_cur_set_before_first(new_page, &cur2);
	
	/* Copy records from the original page to the new page */	

	sup = page_get_supremum_rec(page);
	
	for (;;) {
		rec_t*	cur1_rec = page_cur_get_rec(&cur1);
		if (cur1_rec == sup) {