fil0fil.c 87.9 KB
Newer Older
1
/******************************************************
unknown's avatar
unknown committed
2
The tablespace memory cache
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

(c) 1995 Innobase Oy

Created 10/25/1995 Heikki Tuuri
*******************************************************/

#include "fil0fil.h"

#include "mem0mem.h"
#include "sync0sync.h"
#include "hash0hash.h"
#include "os0file.h"
#include "os0sync.h"
#include "mach0data.h"
#include "ibuf0ibuf.h"
#include "buf0buf.h"
unknown's avatar
unknown committed
19 20
#include "buf0flu.h"
#include "buf0lru.h"
21 22 23
#include "log0log.h"
#include "log0recv.h"
#include "fsp0fsp.h"
24
#include "srv0srv.h"
unknown's avatar
unknown committed
25
#include "srv0start.h"
26 27

/*
unknown's avatar
unknown committed
28 29
		IMPLEMENTATION OF THE TABLESPACE MEMORY CACHE
		=============================================
30

unknown's avatar
unknown committed
31
The tablespace cache is responsible for providing fast read/write access to
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
tablespaces and logs of the database. File creation and deletion is done
in other modules which know more of the logic of the operation, however.

A tablespace consists of a chain of files. The size of the files does not
have to be divisible by the database block size, because we may just leave
the last incomplete block unused. When a new file is appended to the
tablespace, the maximum size of the file is also specified. At the moment,
we think that it is best to extend the file to its maximum size already at
the creation of the file, because then we can avoid dynamically extending
the file when more space is needed for the tablespace.

A block's position in the tablespace is specified with a 32-bit unsigned
integer. The files in the chain are thought to be catenated, and the block
corresponding to an address n is the nth block in the catenated file (where
the first block is named the 0th block, and the incomplete block fragments
at the end of files are not taken into account). A tablespace can be extended
by appending a new file at the end of the chain.

Our tablespace concept is similar to the one of Oracle.

To acquire more speed in disk transfers, a technique called disk striping is
sometimes used. This means that logical block addresses are divided in a
round-robin fashion across several disks. Windows NT supports disk striping,
so there we do not need to support it in the database. Disk striping is
implemented in hardware in RAID disks. We conclude that it is not necessary
to implement it in the database. Oracle 7 does not support disk striping,
either.

Another trick used at some database sites is replacing tablespace files by
raw disks, that is, the whole physical disk drive, or a partition of it, is
opened as a single file, and it is accessed through byte offsets calculated
from the start of the disk or the partition. This is recommended in some
books on database tuning to achieve more speed in i/o. Using raw disk
certainly prevents the OS from fragmenting disk space, but it is not clear
if it really adds speed. We measured on the Pentium 100 MHz + NT + NTFS file
system + EIDE Conner disk only a negligible difference in speed when reading
from a file, versus reading from a raw disk. 

To have fast access to a tablespace or a log file, we put the data structures
to a hash table. Each tablespace and log file is given an unique 32-bit
identifier.

Some operating systems do not support many open files at the same time,
though NT seems to tolerate at least 900 open files. Therefore, we put the
open files in an LRU-list. If we need to open another file, we may close the
file at the end of the LRU-list. When an i/o-operation is pending on a file,
the file cannot be closed. We take the file nodes with pending i/o-operations
out of the LRU-list and keep a count of pending operations. When an operation
completes, we decrement the count and return the file node to the LRU-list if
the count drops to zero. */

83 84 85
ulint	fil_n_pending_log_flushes		= 0;
ulint	fil_n_pending_tablespace_flushes	= 0;

86 87 88
/* Null file address */
fil_addr_t	fil_addr_null = {FIL_NULL, 0};

unknown's avatar
unknown committed
89
/* File node of a tablespace or the log data space */
90 91
typedef	struct fil_node_struct	fil_node_t;
struct fil_node_struct {
unknown's avatar
unknown committed
92 93 94
	fil_space_t*	space;	/* backpointer to the space where this node
				belongs */
	char*		name;	/* path to the file */
95 96
	ibool		open;	/* TRUE if file open */
	os_file_t	handle;	/* OS handle to the file, if file open */
unknown's avatar
unknown committed
97 98 99 100 101
	ibool		is_raw_disk;/* TRUE if the 'file' is actually a raw
				device or a raw disk partition */
	ulint		size;	/* size of the file in database pages, 0 if
				not known yet; the possible last incomplete
				megabyte is ignored if space == 0 */
102
	ulint		n_pending;
unknown's avatar
unknown committed
103 104 105 106 107 108 109 110 111 112 113
				/* count of pending i/o's on this file;
				closing of the file is not allowed if
				this is > 0 */
	ulint		n_pending_flushes;
				/* count of pending flushes on this file;
				closing of the file is not allowed if
				this is > 0 */	
	ib_longlong	modification_counter;/* when we write to the file we
				increment this by one */
	ib_longlong	flush_counter;/* up to what modification_counter value
				we have flushed the modifications to disk */
114 115 116 117 118 119 120 121 122
	UT_LIST_NODE_T(fil_node_t) chain;
				/* link field for the file chain */
	UT_LIST_NODE_T(fil_node_t) LRU;
				/* link field for the LRU list */
	ulint		magic_n;
};

#define	FIL_NODE_MAGIC_N	89389

unknown's avatar
unknown committed
123
/* Tablespace or log data space: let us call them by a common name space */
124
struct fil_space_struct {
unknown's avatar
unknown committed
125 126
	char*		name;	/* space name = the path to the first file in
				it */
127
	ulint		id;	/* space id */
unknown's avatar
unknown committed
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
	ib_longlong	tablespace_version;
				/* in DISCARD/IMPORT this timestamp is used to
				check if we should ignore an insert buffer
				merge request for a page because it actually
				was for the previous incarnation of the
				space */
	ibool		mark;	/* this is set to TRUE at database startup if
				the space corresponds to a table in the InnoDB
				data dictionary; so we can print a warning of
				orphaned tablespaces */
	ibool		stop_ios;/* TRUE if we want to rename the .ibd file of
				tablespace and want to stop temporarily
				posting of new i/o requests on the file */
	ibool		stop_ibuf_merges;
				/* we set this TRUE when we start deleting a
				single-table tablespace */
	ibool		is_being_deleted;
				/* this is set to TRUE when we start
				deleting a single-table tablespace and its
				file; when this flag is set no further i/o
				or flush requests can be placed on this space,
				though there may be such requests still being
				processed on this space */
151 152 153
	ulint		purpose;/* FIL_TABLESPACE, FIL_LOG, or FIL_ARCH_LOG */
	UT_LIST_BASE_NODE_T(fil_node_t) chain;
				/* base node for the file chain */
unknown's avatar
unknown committed
154 155
	ulint		size;	/* space size in pages; 0 if a single-table
				tablespace whose size we do not know yet */
156 157 158
	ulint		n_reserved_extents;
				/* number of reserved free extents for
				ongoing operations like B-tree page split */
unknown's avatar
unknown committed
159 160 161 162 163 164 165 166
	ulint		n_pending_flushes; /* this is > 0 when flushing
				the tablespace to disk; dropping of the
				tablespace is forbidden if this is > 0 */
	ulint		n_pending_ibuf_merges;/* this is > 0 when merging
				insert buffer entries to a page so that we
				may need to access the ibuf bitmap page in the
				tablespade: dropping of the tablespace is
				forbidden if this is > 0 */
167
	hash_node_t	hash; 	/* hash chain node */
unknown's avatar
unknown committed
168
	hash_node_t	name_hash;/* hash chain the name_hash table */
169 170 171 172 173 174 175 176 177 178 179
	rw_lock_t	latch;	/* latch protecting the file space storage
				allocation */
	UT_LIST_NODE_T(fil_space_t) space_list;
				/* list of all spaces */
	ibuf_data_t*	ibuf_data;
				/* insert buffer data */
	ulint		magic_n;
};

#define	FIL_SPACE_MAGIC_N	89472

unknown's avatar
unknown committed
180 181 182
/* The tablespace memory cache; also the totality of logs = the log data space,
is stored here; below we talk about tablespaces, but also the ib_logfiles
form a 'space' and it is handled here */
183 184 185

typedef	struct fil_system_struct	fil_system_t;
struct fil_system_struct {
unknown's avatar
unknown committed
186
	mutex_t		mutex;		/* The mutex protecting the cache */
187
	hash_table_t*	spaces;		/* The hash table of spaces in the
unknown's avatar
unknown committed
188 189 190 191
					system; they are hashed on the space
					id */
	hash_table_t*	name_hash;	/* hash table based on the space
					name */
192 193
	UT_LIST_BASE_NODE_T(fil_node_t) LRU;
					/* base node for the LRU list of the
unknown's avatar
unknown committed
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221
					most recently used open files with no
					pending i/o's; if we start an i/o on
					the file, we first remove it from this
					list, and return it to the start of
					the list when the i/o ends;
					log files and the system tablespace are
					not put to this list: they are opened
					after the startup, and kept open until
					shutdown */
	ulint		n_open;		/* number of files currently open */
	ulint		max_n_open;	/* n_open is not allowed to exceed
					this */
	ib_longlong	modification_counter;/* when we write to a file we
					increment this by one */
	ulint		max_assigned_id;/* maximum space id in the existing
					tables, or assigned during the time
					mysqld has been up; at an InnoDB
					startup we scan the data dictionary
					and set here the maximum of the
					space id's of the tables there */
	ib_longlong	tablespace_version;
					/* a counter which is incremented for
					every space object memory creation;
					every space mem object gets a
					'timestamp' from this; in DISCARD/
					IMPORT this is used to check if we
					should ignore an insert buffer merge
					request */
222 223 224 225
	UT_LIST_BASE_NODE_T(fil_space_t) space_list;
					/* list of all file spaces */
};

unknown's avatar
unknown committed
226 227
/* The tablespace memory cache. This variable is NULL before the module is
initialized. */
228 229
fil_system_t*	fil_system	= NULL;

unknown's avatar
unknown committed
230 231
/* The tablespace memory cache hash table size */
#define	FIL_SYSTEM_HASH_SIZE	50 /* TODO: make bigger! */
232 233


unknown's avatar
unknown committed
234 235
/************************************************************************
NOTE: you must call fil_mutex_enter_and_prepare_for_io() first!
236

unknown's avatar
unknown committed
237 238 239 240 241
Prepares a file node for i/o. Opens the file if it is closed. Updates the
pending i/o's field in the node and the system appropriately. Takes the node
off the LRU list if it is in the LRU list. The caller must hold the fil_sys
mutex. */
static
242
void
unknown's avatar
unknown committed
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
fil_node_prepare_for_io(
/*====================*/
	fil_node_t*	node,	/* in: file node */
	fil_system_t*	system,	/* in: tablespace memory cache */
	fil_space_t*	space);	/* in: space */
/************************************************************************
Updates the data structures when an i/o operation finishes. Updates the
pending i/o's field in the node appropriately. */
static
void
fil_node_complete_io(
/*=================*/
	fil_node_t*	node,	/* in: file node */
	fil_system_t*	system,	/* in: tablespace memory cache */
	ulint		type);	/* in: OS_FILE_WRITE or OS_FILE_READ; marks
				the node as modified if
				type == OS_FILE_WRITE */
260 261


unknown's avatar
unknown committed
262 263
/***********************************************************************
Returns the version number of a tablespace, -1 if not found. */
264

unknown's avatar
unknown committed
265 266 267 268 269 270 271 272 273 274
ib_longlong
fil_space_get_version(
/*==================*/
			/* out: version number, -1 if the tablespace does not
			exist in the memory cache */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;
	ib_longlong	version		= -1;
275

unknown's avatar
unknown committed
276
	ut_ad(system);
277

unknown's avatar
unknown committed
278
	mutex_enter(&(system->mutex));
279

unknown's avatar
unknown committed
280
	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);
281

unknown's avatar
unknown committed
282 283
	if (space) {
		version = space->tablespace_version;
284 285
	}

unknown's avatar
unknown committed
286
	mutex_exit(&(system->mutex));
287

unknown's avatar
unknown committed
288
	return(version);
289 290 291 292 293 294 295 296 297 298 299 300
}

/***********************************************************************
Returns the latch of a file space. */

rw_lock_t*
fil_space_get_latch(
/*================*/
			/* out: latch protecting storage allocation */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
301
	fil_space_t*	space;
302 303 304 305 306 307 308

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
309 310
	ut_a(space);

311 312 313 314 315 316 317 318 319 320 321 322 323 324 325
	mutex_exit(&(system->mutex));

	return(&(space->latch));
}

/***********************************************************************
Returns the type of a file space. */

ulint
fil_space_get_type(
/*===============*/
			/* out: FIL_TABLESPACE or FIL_LOG */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
326
	fil_space_t*	space;
327 328 329 330 331 332 333

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
334 335
	ut_a(space);

336 337 338 339 340 341 342 343 344 345 346 347 348 349
	mutex_exit(&(system->mutex));

	return(space->purpose);
}

/***********************************************************************
Returns the ibuf data of a file space. */

ibuf_data_t*
fil_space_get_ibuf_data(
/*====================*/
			/* out: ibuf data for this space */
	ulint	id)	/* in: space id */
{
unknown's avatar
unknown committed
350
	fil_system_t*	system		= fil_system;
351 352 353 354
	fil_space_t*	space;

	ut_ad(system);

unknown's avatar
unknown committed
355 356
	ut_a(id == 0);

357 358 359 360 361 362
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	mutex_exit(&(system->mutex));

unknown's avatar
unknown committed
363 364
	ut_a(space);

365 366 367 368 369 370 371 372 373 374 375 376
	return(space->ibuf_data);
}

/***********************************************************************
Appends a new file to the chain of files of a space. File must be closed. */

void
fil_node_create(
/*============*/
	char*	name,	/* in: file name (file must be closed) */
	ulint	size,	/* in: file size in database blocks, rounded downwards
			to an integer */
unknown's avatar
unknown committed
377 378
	ulint	id,	/* in: space id where to append */
	ibool	is_raw)	/* in: TRUE if a raw device or a raw disk partition */
379
{
unknown's avatar
unknown committed
380
	fil_system_t*	system	= fil_system;
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
	fil_node_t*	node;
	fil_space_t*	space;
	char*		name2;

	ut_a(system);
	ut_a(name);

	mutex_enter(&(system->mutex));

	node = mem_alloc(sizeof(fil_node_t));

	name2 = mem_alloc(ut_strlen(name) + 1);

	ut_strcpy(name2, name);

	node->name = name2;
	node->open = FALSE;
unknown's avatar
unknown committed
398 399 400 401

	ut_a(!is_raw || srv_start_raw_disk_in_use);

	node->is_raw_disk = is_raw;
402 403 404
	node->size = size;
	node->magic_n = FIL_NODE_MAGIC_N;
	node->n_pending = 0;
unknown's avatar
unknown committed
405
	node->n_pending_flushes = 0;
406

unknown's avatar
unknown committed
407 408
	node->modification_counter = 0;
	node->flush_counter = 0;
409 410 411
	
	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
412 413 414 415 416 417 418 419 420 421 422 423 424 425
	if (!space) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Error: Could not find tablespace %lu for\n"
"InnoDB: file %s from the tablespace memory cache.\n", id, name);
		mem_free(name2);

		mem_free(node);

		mutex_exit(&(system->mutex));

		return;
	}

426 427
	space->size += size;

unknown's avatar
unknown committed
428 429
	node->space = space;

430 431 432 433 434
	UT_LIST_ADD_LAST(chain, space->chain, node);
				
	mutex_exit(&(system->mutex));
}

unknown's avatar
unknown committed
435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
/************************************************************************
Opens a the file of a node of a tablespace. The caller must own the fil_system
mutex. */
static
void
fil_node_open_file(
/*===============*/
	fil_node_t*	node,	/* in: file node */
	fil_system_t*	system,	/* in: tablespace memory cache */
	fil_space_t*	space)	/* in: space */
{
	ib_longlong	size_bytes;
	ulint		size_low;
	ulint		size_high;
	ibool		ret;

	ut_ad(mutex_own(&(system->mutex)));

	ut_a(node->n_pending == 0);
	ut_a(node->open == FALSE);

	/* printf("Opening file %s\n", node->name); */

	if (space->purpose == FIL_LOG) {	
		node->handle = os_file_create(node->name, OS_FILE_OPEN,
					OS_FILE_AIO, OS_LOG_FILE, &ret);
	} else if (node->is_raw_disk) {
		node->handle = os_file_create(node->name,
				        OS_FILE_OPEN_RAW,
					OS_FILE_AIO, OS_DATA_FILE, &ret);
	} else {
		node->handle = os_file_create(node->name, OS_FILE_OPEN,
					OS_FILE_AIO, OS_DATA_FILE, &ret);
	}
		
	ut_a(ret);
		
	node->open = TRUE;

	system->n_open++;

	if (node->size == 0) {
		/* It must be a single-table tablespace and we do not know the
		size of the file yet */

		ut_a(space->id != 0);
			
		os_file_get_size(node->handle, &size_low, &size_high);

		size_bytes = (((ib_longlong)size_high) << 32)
				     + (ib_longlong)size_low;

		if (size_bytes >= FSP_EXTENT_SIZE * UNIV_PAGE_SIZE) {
			node->size = (ulint) ((size_bytes / (1024 * 1024))
					   * ((1024 * 1024) / UNIV_PAGE_SIZE));
		} else {
			node->size = (ulint) (size_bytes / UNIV_PAGE_SIZE);
		}

		space->size = node->size;
	}

	if (space->purpose == FIL_TABLESPACE && space->id != 0) {
		/* Put the node to the LRU list */
		UT_LIST_ADD_FIRST(LRU, system->LRU, node);
	}
}

503 504 505 506
/**************************************************************************
Closes a file. */
static
void
unknown's avatar
unknown committed
507 508
fil_node_close_file(
/*================*/
509
	fil_node_t*	node,	/* in: file node */
unknown's avatar
unknown committed
510
	fil_system_t*	system)	/* in: tablespace memory cache */
511 512 513 514 515 516 517
{
	ibool	ret;

	ut_ad(node && system);
	ut_ad(mutex_own(&(system->mutex)));
	ut_a(node->open);
	ut_a(node->n_pending == 0);
unknown's avatar
unknown committed
518
	ut_a(node->n_pending_flushes == 0);
519 520 521 522

	ret = os_file_close(node->handle);
	ut_a(ret);

unknown's avatar
unknown committed
523 524
	/* printf("Closing file %s\n", node->name); */

525
	node->open = FALSE;
unknown's avatar
unknown committed
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
	ut_a(system->n_open > 0);
	system->n_open--;

	if (node->space->purpose == FIL_TABLESPACE && node->space->id != 0) {
		ut_a(UT_LIST_GET_LEN(system->LRU) > 0);

		/* The node is in the LRU list, remove it */
		UT_LIST_REMOVE(LRU, system->LRU, node);
	}
}

/************************************************************************
Tries to close a file in the LRU list. The caller must hold the fil_sys
mutex. */
static
ibool
fil_try_to_close_file_in_LRU(
/*=========================*/
				/* out: TRUE if success, FALSE if should retry
				later; since i/o's generally complete in < 
				100 ms, and as InnoDB writes at most 128 pages
				from the buffer pool in a batch, and then
				immediately flushes the files, there is a good
				chance that the next time we find a suitable
				node from the LRU list */
	ibool	print_info)	/* in: if TRUE, prints information why it
				cannot close a file */
{
	fil_system_t*	system		= fil_system;
	fil_node_t*	node;

	ut_ad(mutex_own(&(system->mutex)));

	node = UT_LIST_GET_LAST(system->LRU);

	if (print_info) {
		fprintf(stderr,
"InnoDB: fil_sys open file LRU len %lu\n", UT_LIST_GET_LEN(system->LRU));
	}

	while (node != NULL) {
		if (node->modification_counter == node->flush_counter
		    && node->n_pending_flushes == 0) {
569

unknown's avatar
unknown committed
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706
			fil_node_close_file(node, system);
			
			return(TRUE);
		}
		
		if (print_info && node->n_pending_flushes > 0) {
			fprintf(stderr,
"InnoDB: cannot close file %s, because n_pending_flushes %lu\n", node->name,
				       node->n_pending_flushes);
		}

		if (print_info
		    && node->modification_counter != node->flush_counter) {
			fprintf(stderr,
"InnoDB: cannot close file %s, because mod_count %lld != fl_count %lld\n",
				node->name, node->modification_counter,
				node->flush_counter);
		}

		node = UT_LIST_GET_PREV(LRU, node);
	}

	return(FALSE);
}

/***********************************************************************
Reserves the fil_system mutex and tries to make sure we can open at least one
file while holding it. This should be called before calling
fil_node_prepare_for_io(), because that function may need to open a file. */
static
void
fil_mutex_enter_and_prepare_for_io(
/*===============================*/
	ulint	space_id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;
	ibool		success;
	ibool		print_info	= FALSE;
	ulint		count		= 0;
	ulint		count2		= 0;

	ut_ad(!mutex_own(&(system->mutex)));
retry:
	mutex_enter(&(system->mutex));

	if (space_id == 0 || space_id >= SRV_LOG_SPACE_FIRST_ID) {
		/* We keep log files and system tablespace files always open;
		this is important in preventing deadlocks in this module, as
		a page read completion often performs another read from the
		insert buffer. The insert buffer is in tablespace 0, and we
		cannot end up waiting in this function. */

		return;
	}

	if (system->n_open < system->max_n_open) {

		return;
	}

	HASH_SEARCH(hash, system->spaces, space_id, space,
							space->id == space_id);
	if (space != NULL && space->stop_ios) {
		/* We are going to do a rename file and want to stop new i/o's
		for a while */

		if (count2 > 20000) {
			fprintf(stderr,
"InnoDB: Warning: tablespace %s has i/o ops stopped for a long time %lu\n",
						    space->name, count2);
		}

		mutex_exit(&(system->mutex));

		os_thread_sleep(20000);

		count2++;

		goto retry;
	}

	/* If the file is already open, no need to do anything; if the space
	does not exist, we handle the situation in the function which called
	this function */

	if (!space || UT_LIST_GET_FIRST(space->chain)->open) {

		return;
	}

	if (count > 1) {
		print_info = TRUE;
	}

	/* Too many files are open, try to close some */
close_more:
	success = fil_try_to_close_file_in_LRU(print_info);

	if (success && system->n_open >= system->max_n_open) {

		goto close_more;
	}

	if (system->n_open < system->max_n_open) {
		/* Ok */

		return;
	}

	if (count >= 2) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Warning: too many (%lu) files stay open while the maximum\n"
"InnoDB: allowed value would be %lu.\n"
"InnoDB: You may need to raise the value of innodb_max_files_open in\n"
"InnoDB: my.cnf.\n", system->n_open, system->max_n_open);

		return;
	}

	mutex_exit(&(system->mutex));

	/* Wake the i/o-handler threads to make sure pending i/o's are
	performed */
	os_aio_simulated_wake_handler_threads();

	os_thread_sleep(20000);

	/* Flush tablespaces so that we can close modified files in the LRU
	list */

	fil_flush_file_spaces(FIL_TABLESPACE);		

	count++;

	goto retry;
707 708 709
}

/***********************************************************************
unknown's avatar
unknown committed
710
Frees a file node object from a tablespace memory cache. */
711 712 713 714 715
static
void
fil_node_free(
/*==========*/
	fil_node_t*	node,	/* in, own: file node */
unknown's avatar
unknown committed
716
	fil_system_t*	system,	/* in: tablespace memory cache */
717 718 719 720 721
	fil_space_t*	space)	/* in: space where the file node is chained */
{
	ut_ad(node && system && space);
	ut_ad(mutex_own(&(system->mutex)));
	ut_a(node->magic_n == FIL_NODE_MAGIC_N);
unknown's avatar
unknown committed
722
	ut_a(node->n_pending == 0);
723 724

	if (node->open) {
unknown's avatar
unknown committed
725
		fil_node_close_file(node, system);
726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747
	}

	space->size -= node->size;
	
	UT_LIST_REMOVE(chain, space->chain, node);

	mem_free(node->name);
	mem_free(node);
}

/********************************************************************
Drops files from the start of a file space, so that its size is cut by
the amount given. */

void
fil_space_truncate_start(
/*=====================*/
	ulint	id,		/* in: space id */
	ulint	trunc_len)	/* in: truncate by this much; it is an error
				if this does not equal to the combined size of
				some initial files in the space */
{
unknown's avatar
unknown committed
748
	fil_system_t*	system		= fil_system;
749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768
	fil_node_t*	node;
	fil_space_t*	space;

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	ut_a(space);
	
	while (trunc_len > 0) {
		node = UT_LIST_GET_FIRST(space->chain);

		ut_a(node->size * UNIV_PAGE_SIZE >= trunc_len);

		trunc_len -= node->size * UNIV_PAGE_SIZE;

		fil_node_free(node, system, space);
	}	
				
	mutex_exit(&(system->mutex));
unknown's avatar
unknown committed
769
}
770

unknown's avatar
unknown committed
771 772 773 774 775 776 777 778 779 780 781
/***********************************************************************
Creates a space memory object and puts it to the tablespace memory cache. If
there is an error, prints an error message to the .err log. */

ibool
fil_space_create(
/*=============*/
			/* out: TRUE if success */
	char*	name,	/* in: space name */
	ulint	id,	/* in: space id */
	ulint	purpose)/* in: FIL_TABLESPACE, or FIL_LOG if log */
782
{
unknown's avatar
unknown committed
783 784 785 786 787 788 789 790
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;	
	char*		name2;
	ulint		namesake_id;
try_again:
	/*printf(
	"InnoDB: Adding tablespace %lu of name %s, purpose %lu\n", id, name,
	  purpose);*/
791

unknown's avatar
unknown committed
792 793
	ut_a(system);
	ut_a(name);
794

unknown's avatar
unknown committed
795
	mutex_enter(&(system->mutex));
796

unknown's avatar
unknown committed
797 798 799 800 801 802 803 804 805 806
	HASH_SEARCH(name_hash, system->name_hash, ut_fold_string(name), space,
					0 == strcmp(name, space->name));
	if (space != NULL) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Warning: trying to init to the tablespace memory cache\n"
"InnoDB: a tablespace %lu of name %s,\n"
"InnoDB: but a tablespace %lu of the same name %s\n"
"InnoDB: already exists in the tablespace memory cache!\n",
			       id, name, space->id, space->name);
807

unknown's avatar
unknown committed
808
		if (id == 0 || purpose != FIL_TABLESPACE) {
809

unknown's avatar
unknown committed
810
			mutex_exit(&(system->mutex));
811

unknown's avatar
unknown committed
812 813
			return(FALSE);
		}
814

unknown's avatar
unknown committed
815 816 817 818 819 820 821 822
		fprintf(stderr,
"InnoDB: We assume that InnoDB did a crash recovery, and you had\n"
"InnoDB: an .ibd file for which the table did not exist in the\n"
"InnoDB: InnoDB internal data dictionary in the ibdata files.\n"
"InnoDB: We assume that you later removed the .ibd and .frm files,\n"
"InnoDB: and are now trying to recreate the table. We now remove the\n"
"InnoDB: conflicting tablespace object from the memory cache and try\n"
"InnoDB: the init again.\n");
823

unknown's avatar
unknown committed
824
		namesake_id = space->id;
825

unknown's avatar
unknown committed
826
		mutex_exit(&(system->mutex));
827

unknown's avatar
unknown committed
828
		fil_space_free(namesake_id);
829

unknown's avatar
unknown committed
830 831
		goto try_again;
	}
832

unknown's avatar
unknown committed
833
	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);
834

unknown's avatar
unknown committed
835 836 837 838 839 840
	if (space != NULL) {
		fprintf(stderr,
"InnoDB: Error: trying to add tablespace %lu of name %s\n"
"InnoDB: to the tablespace memory cache, but tablespace\n"
"InnoDB: %lu of name %s already exists in the tablespace\n"
"InnoDB: memory cache!\n", id, name, space->id, space->name);
841

unknown's avatar
unknown committed
842
		mutex_exit(&(system->mutex));
843

unknown's avatar
unknown committed
844
		return(FALSE);
845 846
	}

unknown's avatar
unknown committed
847
	space = mem_alloc(sizeof(fil_space_t));
848

unknown's avatar
unknown committed
849
	name2 = mem_alloc(ut_strlen(name) + 1);
850

unknown's avatar
unknown committed
851
	ut_strcpy(name2, name);
852

unknown's avatar
unknown committed
853 854
	space->name = name2;
	space->id = id;
855

unknown's avatar
unknown committed
856 857 858 859
	system->tablespace_version++;
	space->tablespace_version =
				system->tablespace_version;
	space->mark = FALSE;
860

unknown's avatar
unknown committed
861 862
	if (purpose == FIL_TABLESPACE && id > system->max_assigned_id) {
		system->max_assigned_id = id;
863 864
	}

unknown's avatar
unknown committed
865 866 867 868 869
	space->stop_ios = FALSE;
	space->stop_ibuf_merges = FALSE;
	space->is_being_deleted = FALSE;
	space->purpose = purpose;
	space->size = 0;
870

unknown's avatar
unknown committed
871
	space->n_reserved_extents = 0;
unknown's avatar
unknown committed
872
	
unknown's avatar
unknown committed
873 874
	space->n_pending_flushes = 0;
	space->n_pending_ibuf_merges = 0;
875

unknown's avatar
unknown committed
876 877
	UT_LIST_INIT(space->chain);
	space->magic_n = FIL_SPACE_MAGIC_N;
878

unknown's avatar
unknown committed
879 880 881 882 883 884
	space->ibuf_data = NULL;
	
	rw_lock_create(&(space->latch));
	rw_lock_set_level(&(space->latch), SYNC_FSP);
	
	HASH_INSERT(fil_space_t, hash, system->spaces, id, space);
885

unknown's avatar
unknown committed
886 887 888 889 890
	HASH_INSERT(fil_space_t, name_hash, system->name_hash,
						ut_fold_string(name), space);
	UT_LIST_ADD_LAST(space_list, system->space_list, space);
				
	mutex_exit(&(system->mutex));
891

unknown's avatar
unknown committed
892
	return(TRUE);
893 894 895
}

/***********************************************************************
unknown's avatar
unknown committed
896 897 898 899 900 901 902 903 904
Assigns a new space id for a new single-table tablespace. This works simply by
incrementing the global counter. If 4 billion id's is not enough, we may need
to recycle id's. */
static
ulint
fil_assign_new_space_id(void)
/*=========================*/
			/* out: new tablespace id; ULINT_UNDEFINED if could
			not assign an id */
905 906
{
	fil_system_t*	system = fil_system;
unknown's avatar
unknown committed
907
	ulint		id;
908 909 910

	mutex_enter(&(system->mutex));

unknown's avatar
unknown committed
911
	system->max_assigned_id++;
912

unknown's avatar
unknown committed
913
	id = system->max_assigned_id;
914

unknown's avatar
unknown committed
915 916 917 918 919 920 921 922 923
	if (id > (SRV_LOG_SPACE_FIRST_ID / 2) && (id % 1000000UL == 0)) {
	        ut_print_timestamp(stderr);
	        fprintf(stderr,
"InnoDB: Warning: you are running out of new single-table tablespace id's.\n"
"InnoDB: Current counter is %lu and it must not exceed %lu!\n"
"InnoDB: To reset the counter to zero you have to dump all your tables and\n"
"InnoDB: recreate the whole InnoDB installation.\n", id,
						     SRV_LOG_SPACE_FIRST_ID);
	}
924

unknown's avatar
unknown committed
925 926 927 928 929 930 931 932
	if (id >= SRV_LOG_SPACE_FIRST_ID) {
	        ut_print_timestamp(stderr);
	        fprintf(stderr,
"InnoDB: You have run out of single-table tablespace id's!\n"
"InnoDB: Current counter is %lu.\n"
"InnoDB: To reset the counter to zero you have to dump all your tables and\n"
"InnoDB: recreate the whole InnoDB installation.\n", id);
		system->max_assigned_id--;
933

unknown's avatar
unknown committed
934 935
		id = ULINT_UNDEFINED;
	}
936 937

	mutex_exit(&(system->mutex));
unknown's avatar
unknown committed
938 939

	return(id);
940 941 942
}

/***********************************************************************
unknown's avatar
unknown committed
943 944 945
Frees a space object from the tablespace memory cache. Closes the files in
the chain but does not delete them. There must not be any pending i/o's or
flushes on the files. */
946

unknown's avatar
unknown committed
947
ibool
948 949
fil_space_free(
/*===========*/
unknown's avatar
unknown committed
950
			/* out: TRUE if success */
951 952
	ulint	id)	/* in: space id */
{
unknown's avatar
unknown committed
953
	fil_system_t*	system = fil_system;
954
	fil_space_t*	space;
unknown's avatar
unknown committed
955
	fil_space_t*	namespace;
956
	fil_node_t*	fil_node;
unknown's avatar
unknown committed
957

958 959 960 961
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
962 963 964 965 966 967 968 969 970 971 972
	if (!space) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Error: trying to remove tablespace %lu from the cache but\n"
"InnoDB: it is not there.\n", id);

		mutex_exit(&(system->mutex));
		
		return(FALSE);
	}

973 974
	HASH_DELETE(fil_space_t, hash, system->spaces, id, space);

unknown's avatar
unknown committed
975 976 977 978 979 980 981 982
	HASH_SEARCH(name_hash, system->name_hash, ut_fold_string(space->name),
		    namespace, 0 == strcmp(space->name, namespace->name));
	ut_a(namespace);
	ut_a(space == namespace);

	HASH_DELETE(fil_space_t, name_hash, system->name_hash,
					   ut_fold_string(space->name), space);

983 984 985
	UT_LIST_REMOVE(space_list, system->space_list, space);

	ut_a(space->magic_n == FIL_SPACE_MAGIC_N);
unknown's avatar
unknown committed
986
	ut_a(0 == space->n_pending_flushes);
987 988 989 990 991 992 993 994 995

	fil_node = UT_LIST_GET_FIRST(space->chain);

	while (fil_node != NULL) {
		fil_node_free(fil_node, system, space);

		fil_node = UT_LIST_GET_FIRST(space->chain);
	}	
	
unknown's avatar
unknown committed
996
	ut_a(0 == UT_LIST_GET_LEN(space->chain));
997 998 999

	mutex_exit(&(system->mutex));

unknown's avatar
unknown committed
1000 1001
	rw_lock_free(&(space->latch));

1002 1003
	mem_free(space->name);
	mem_free(space);
unknown's avatar
unknown committed
1004 1005

	return(TRUE);
1006 1007 1008
}

/***********************************************************************
unknown's avatar
unknown committed
1009 1010
Returns the size of the space in pages. The tablespace must be cached in the
memory cache. */
1011 1012 1013 1014

ulint
fil_space_get_size(
/*===============*/
unknown's avatar
unknown committed
1015
			/* out: space size, 0 if space not found */
1016 1017
	ulint	id)	/* in: space id */
{
unknown's avatar
unknown committed
1018 1019
	fil_system_t*	system 		= fil_system;
	fil_node_t*	node;
1020 1021 1022 1023 1024
	fil_space_t*	space;
	ulint		size;

	ut_ad(system);

unknown's avatar
unknown committed
1025
	fil_mutex_enter_and_prepare_for_io(id);
1026 1027 1028

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
	if (space == NULL) {
		mutex_exit(&(system->mutex));

		return(0);
	}

	if (space->size == 0 && space->purpose == FIL_TABLESPACE) {
		ut_a(id != 0);

		ut_a(1 == UT_LIST_GET_LEN(space->chain));

		node = UT_LIST_GET_FIRST(space->chain);

		/* It must be a single-table tablespace and we have not opened
		the file yet; the following calls will open it and update the
		size fields */

		fil_node_prepare_for_io(node, system, space);
		fil_node_complete_io(node, system, OS_FILE_READ);
	}

1050 1051 1052 1053
	size = space->size;
	
	mutex_exit(&(system->mutex));

unknown's avatar
unknown committed
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
	return(size);
}

/***********************************************************************
Checks if the pair space, page_no refers to an existing page in a tablespace
file space. The tablespace must be cached in the memory cache. */

ibool
fil_check_adress_in_tablespace(
/*===========================*/
			/* out: TRUE if the address is meaningful */
	ulint	id,	/* in: space id */
	ulint	page_no)/* in: page number */
{
	if (fil_space_get_size(id) > page_no) {

		return(TRUE);
	}

	return(FALSE);
}		

/********************************************************************
Creates a the tablespace memory cache. */
static
fil_system_t*
fil_system_create(
/*==============*/
				/* out, own: tablespace memory cache */
	ulint	hash_size,	/* in: hash table size */
	ulint	max_n_open)	/* in: maximum number of open files; must be
				> 10 */
{
	fil_system_t*	system;

	ut_a(hash_size > 0);
	ut_a(max_n_open > 0);

	system = mem_alloc(sizeof(fil_system_t));

	mutex_create(&(system->mutex));

	mutex_set_level(&(system->mutex), SYNC_ANY_LATCH);

	system->spaces = hash_create(hash_size);
	system->name_hash = hash_create(hash_size);

	UT_LIST_INIT(system->LRU);

	system->n_open = 0;
	system->max_n_open = max_n_open;

	system->modification_counter = 0;
	system->max_assigned_id = 0;

	system->tablespace_version = 0;

	UT_LIST_INIT(system->space_list);

	return(system);
}

/********************************************************************
Initializes the tablespace memory cache. */

void
fil_init(
/*=====*/
	ulint	max_n_open)	/* in: max number of open files */
{
	ut_a(fil_system == NULL);

	/*printf("Initializing the tablespace cache with max %lu open files\n",
							       max_n_open); */
	fil_system = fil_system_create(FIL_SYSTEM_HASH_SIZE, max_n_open);
}

/***********************************************************************
Opens all log files and system tablespace data files. They stay open until the
database server shutdown. This should be called at a server startup after the
space objects for the log and the system tablespace have been created. The
purpose of this operation is to make sure we never run out of file descriptors
if we need to read from the insert buffer or to write to the log. */

void
fil_open_log_and_system_tablespace_files(void)
/*==========================================*/
{
	fil_system_t*	system = fil_system;
	fil_space_t*	space;
	fil_node_t*	node;

	mutex_enter(&(system->mutex));

	space = UT_LIST_GET_FIRST(system->space_list);

	while (space != NULL) {
		if (space->purpose != FIL_TABLESPACE || space->id == 0) {
			node = UT_LIST_GET_FIRST(space->chain);

			while (node != NULL) {
				if (!node->open) {
					fil_node_open_file(node, system,
									space);
				}
				if (system->max_n_open < 10 + system->n_open) {
					fprintf(stderr,
"InnoDB: Warning: you must raise the value of innodb_max_open_files in\n"
"InnoDB: my.cnf! Remember that InnoDB keeps all log files and all system\n"
"InnoDB: tablespace files open for the whole time mysqld is running, and\n"
"InnoDB: needs to open also some .ibd files if the file-per-table storage\n"
"InnoDB: model is used. Current open files %lu, max allowed open files %lu.\n",
				     system->n_open, system->max_n_open);
				}
				node = UT_LIST_GET_NEXT(chain, node);
			}
		}
		space = UT_LIST_GET_NEXT(space_list, space);
	}

	mutex_exit(&(system->mutex));
}

/***********************************************************************
Closes all open files. There must not be any pending i/o's or not flushed
modifications in the files. */

void
fil_close_all_files(void)
/*=====================*/
{
	fil_system_t*	system = fil_system;
	fil_space_t*	space;
	fil_node_t*	node;

	mutex_enter(&(system->mutex));

	space = UT_LIST_GET_FIRST(system->space_list);

	while (space != NULL) {
		node = UT_LIST_GET_FIRST(space->chain);

		while (node != NULL) {
			if (node->open) {
				fil_node_close_file(node, system);
			}
			node = UT_LIST_GET_NEXT(chain, node);
		}
		space = UT_LIST_GET_NEXT(space_list, space);
	}

	mutex_exit(&(system->mutex));
}

/***********************************************************************
Sets the max tablespace id counter if the given number is bigger than the
previous value. */

void
fil_set_max_space_id_if_bigger(
/*===========================*/
	ulint	max_id)	/* in: maximum known id */
{
	fil_system_t*	system = fil_system;

	if (max_id >= SRV_LOG_SPACE_FIRST_ID) {
		fprintf(stderr,
"InnoDB: Fatal error: max tablespace id is too high, %lu\n", max_id);
		ut_a(0);
	}

	mutex_enter(&(system->mutex));

	if (system->max_assigned_id < max_id) {

		system->max_assigned_id = max_id;
	}

	mutex_exit(&(system->mutex));
}

/********************************************************************
Initializes the ibuf data structure for space 0 == the system tablespace.
This can be called after the file space headers have been created and the
dictionary system has been initialized. */

void
fil_ibuf_init_at_db_start(void)
/*===========================*/
{
	fil_space_t*	space;

	space = UT_LIST_GET_FIRST(fil_system->space_list);

	ut_a(space);
        ut_a(space->purpose == FIL_TABLESPACE);	

	space->ibuf_data = ibuf_data_init_for_space(space->id);
}

/********************************************************************
Writes the flushed lsn and the latest archived log number to the page header
of the first page of a data file. */
static
ulint
fil_write_lsn_and_arch_no_to_file(
/*==============================*/
	ulint	space_id,	/* in: space number */
	ulint	sum_of_sizes,	/* in: combined size of previous files in
				space, in database pages */
	dulint	lsn,		/* in: lsn to write */
	ulint	arch_log_no)	/* in: archived log number to write */
{
	byte*	buf1;
	byte*	buf;

	UT_NOT_USED(arch_log_no);

	buf1 = mem_alloc(2 * UNIV_PAGE_SIZE);
	buf = ut_align(buf1, UNIV_PAGE_SIZE);

	fil_read(TRUE, space_id, sum_of_sizes, 0, UNIV_PAGE_SIZE, buf, NULL);

	mach_write_to_8(buf + FIL_PAGE_FILE_FLUSH_LSN, lsn);

	fil_write(TRUE, space_id, sum_of_sizes, 0, UNIV_PAGE_SIZE, buf, NULL);

	return(DB_SUCCESS);	
}

/********************************************************************
Writes the flushed lsn and the latest archived log number to the page
header of the first page of each data file in the system tablespace. */

ulint
fil_write_flushed_lsn_to_data_files(
/*================================*/
				/* out: DB_SUCCESS or error number */
	dulint	lsn,		/* in: lsn to write */
	ulint	arch_log_no)	/* in: latest archived log file number */
{
	fil_space_t*	space;
	fil_node_t*	node;
	ulint		sum_of_sizes;
	ulint		err;

	mutex_enter(&(fil_system->mutex));
	
	space = UT_LIST_GET_FIRST(fil_system->space_list);
	
	while (space) {
		/* We only write the lsn to the system tablespace
		(space id == 0) files */

		if (space->id == 0) {
			ut_a(space->purpose == FIL_TABLESPACE);
			sum_of_sizes = 0;

			node = UT_LIST_GET_FIRST(space->chain);
			while (node) {
				mutex_exit(&(fil_system->mutex));

				err = fil_write_lsn_and_arch_no_to_file(
						space->id, sum_of_sizes,
						lsn, arch_log_no);
				if (err != DB_SUCCESS) {

					return(err);
				}

				mutex_enter(&(fil_system->mutex));

				sum_of_sizes += node->size;
				node = UT_LIST_GET_NEXT(chain, node);
			}

			break; /* there is only one space with id == 0 */
		}
		space = UT_LIST_GET_NEXT(space_list, space);
	}

	mutex_exit(&(fil_system->mutex));

	return(DB_SUCCESS);
}

/***********************************************************************
Reads the flushed lsn and arch no fields from a data file at database
startup. */

void
fil_read_flushed_lsn_and_arch_log_no(
/*=================================*/
	os_file_t data_file,		/* in: open data file */
	ibool	one_read_already,	/* in: TRUE if min and max parameters
					below already contain sensible data */
	dulint*	min_flushed_lsn,	/* in/out: */
	ulint*	min_arch_log_no,	/* in/out: */
	dulint*	max_flushed_lsn,	/* in/out: */
	ulint*	max_arch_log_no)	/* in/out: */
{
	byte*	buf;
	byte*	buf2;
	dulint	flushed_lsn;
        ulint   arch_log_no     = 0;    /* since InnoDB does not archive
                                        its own logs under MySQL, this
                                        parameter is not relevant */
	buf2 = ut_malloc(2 * UNIV_PAGE_SIZE);
	/* Align the memory for a possible read from a raw device */
	buf = ut_align(buf2, UNIV_PAGE_SIZE);
	
	os_file_read(data_file, buf, 0, 0, UNIV_PAGE_SIZE);

	flushed_lsn = mach_read_from_8(buf + FIL_PAGE_FILE_FLUSH_LSN);

	ut_free(buf2);

	if (!one_read_already) {
		*min_flushed_lsn = flushed_lsn;
		*max_flushed_lsn = flushed_lsn;
		*min_arch_log_no = arch_log_no;
		*max_arch_log_no = arch_log_no;

		return;
	}

	if (ut_dulint_cmp(*min_flushed_lsn, flushed_lsn) > 0) {
		*min_flushed_lsn = flushed_lsn;
	}
	if (ut_dulint_cmp(*max_flushed_lsn, flushed_lsn) < 0) {
		*max_flushed_lsn = flushed_lsn;
	}
	if (*min_arch_log_no > arch_log_no) {
		*min_arch_log_no = arch_log_no;
	}
	if (*max_arch_log_no < arch_log_no) {
		*max_arch_log_no = arch_log_no;
	}
}

/*================ SINGLE-TABLE TABLESPACES ==========================*/

/***********************************************************************
Increments the count of pending insert buffer page merges, if space is not
being deleted. */

ibool
fil_inc_pending_ibuf_merges(
/*========================*/
			/* out: TRUE if being deleted, and ibuf merges should
			be skipped */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;
	
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL) {
		fprintf(stderr,
"InnoDB: Error: trying to do ibuf merge to a dropped tablespace %lu\n", id);
	}

	if (space == NULL || space->stop_ibuf_merges) {
		mutex_exit(&(system->mutex));

		return(TRUE);
	}

	space->n_pending_ibuf_merges++;

	mutex_exit(&(system->mutex));

	return(FALSE);
}

/***********************************************************************
Decrements the count of pending insert buffer page merges. */

void
fil_decr_pending_ibuf_merges(
/*========================*/
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;
	
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL) {
		fprintf(stderr,
"InnoDB: Error: decrementing ibuf merge of a dropped tablespace %lu\n", id);
	}

	if (space != NULL) {
		space->n_pending_ibuf_merges--;
	}

	mutex_exit(&(system->mutex));
}

/***********************************************************************
Deletes a single-table tablespace. The tablespace must be cached in the
memory cache. */

ibool
fil_delete_tablespace(
/*==================*/
			/* out: TRUE if success */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	ibool		success;
	fil_space_t*	space;
	fil_node_t*	node;
	ulint		count		= 0;
	char		path[OS_FILE_MAX_PATH];

	ut_a(id != 0);
stop_ibuf_merges:
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space != NULL) {
		space->stop_ibuf_merges = TRUE;

		if (space->n_pending_ibuf_merges == 0) {
			mutex_exit(&(system->mutex));

			count = 0;

			goto try_again;
		} else {
			if (count > 5000) {
			   ut_print_timestamp(stderr);
			   fprintf(stderr,
"  InnoDB: Warning: trying to delete tablespace %s,\n"
"InnoDB: but there are %lu pending ibuf merges on it.\n"
"InnoDB: Loop %lu.\n", space->name, space->n_pending_ibuf_merges, count);
			}

			mutex_exit(&(system->mutex));

			os_thread_sleep(20000);
			count++;

			goto stop_ibuf_merges;
		}
	}

	mutex_exit(&(system->mutex));
	count = 0;

try_again:
	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Error: cannot delete tablespace %lu because it is not found\n"
"InnoDB: in the tablespace memory cache.\n", id);

		mutex_exit(&(system->mutex));
	
		return(FALSE);
	}	

	ut_a(space);
	ut_a(strlen(space->name) < OS_FILE_MAX_PATH);
	ut_a(space->n_pending_ibuf_merges == 0);

	strcpy(path, space->name);

	space->is_being_deleted = TRUE;

	ut_a(UT_LIST_GET_LEN(space->chain) == 1);
	node = UT_LIST_GET_FIRST(space->chain);

	if (space->n_pending_flushes > 0 || node->n_pending > 0) {
		if (count > 1000) {
			ut_print_timestamp(stderr);
			fprintf(stderr,
"  InnoDB: Warning: trying to delete tablespace %s,\n"
"InnoDB: but there are %lu flushes and %lu pending i/o's on it\n"
"InnoDB: Loop %lu.\n", space->name, space->n_pending_flushes, node->n_pending,
									count);
		}
		mutex_exit(&(system->mutex));
		os_thread_sleep(20000);

		count++;

		goto try_again;
	}

	mutex_exit(&(system->mutex));

	/* Invalidate in the buffer pool all pages belonging to the
	tablespace. Since we have set space->is_being_deleted = TRUE, readahead
	or ibuf merge can no longer read more pages of this tablespace to the
	buffer pool. Thus we can clean the tablespace out of the buffer pool
	completely and permanently. The flag is_being_deleted also prevents
	fil_flush() from being applied to this tablespace. */

	buf_LRU_invalidate_tablespace(id);

	success = fil_space_free(id);

	if (success) {
		success = os_file_delete(path);

		if (success) {

			return(TRUE);
		}
	}

	return(FALSE);
}

/***********************************************************************
Discards a single-table tablespace. The tablespace must be cached in the
memory cache. Discarding is like deleting a tablespace, but
1) we do not drop the table from the data dictionary;
2) we remove all insert buffer entries for the tablespace immediately; in DROP
TABLE they are only removed gradually in the background;
3) when the user does IMPORT TABLESPACE, the tablespace will have the same id
as it originally had. */

ibool
fil_discard_tablespace(
/*===================*/
			/* out: TRUE if success */
	ulint	id)	/* in: space id */
{
	ibool	success;

	success = fil_delete_tablespace(id);

	if (!success) {
		fprintf(stderr,
"InnoDB: Warning: cannot delete tablespace %lu in DISCARD TABLESPACE.\n"
"InnoDB: But let us remove the insert buffer entries for this tablespace.\n",
									id); 
	}

	/* Remove all insert buffer entries for the tablespace */

	ibuf_delete_for_discarded_space(id);

	return(TRUE);
}

/***********************************************************************
Renames the memory cache structures of a single-table tablespace. */
static
ibool
fil_rename_tablespace_in_mem(
/*=========================*/
				/* out: TRUE if success */
	fil_space_t*	space,	/* in: tablespace memory object */
	fil_node_t*	node,	/* in: file node of that tablespace */
	char*		path)	/* in: new name */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space2;
	char*		old_name	= space->name;
	
	HASH_SEARCH(name_hash, system->name_hash, ut_fold_string(old_name),
			       space2, 0 == strcmp(old_name, space2->name));
	if (space != space2) {
		fprintf(stderr,
"InnoDB: Error: cannot find %s in tablespace memory cache\n", old_name);

		return(FALSE);
	}

	HASH_SEARCH(name_hash, system->name_hash, ut_fold_string(path),
			       space2, 0 == strcmp(path, space2->name));
	if (space2 != NULL) {
		fprintf(stderr,
"InnoDB: Error: %s is already in tablespace memory cache\n", path);
		
		return(FALSE);
	}

	HASH_DELETE(fil_space_t, name_hash, system->name_hash,
					   ut_fold_string(space->name), space);
	mem_free(space->name);
	mem_free(node->name);

	space->name = mem_alloc(strlen(path) + 1);
	node->name = mem_alloc(strlen(path) + 1);

	strcpy(space->name, path);
	strcpy(node->name, path);

	HASH_INSERT(fil_space_t, name_hash, system->name_hash,
						ut_fold_string(path), space);
	return(TRUE);
}

/***********************************************************************
Renames a single-table tablespace. The tablespace must be cached in the
tablespace memory cache. */

ibool
fil_rename_tablespace(
/*==================*/
				/* out: TRUE if success */
	char*	old_name,	/* in: old table name in the standard
				databasename/tablename format of InnoDB */
	ulint	id,		/* in: space id */
	char*	new_name)	/* in: new table name in the standard
				databasename/tablename format of InnoDB */
{
	fil_system_t*	system		= fil_system;
	ibool		success;
	fil_space_t*	space;
	fil_node_t*	node;
	ulint		count		= 0;
	char		old_path[OS_FILE_MAX_PATH];
	char		path[OS_FILE_MAX_PATH];

	ut_a(id != 0);
retry:
	count++;

	if (count > 1000) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Warning: problems renaming %s to %s, %lu iterations\n",
						old_name, new_name, count);
	}

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL) {
		fprintf(stderr,
"InnoDB: Error: cannot find space id %lu from the tablespace memory cache\n"
"InnoDB: though the table %s in a rename operation should have that id\n",
							  id, old_name);
		mutex_exit(&(system->mutex));

		return(FALSE);
	}

	if (count > 25000) {
		space->stop_ios = FALSE;
		mutex_exit(&(system->mutex));

		return(FALSE);
	}

	/* We temporarily close the .ibd file because we do not trust that
	operating systems can rename an open file. For the closing we have to
	wait until there are no pending i/o's or flushes on the file. */

	space->stop_ios = TRUE;

	ut_a(UT_LIST_GET_LEN(space->chain) == 1);
	node = UT_LIST_GET_FIRST(space->chain);

	if (node->n_pending > 0 || node->n_pending_flushes > 0) {
		/* There are pending i/o's or flushes, sleep for a while and
		retry */

		mutex_exit(&(system->mutex));

		os_thread_sleep(20000);

		goto retry;

	} else if (node->modification_counter > node->flush_counter) {
		/* Flush the space */

		mutex_exit(&(system->mutex));

		os_thread_sleep(20000);

		fil_flush(id);

		goto retry;

	} else if (node->open) {
		/* Close the file */

		fil_node_close_file(node, system);
	}

	/* Check that the old name in the space is right */
	
	ut_a(strlen(old_name) < OS_FILE_MAX_PATH - 10);

	sprintf(old_path, "./%s.ibd", old_name);

	srv_normalize_path_for_win(old_path);

	ut_a(strcmp(space->name, old_path) == 0);
	ut_a(strcmp(node->name, old_path) == 0);

	/* Rename the tablespace and the node in the memory cache */
	
	ut_a(strlen(new_name) < OS_FILE_MAX_PATH - 10);

	sprintf(path, "./%s.ibd", new_name);

	srv_normalize_path_for_win(path);

	success = fil_rename_tablespace_in_mem(space, node, path);

	if (!success) {

		goto func_exit;	
	}

	success = os_file_rename(old_path, path);

	if (!success) {
		/* We have to revert the changes we made to the tablespace
		memory cache */

		ut_a(fil_rename_tablespace_in_mem(space, node, old_path));
	}
func_exit:
	space->stop_ios = FALSE;

	mutex_exit(&(system->mutex));

	return(success);
}

/***********************************************************************
Creates a new single-table tablespace to a database directory of MySQL.
Database directories are under the 'datadir' of MySQL. The datadir is the
directory of a running mysqld program. We can refer to it by simply the
path '.'. */

ulint
fil_create_new_single_table_tablespace(
/*===================================*/
				/* out: DB_SUCCESS or error code */
	ulint*	space_id,	/* out: space id */
	char*	tablename,	/* in: the table name in the usual
				databasename/tablename format of InnoDB */
	ulint	size)		/* in: the initial size of the tablespace file
				in pages, must be > 0 */
{
	os_file_t       file;
	ibool		ret;
	ulint		err;
	byte*		page;
	ibool		success;
	char		path[OS_FILE_MAX_PATH];

	ut_a(strlen(tablename) < OS_FILE_MAX_PATH - 10);

	sprintf(path, "./%s.ibd", tablename);

	srv_normalize_path_for_win(path);
	
	file = os_file_create(path, OS_FILE_CREATE, OS_FILE_NORMAL,
						    OS_DATA_FILE, &ret);
	if (ret == FALSE) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Error creating file %s.\n", path);

		/* The following call will print an error message */
		 
		err = os_file_get_last_error(TRUE);
		
		if (err == OS_FILE_ALREADY_EXISTS) {
		        fprintf(stderr,
"InnoDB: The file already exists though the corresponding table did not\n"
"InnoDB: exist in the InnoDB data dictionary. Have you moved InnoDB\n"
"InnoDB: .ibd files around without using the SQL commands\n"
"InnoDB: DISCARD TABLESPACE and IMPORT TABLESPACE, or did\n"
"InnoDB: mysqld crash in the middle of CREATE TABLE? You can\n"
"InnoDB: resolve the problem by removing the file %s\n"
"InnoDB: under the 'datadir' of MySQL.\n", path);

			return(DB_TABLESPACE_ALREADY_EXISTS);
		}

		if (err == OS_FILE_DISK_FULL) {

			return(DB_OUT_OF_FILE_SPACE);
		}

		return(DB_ERROR);
	}

	page = ut_malloc(UNIV_PAGE_SIZE);

	ret = os_file_set_size(path, file, size * UNIV_PAGE_SIZE, 0);
	
	if (!ret) {
		ut_free(page);
		os_file_close(file);
		os_file_delete(path);

		return(DB_OUT_OF_FILE_SPACE);
	}

	*space_id = fil_assign_new_space_id();

	if (*space_id == ULINT_UNDEFINED) {
		ut_free(page);
		os_file_close(file);
		os_file_delete(path);

		return(DB_ERROR);
	}

	/* We have to write the space id to the file immediately and flush the
	file to disk. This is because in crash recovery we must be aware what
	tablespaces exist and what are their space id's, so that we can apply
	the log records to the right file. It may take quite a while until
	buffer pool flush algorithms write anything to the file and flush it to
	disk. If we would not write here anything, the file would be filled
	with zeros from the call of os_file_set_size(), until a buffer pool
	flush would write to it. */

	memset(page, '\0', UNIV_PAGE_SIZE);

	fsp_header_write_space_id(page, *space_id);		

	buf_flush_init_for_writing(page, ut_dulint_zero, *space_id, 0);

	ret = os_file_write(path, file, page, 0, 0, UNIV_PAGE_SIZE);

	ut_free(page);

	if (!ret) {
		fprintf(stderr,
"InnoDB: Error: could not write the first page to tablespace %s\n", path);

		os_file_close(file);
		os_file_delete(path);

		return(DB_ERROR);
	}

	ret = os_file_flush(file);

	if (!ret) {
		fprintf(stderr,
"InnoDB: Error: file flush of tablespace %s failed\n", path);

		os_file_close(file);
		os_file_delete(path);

		return(DB_ERROR);
	}

	os_file_close(file);

	if (*space_id == ULINT_UNDEFINED) {
		os_file_delete(path);

		return(DB_ERROR);
	}

	success = fil_space_create(path, *space_id, FIL_TABLESPACE);
	
	if (!success) {
		os_file_delete(path);

		return(DB_ERROR);
	}	

	fil_node_create(path, size, *space_id, FALSE);

	return(DB_SUCCESS);
}

/************************************************************************
Tries to open a single-table tablespace and checks the space id is right in
it. If does not succeed, prints an error message to the .err log. This
function is used to open the tablespace when we load a table definition
to the dictionary cache. NOTE that we assume this operation is used under the
protection of the dictionary mutex, so that two users cannot race here. This
operation does not leave the file associated with the tablespace open, but
closes it after we have looked at the space id in it. */

ibool
fil_open_single_table_tablespace(
/*=============================*/
			/* out: TRUE if success */
	ulint	id,	/* in: space id */
	char*	name)	/* in: table name in the databasename/tablename
			format */
{
	os_file_t	file;
	char*		filepath;
	ibool		success;
	byte*		page;
	ulint		space_id;
	ibool		ret		= TRUE;

	filepath = ut_malloc(OS_FILE_MAX_PATH);

	ut_a(strlen(name) < OS_FILE_MAX_PATH - 10);

	sprintf(filepath, "./%s.ibd", name);
					
	srv_normalize_path_for_win(filepath);

	file = os_file_create_simple_no_error_handling(filepath, OS_FILE_OPEN,
						OS_FILE_READ_ONLY, &success);
	if (!success) {
		/* The following call prints an error message */
		os_file_get_last_error(TRUE);

		ut_print_timestamp(stderr);

	        fprintf(stderr,
"  InnoDB: Error: trying to open a table, but could not\n"
"InnoDB: open the tablespace file %s!\n", filepath);
		fprintf(stderr,
"InnoDB: have you moved InnoDB .ibd files around without using the\n"
"InnoDB: commands DISCARD TABLESPACE and IMPORT TABLESPACE?\n");

		ut_free(filepath);

		return(FALSE);
	}

	/* Read the first page of the tablespace */

	page = ut_malloc(UNIV_PAGE_SIZE);

	success = os_file_read(file, page, 0, 0, UNIV_PAGE_SIZE);

	/* We have to read the tablespace id from the file */

	space_id = fsp_header_get_space_id(page);

	if (space_id != id) {
		ut_print_timestamp(stderr);

	        fprintf(stderr,
"  InnoDB: Error: tablespace id in file %s is %lu, but in the InnoDB\n"
"InnoDB: data dictionary it is %lu.\n", filepath, space_id, id);
		fprintf(stderr,
"InnoDB: Have you moved InnoDB .ibd files around without using the\n"
"InnoDB: commands DISCARD TABLESPACE and IMPORT TABLESPACE?\n");

		ret = FALSE;

		goto func_exit;
	}

	success = fil_space_create(filepath, space_id, FIL_TABLESPACE);

	if (!success) {
		goto func_exit;
	}

	/* We do not measure the size of the file, that is why we pass the 0
	below */

	fil_node_create(filepath, 0, space_id, FALSE);
func_exit:
	os_file_close(file);
	ut_free(page);
	ut_free(filepath);

	return(ret);
}

/************************************************************************
Opens an .ibd file and adds the associated single-table tablespace to the
InnoDB fil0fil.c data structures. */
static
void
fil_load_single_table_tablespace(
/*=============================*/
	char*	dbname,		/* in: database name */
	char*	filename)	/* in: file name (not a path), including the
				.ibd extension */
{
	os_file_t	file;
	char*		filepath;
	ibool		success;
	byte*		page;
	ulint		space_id;
	ulint		size_low;
	ulint		size_high;
	ib_longlong	size;

	filepath = ut_malloc(OS_FILE_MAX_PATH);

	ut_a(strlen(dbname) + strlen(filename) < OS_FILE_MAX_PATH - 10);

	sprintf(filepath, "./%s/%s", dbname, filename);
					
	srv_normalize_path_for_win(filepath);

	file = os_file_create_simple_no_error_handling(filepath, OS_FILE_OPEN,
						OS_FILE_READ_ONLY, &success);
	if (!success) {
		/* The following call prints an error message */
		os_file_get_last_error(TRUE);

	        fprintf(stderr,
"InnoDB: Error: could not open single-table tablespace file\n"
"InnoDB: %s!", filepath);

		ut_free(filepath);

		return;
	}

	success = os_file_get_size(file, &size_low, &size_high);

	if (!success) {
		/* The following call prints an error message */
		os_file_get_last_error(TRUE);

	        fprintf(stderr,
"InnoDB: Error: could not measure the size of single-table tablespace file\n"
"InnoDB: %s!", filepath);

		os_file_close(file);
		ut_free(filepath);

		return;
	}

	size = (((ib_longlong)size_high) << 32) + (ib_longlong)size_low;

	if (size < 4 * UNIV_PAGE_SIZE) {
	        fprintf(stderr,
"InnoDB: Error: the size of single-table tablespace file %s\n"
"InnoDB: is only %lu %lu, should be at least %lu!", filepath, size_high,
				       size_low, (ulint)4 * UNIV_PAGE_SIZE);
		os_file_close(file);
		ut_free(filepath);

		return;
	}

	/* Read the first page of the tablespace */

	page = ut_malloc(UNIV_PAGE_SIZE);

	success = os_file_read(file, page, 0, 0, UNIV_PAGE_SIZE);

	/* We have to read the tablespace id from the file */

	space_id = fsp_header_get_space_id(page);

	if (space_id == ULINT_UNDEFINED || space_id == 0) {
	        fprintf(stderr,
"InnoDB: Error: tablespace id %lu in file %s is not sensible\n", space_id,
								 filepath);
		goto func_exit;
	}

	success = fil_space_create(filepath, space_id, FIL_TABLESPACE);

	if (!success) {

		goto func_exit;
	}

	/* We do not measure the size of the file, that is why we pass the 0
	below */

	fil_node_create(filepath, 0, space_id, FALSE);
func_exit:
	os_file_close(file);
	ut_free(page);
	ut_free(filepath);
}

/************************************************************************
At the server startup, if we need crash recovery, scans the database
directories under the MySQL datadir, looking for .ibd files. Those files are
single-table tablespaces. We need to know the space id in each of them so that
we know into which file we should look to check the contents of a page stored
in the doublewrite buffer, also to know where to apply log records where the
space id is != 0. */

ulint
fil_load_single_table_tablespaces(void)
/*===================================*/
			/* out: DB_SUCCESS or error number */
{
	int		ret;
	char*		dbpath;
	os_file_dir_t	dir;
	os_file_dir_t	dbdir;
	os_file_stat_t	dbinfo;
	os_file_stat_t	fileinfo;

	/* The datadir of MySQL is always the default directory of mysqld */

	dir = os_file_opendir((char*)".", TRUE);

	if (dir == NULL) {

		return(DB_ERROR);
	}

	dbpath = ut_malloc(OS_FILE_MAX_PATH);

	/* Scan all directories under the datadir. They are the database
	directories of MySQL. */

	ret = os_file_readdir_next_file((char*)".", dir, &dbinfo);

	while (ret == 0) {
		/* printf("Looking at %s in datadir\n", dbinfo.name); */

		if (dbinfo.type == OS_FILE_TYPE_FILE
		    || dbinfo.type == OS_FILE_TYPE_UNKNOWN) {

		        goto next_datadir_item;
		}

		/* We found a symlink or a directory; try opening it to see
		if a symlink is a directory */
		
		ut_a(strlen(dbinfo.name) < OS_FILE_MAX_PATH - 10);

		sprintf(dbpath, "./%s", dbinfo.name);

		srv_normalize_path_for_win(dbpath);

		dbdir = os_file_opendir(dbpath, FALSE);

		if (dbdir != NULL) {
			/* printf("Opened dir %s\n", dbinfo.name); */

			/* We found a database directory; loop through it,
			looking for possible .ibd files in it */

			ret = os_file_readdir_next_file(dbpath, dbdir,
								&fileinfo);
			while (ret == 0) {
				/* printf(
"     Looking at file %s\n", fileinfo.name); */

			        if (fileinfo.type == OS_FILE_TYPE_DIR
				    || dbinfo.type == OS_FILE_TYPE_UNKNOWN) {
				        goto next_file_item;
				}

				/* We found a symlink or a file */
				if (strlen(fileinfo.name) > 4
				    && 0 == strcmp(fileinfo.name + 
						strlen(fileinfo.name) - 4,
						".ibd")) {
				        /* The name ends in .ibd; try opening
					the file */
					fil_load_single_table_tablespace(
						dbinfo.name, fileinfo.name);
				}
next_file_item:
				ret = os_file_readdir_next_file(dbpath, dbdir,
								&fileinfo);
			}

			if (0 != os_file_closedir(dbdir)) {
				 fprintf(stderr,
"InnoDB: Warning: could not close database directory %s\n", dbpath);
			}
		}
		
next_datadir_item:
		ret = os_file_readdir_next_file((char*)".", dir, &dbinfo);
	}

	ut_free(dbpath);

	/* At the end of directory we should get 1 as the return value, -1
	if there was an error */
	if (ret != 1) {
		fprintf(stderr,
"InnoDB: Error: os_file_readdir_next_file returned %d in MySQL datadir\n",
							       ret);
		os_file_closedir(dir);

		return(DB_ERROR);
	}

	if (0 != os_file_closedir(dir)) {
		fprintf(stderr,
"InnoDB: Error: could not close MySQL datadir\n");

		return(DB_ERROR);
	}

	return(DB_SUCCESS);
}

/************************************************************************
If we need crash recovery, and we have called
fil_load_single_table_tablespaces() and dict_load_single_table_tablespaces(),
we can call this function to print an error message of orphaned .ibd files
for which there is not a data dictionary entry with a matching table name
and space id. */

void
fil_print_orphaned_tablespaces(void)
/*================================*/
{
	fil_system_t*	system 		= fil_system;
	fil_space_t*	space;

	mutex_enter(&(system->mutex));

	space = UT_LIST_GET_FIRST(system->space_list);

	while (space) {
	        if (space->purpose == FIL_TABLESPACE && space->id != 0
							  && !space->mark) {
			fprintf(stderr,
"InnoDB: Warning: tablespace %s of id %lu has no matching table in\n"
"InnoDB: the InnoDB data dixtionary.\n", space->name, space->id);
		}

		space = UT_LIST_GET_NEXT(space_list, space);

	}

	mutex_exit(&(system->mutex));	
}

/***********************************************************************
Returns TRUE if a single-table tablespace does not exist in the memory cache,
or is being deleted there. */

ibool
fil_tablespace_deleted_or_being_deleted_in_mem(
/*===========================================*/
				/* out: TRUE if does not exist or is being\
				deleted */
	ulint		id,	/* in: space id */
	ib_longlong	version)/* in: tablespace_version should be this; if
				you pass -1 as the value of this, then this
				parameter is ignored */
{
	fil_system_t*	system	= fil_system;
	fil_space_t*	space;

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL || space->is_being_deleted) {
		mutex_exit(&(system->mutex));

		return(TRUE);
	}

	if (version != -1LL && space->tablespace_version != version) {
		mutex_exit(&(system->mutex));

		return(TRUE);
	}

	mutex_exit(&(system->mutex));

	return(FALSE);
}

/***********************************************************************
Returns TRUE if a single-table tablespace exists in the memory cache. */

ibool
fil_tablespace_exists_in_mem(
/*=========================*/
			/* out: TRUE if exists */
	ulint	id)	/* in: space id */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	space;

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	if (space == NULL) {
		mutex_exit(&(system->mutex));

		return(FALSE);
	}

	mutex_exit(&(system->mutex));

	return(TRUE);
}

/***********************************************************************
Returns TRUE if a matching tablespace exists in the InnoDB tablespace memory
cache. Note that if we have not done a crash recovery at the database startup,
there may be many tablespaces which are not yet in the memory cache. */

ibool
fil_space_for_table_exists_in_mem(
/*==============================*/
				/* out: TRUE if a matching tablespace
				exists in the memory cache */
	ulint	id,		/* in: space id */
	char*	name,		/* in: table name in the standard
				'databasename/tablename' format */
	ibool	mark_space,	/* in: in crash recovery, at database startup
				we mark all spaces which have an associated
				table in the InnoDB data dictionary, so that
				we can print a warning about orphaned
				tablespaces */
	ibool	print_error_if_does_not_exist)
				/* in: print detailed error information to
				the .err log if a matching tablespace is
				not found from memory */
{
	fil_system_t*	system		= fil_system;
	fil_space_t*	namespace;
	fil_space_t*	space;
	char		path[OS_FILE_MAX_PATH];

	ut_ad(system);

	mutex_enter(&(system->mutex));

	sprintf(path, "./%s.ibd", name);
	srv_normalize_path_for_win(path);

	/* Look if there is a space with the same id */

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

	/* Look if there is a space with the same name; the name is the
	directory path from the datadir to the file */

	HASH_SEARCH(name_hash, system->name_hash,
					ut_fold_string(path), namespace,
					0 == strcmp(namespace->name, path));
	if (!print_error_if_does_not_exist) {
		if (space && space == namespace) {
			if (mark_space) {
				space->mark = TRUE;
			}

			mutex_exit(&(system->mutex));

			return(TRUE);
		}
		
		mutex_exit(&(system->mutex));
		
		return(FALSE);
	}

	if (space == NULL) {
		if (namespace == NULL) {
		        ut_print_timestamp(stderr);
			fprintf(stderr,
"  InnoDB: Error: table %s in InnoDB data dictionary has tablespace\n"
"InnoDB: id %lu, but tablespace with that id or name does not exist. Have\n"
"InnoDB: you deleted or moved .ibd files? We cannot open table %s now.\n",
			       name, id, name);
		} else {
		        ut_print_timestamp(stderr);
			fprintf(stderr,
"  InnoDB: Error: table %s in InnoDB data dictionary has tablespace\n"
"InnoDB: id %lu, but tablespace with that id does not exist. There is\n"
"InnoDB: a tablespace of name %s and id %lu, though. Have\n"
"InnoDB: you deleted or moved .ibd files? We cannot open table %s now.\n",
			       name, id, namespace->name, namespace->id, name);
		}

		mutex_exit(&(system->mutex));

		return(FALSE);
	}

	if (0 != strcmp(space->name, path)) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Error: table %s in InnoDB data dictionary has tablespace\n"
"InnoDB: id %lu, but tablespace with that id has name %s. Have you\n"
"InnoDB: deleted or moved .ibd files? We cannot open table %s now.\n",
					name, id, space->name, name);
		if (namespace != NULL) {
			fprintf(stderr,
"InnoDB: There is a tablespace with the right name %s, but its id is %lu.\n",
					namespace->name, namespace->id);
		}

		mutex_exit(&(system->mutex));

		return(FALSE);
	}

	ut_a(space == namespace);

	if (mark_space) {
		space->mark = TRUE;
	}

	mutex_exit(&(system->mutex));

	return(TRUE);
}

/**************************************************************************
Tries to extend a data file by the number of pages given. Fractions of 1 MB
are ignored. The tablespace must be cached in the memory cache. */

ibool
fil_extend_last_data_file(
/*======================*/
				/* out: TRUE if success, also if we run
				out of disk space we may return TRUE */
	ulint*	actual_increase,/* out: number of pages we were able to
				extend, here the original size of the file and
				the resulting size of the file are rounded
				downwards to a full megabyte, and the
				difference expressed in pages is returned */
	ulint	space_id,	/* in: space id */
	ulint	size,		/* in: current size of the space in pages, as
				stored in the fsp header */
	ulint	size_increase)	/* in: try to extend this many pages */
{
	fil_system_t*	system		= fil_system;
	fil_node_t*	node;
	fil_space_t*	space;
	byte*		buf2;
	byte*		buf;
	ibool		success;
	ulint		i;

	fil_mutex_enter_and_prepare_for_io(space_id);

	HASH_SEARCH(hash, system->spaces, space_id, space,
						space->id == space_id);
	ut_a(space);
	
	node = UT_LIST_GET_LAST(space->chain);

	fil_node_prepare_for_io(node, system, space);

	if (UT_LIST_GET_LEN(space->chain) == 1 && node->size < size) {
	        ut_print_timestamp(stderr);
		fprintf(stderr,
"InnoDB: Fatal error: space %s id %lu size stored in header is %lu pages\n"
"InnoDB: but actual size is only %lu pages (possibly rounded downwards)!\n"
"InnoDB: Cannot continue operation!\n", space->name, space->id, size,
								node->size);
		exit(1);
	}

	buf2 = mem_alloc(1024 * 1024 + UNIV_PAGE_SIZE);
	buf = ut_align(buf2, UNIV_PAGE_SIZE);

	memset(buf, '\0', 1024 * 1024);

	for (i = 0; i < size_increase / ((1024 * 1024) / UNIV_PAGE_SIZE);
									i++) {
		/* If we use native Windows aio, then we use it also in this
		write */

		success = os_aio(OS_FILE_WRITE, OS_AIO_SYNC,
			node->name, node->handle, buf,
			(node->size << UNIV_PAGE_SIZE_SHIFT) & 0xFFFFFFFFUL,
			node->size >> (32 - UNIV_PAGE_SIZE_SHIFT),
			1024 * 1024, NULL, NULL);

		if (!success) {
			break;
		}

		node->size += ((1024 * 1024) / UNIV_PAGE_SIZE);
		space->size += ((1024 * 1024) / UNIV_PAGE_SIZE);

		os_has_said_disk_full = FALSE;
	}

	mem_free(buf2);

	fil_node_complete_io(node, system, OS_FILE_WRITE);

	mutex_exit(&(system->mutex));	

	*actual_increase = i * ((1024 * 1024) / UNIV_PAGE_SIZE);

	fil_flush(space_id);

	if (space_id == 0) {
	        srv_data_file_sizes[srv_n_data_files - 1] += *actual_increase;
	}

	return(TRUE);
2565 2566
}

unknown's avatar
unknown committed
2567 2568 2569
/**************************************************************************
Tries to extend a data file so that it would accommodate the number of pages
given. The tablespace must be cached in the memory cache. */
2570 2571

ibool
unknown's avatar
unknown committed
2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582
fil_extend_data_file_with_pages(
/*============================*/
				/* out: TRUE if success */
	ulint	space_id,	/* in: space id, must be != 0 */
	ulint	size,		/* in: current size of the space in pages, as
				stored in the fsp header */
	ulint	size_after_extend)/* in: desired size in pages after the
				extension, should be less than 4 GB (this
				function is primarily intended for increasing
				the data file size from < 64 pages to up to
				64 pages) */
2583 2584
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599
	fil_node_t*	node;
	fil_space_t*	space;
	byte*		buf2;
	byte*		buf;
	ibool		success;

	ut_a(space_id != 0);
	ut_a(size_after_extend < 64 * 4096);
	ut_a(size_after_extend >= size);

	fil_mutex_enter_and_prepare_for_io(space_id);

	HASH_SEARCH(hash, system->spaces, space_id, space,
						space->id == space_id);
	ut_a(space);
2600
	
unknown's avatar
unknown committed
2601
	node = UT_LIST_GET_LAST(space->chain);
2602

unknown's avatar
unknown committed
2603
	fil_node_prepare_for_io(node, system, space);
2604

unknown's avatar
unknown committed
2605 2606 2607 2608 2609 2610 2611 2612 2613
	if (UT_LIST_GET_LEN(space->chain) == 1 && node->size < size) {
	        ut_print_timestamp(stderr);
		fprintf(stderr,
"InnoDB: Fatal error: space %s id %lu size stored in header is %lu pages\n"
"InnoDB: but actual size is only %lu pages (possibly rounded downwards)!\n"
"InnoDB: Cannot continue operation!\n", space->name, space_id, size,
								node->size);
		exit(1);
	}
2614

unknown's avatar
unknown committed
2615 2616
	buf2 = mem_alloc((1 + size_after_extend - size) * UNIV_PAGE_SIZE);
	buf = ut_align(buf2, UNIV_PAGE_SIZE);
2617

unknown's avatar
unknown committed
2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629
	memset(buf, '\0', (size_after_extend - size) * UNIV_PAGE_SIZE);

	success = os_aio(OS_FILE_WRITE, OS_AIO_SYNC,
			node->name, node->handle, buf,
			UNIV_PAGE_SIZE * size, 0,
			UNIV_PAGE_SIZE * (size_after_extend - size),
			NULL, NULL);
	if (success) {
		node->size = size_after_extend;
		space->size = size_after_extend;

		os_has_said_disk_full = FALSE;
2630 2631
	}

unknown's avatar
unknown committed
2632 2633 2634 2635 2636 2637 2638 2639 2640
	mem_free(buf2);

	fil_node_complete_io(node, system, OS_FILE_WRITE);

	mutex_exit(&(system->mutex));	

	fil_flush(space_id);

	return(success);
2641 2642
}

unknown's avatar
unknown committed
2643 2644
/*========== RESERVE FREE EXTENTS (for a B-tree split, for example) ===*/

2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656
/***********************************************************************
Tries to reserve free extents in a file space. */

ibool
fil_space_reserve_free_extents(
/*===========================*/
				/* out: TRUE if succeed */
	ulint	id,		/* in: space id */
	ulint	n_free_now,	/* in: number of free extents now */
	ulint	n_to_reserve)	/* in: how many one wants to reserve */
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
2657
	fil_space_t*	space;
2658 2659 2660 2661 2662 2663 2664 2665
	ibool		success;

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
2666 2667
	ut_a(space);

2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689
	if (space->n_reserved_extents + n_to_reserve > n_free_now) {
		success = FALSE;
	} else {
		space->n_reserved_extents += n_to_reserve;
		success = TRUE;
	}
	
	mutex_exit(&(system->mutex));

	return(success);
}

/***********************************************************************
Releases free extents in a file space. */

void
fil_space_release_free_extents(
/*===========================*/
	ulint	id,		/* in: space id */
	ulint	n_reserved)	/* in: how many one reserved */
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
2690
	fil_space_t*	space;
2691 2692 2693 2694 2695 2696 2697

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);

unknown's avatar
unknown committed
2698
	ut_a(space);
2699 2700 2701 2702 2703 2704 2705
	ut_a(space->n_reserved_extents >= n_reserved);
	
	space->n_reserved_extents -= n_reserved;
	
	mutex_exit(&(system->mutex));
}

2706 2707 2708 2709 2710 2711 2712 2713 2714 2715
/***********************************************************************
Gets the number of reserved extents. If the database is silent, this number
should be zero. */

ulint
fil_space_get_n_reserved_extents(
/*=============================*/
	ulint	id)		/* in: space id */
{
	fil_system_t*	system		= fil_system;
unknown's avatar
unknown committed
2716
	fil_space_t*	space;
2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733
	ulint		n;

	ut_ad(system);

	mutex_enter(&(system->mutex));

	HASH_SEARCH(hash, system->spaces, id, space, space->id == id);
	
	ut_a(space);

	n = space->n_reserved_extents;
	
	mutex_exit(&(system->mutex));

	return(n);
}

unknown's avatar
unknown committed
2734 2735
/*============================ FILE I/O ================================*/

2736
/************************************************************************
unknown's avatar
unknown committed
2737 2738
NOTE: you must call fil_mutex_enter_and_prepare_for_io() first!

2739 2740
Prepares a file node for i/o. Opens the file if it is closed. Updates the
pending i/o's field in the node and the system appropriately. Takes the node
unknown's avatar
unknown committed
2741 2742
off the LRU list if it is in the LRU list. The caller must hold the fil_sys
mutex. */
2743 2744 2745 2746 2747
static
void
fil_node_prepare_for_io(
/*====================*/
	fil_node_t*	node,	/* in: file node */
unknown's avatar
unknown committed
2748
	fil_system_t*	system,	/* in: tablespace memory cache */
2749 2750 2751 2752 2753
	fil_space_t*	space)	/* in: space */
{
	ut_ad(node && system && space);
	ut_ad(mutex_own(&(system->mutex)));
	
unknown's avatar
unknown committed
2754 2755 2756 2757 2758 2759 2760
	if (system->n_open > system->max_n_open + 5) {
		ut_print_timestamp(stderr);
		fprintf(stderr,
"  InnoDB: Warning: open files %lu exceeds the limit %lu\n",
					system->n_open, system->max_n_open);
	}

2761
	if (node->open == FALSE) {
unknown's avatar
unknown committed
2762
		/* File is closed: open it */
2763 2764
		ut_a(node->n_pending == 0);

unknown's avatar
unknown committed
2765
		fil_node_open_file(node, system, space);
2766 2767
	}

unknown's avatar
unknown committed
2768 2769
	if (node->n_pending == 0 && space->purpose == FIL_TABLESPACE
						      && space->id != 0) {
2770 2771
		/* The node is in the LRU list, remove it */

unknown's avatar
unknown committed
2772
		ut_a(UT_LIST_GET_LEN(system->LRU) > 0);
2773

unknown's avatar
unknown committed
2774
		UT_LIST_REMOVE(LRU, system->LRU, node);
2775
	}
unknown's avatar
unknown committed
2776 2777

	node->n_pending++;
2778 2779 2780 2781
}

/************************************************************************
Updates the data structures when an i/o operation finishes. Updates the
unknown's avatar
unknown committed
2782
pending i/o's field in the node appropriately. */
2783 2784 2785 2786 2787
static
void
fil_node_complete_io(
/*=================*/
	fil_node_t*	node,	/* in: file node */
unknown's avatar
unknown committed
2788 2789 2790 2791
	fil_system_t*	system,	/* in: tablespace memory cache */
	ulint		type)	/* in: OS_FILE_WRITE or OS_FILE_READ; marks
				the node as modified if
				type == OS_FILE_WRITE */
2792 2793 2794 2795
{
	ut_ad(node);
	ut_ad(system);
	ut_ad(mutex_own(&(system->mutex)));
unknown's avatar
unknown committed
2796

2797 2798 2799 2800
	ut_a(node->n_pending > 0);
	
	node->n_pending--;

unknown's avatar
unknown committed
2801 2802 2803
	if (type == OS_FILE_WRITE) {
		system->modification_counter++;
		node->modification_counter = system->modification_counter;
2804 2805
	}
	
unknown's avatar
unknown committed
2806 2807
	if (node->n_pending == 0 && node->space->purpose == FIL_TABLESPACE
					&& node->space->id != 0) {
2808 2809
		/* The node must be put back to the LRU list */
		UT_LIST_ADD_FIRST(LRU, system->LRU, node);
unknown's avatar
unknown committed
2810 2811 2812
	}
}

2813 2814 2815
/************************************************************************
Reads or writes data. This operation is asynchronous (aio). */

unknown's avatar
unknown committed
2816
ulint
2817 2818
fil_io(
/*===*/
unknown's avatar
unknown committed
2819 2820 2821
				/* out: DB_SUCCESS, or DB_TABLESPACE_DELETED
				if we are trying to do i/o on a tablespace
				which does not exist */
2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836
	ulint	type,		/* in: OS_FILE_READ or OS_FILE_WRITE,
				ORed to OS_FILE_LOG, if a log i/o
				and ORed to OS_AIO_SIMULATED_WAKE_LATER
				if simulated aio and we want to post a
				batch of i/os; NOTE that a simulated batch
				may introduce hidden chances of deadlocks,
				because i/os are not actually handled until
				all have been posted: use with great
				caution! */
	ibool	sync,		/* in: TRUE if synchronous aio is desired */
	ulint	space_id,	/* in: space id */
	ulint	block_offset,	/* in: offset in number of blocks */
	ulint	byte_offset,	/* in: remainder of offset in bytes; in
				aio this must be divisible by the OS block
				size */
unknown's avatar
unknown committed
2837 2838 2839
	ulint	len,		/* in: how many bytes to read or write; this
				must not cross a file boundary; in aio this
				must be a block size multiple */
2840 2841 2842 2843 2844 2845
	void*	buf,		/* in/out: buffer where to store read data
				or from where to write; in aio this must be
				appropriately aligned */
	void*	message)	/* in: message for aio handler if non-sync
				aio used, else ignored */
{
unknown's avatar
unknown committed
2846
	fil_system_t*	system		= fil_system;
2847 2848 2849 2850 2851 2852 2853 2854
	ulint		mode;
	fil_space_t*	space;
	fil_node_t*	node;
	ulint		offset_high;
	ulint		offset_low;
	ibool		ret;
	ulint		is_log;
	ulint		wake_later;
2855
	
2856 2857 2858 2859 2860 2861 2862 2863 2864
	is_log = type & OS_FILE_LOG;
	type = type & ~OS_FILE_LOG;

	wake_later = type & OS_AIO_SIMULATED_WAKE_LATER;
	type = type & ~OS_AIO_SIMULATED_WAKE_LATER;
	
	ut_ad(byte_offset < UNIV_PAGE_SIZE);
	ut_ad(buf);
	ut_ad(len > 0);
unknown's avatar
unknown committed
2865
	ut_a((1 << UNIV_PAGE_SIZE_SHIFT) == UNIV_PAGE_SIZE);
2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877
	ut_ad(fil_validate());
#ifndef UNIV_LOG_DEBUG
	/* ibuf bitmap pages must be read in the sync aio mode: */
	ut_ad(recv_no_ibuf_operations || (type == OS_FILE_WRITE)
		|| !ibuf_bitmap_page(block_offset) || sync || is_log);
#ifdef UNIV_SYNC_DEBUG
	ut_ad(!ibuf_inside() || is_log || (type == OS_FILE_WRITE)
					|| ibuf_page(space_id, block_offset));
#endif
#endif
	if (sync) {
		mode = OS_AIO_SYNC;
2878
	} else if (type == OS_FILE_READ && !is_log
2879 2880 2881 2882 2883 2884 2885 2886
				&& ibuf_page(space_id, block_offset)) {
		mode = OS_AIO_IBUF;
	} else if (is_log) {
		mode = OS_AIO_LOG;
	} else {
		mode = OS_AIO_NORMAL;
	}

unknown's avatar
unknown committed
2887 2888
	/* Reserve the fil_system mutex and make sure that we can open at
	least one file while holding it, if the file is not already open */
2889

unknown's avatar
unknown committed
2890
	fil_mutex_enter_and_prepare_for_io(space_id);
2891
	
unknown's avatar
unknown committed
2892 2893 2894
	HASH_SEARCH(hash, system->spaces, space_id, space,
							space->id == space_id);
	if (!space) {
2895 2896
		mutex_exit(&(system->mutex));

unknown's avatar
unknown committed
2897
		ut_print_timestamp(stderr);
2898
		fprintf(stderr,
unknown's avatar
unknown committed
2899 2900 2901
"  InnoDB: Error: trying to do i/o to a tablespace which does not exist.\n"
"InnoDB: i/o type %lu, space id %lu, page no. %lu, i/o length %lu bytes\n",
		       type, space_id, block_offset, len);
2902

unknown's avatar
unknown committed
2903 2904
		return(DB_TABLESPACE_DELETED);
	}
2905 2906 2907 2908 2909 2910

	ut_ad((mode != OS_AIO_IBUF) || (space->purpose == FIL_TABLESPACE));

	node = UT_LIST_GET_FIRST(space->chain);

	for (;;) {
unknown's avatar
unknown committed
2911 2912 2913 2914 2915 2916 2917
		if (space->id != 0 && node->size == 0) {
			/* We do not know the size of a single-table tablespace
			before we open the file */

			break;
		}

unknown's avatar
unknown committed
2918 2919
		if (node == NULL) {
			fprintf(stderr,
unknown's avatar
unknown committed
2920 2921
	"InnoDB: Error: trying to access page number %lu in space %lu,\n"
	"InnoDB: space name %s,\n"
unknown's avatar
unknown committed
2922 2923
	"InnoDB: which is outside the tablespace bounds.\n"
	"InnoDB: Byte offset %lu, len %lu, i/o type %lu\n", 
unknown's avatar
unknown committed
2924 2925
 			block_offset, space_id, space->name, byte_offset, len,
			type);
unknown's avatar
unknown committed
2926 2927 2928
 			
			ut_a(0);
		}
2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941

		if (node->size > block_offset) {
			/* Found! */
			break;
		} else {
			block_offset -= node->size;
			node = UT_LIST_GET_NEXT(chain, node);
		}
	}		
	
	/* Open file if closed */
	fil_node_prepare_for_io(node, system, space);

unknown's avatar
unknown committed
2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956
	/* Check that at least the start offset is within the bounds of a
	single-table tablespace */
	if (space->purpose == FIL_TABLESPACE && space->id != 0
	    && node->size <= block_offset) {

	        fprintf(stderr,
	"InnoDB: Error: trying to access page number %lu in space %lu,\n"
	"InnoDB: space name %s,\n"
	"InnoDB: which is outside the tablespace bounds.\n"
	"InnoDB: Byte offset %lu, len %lu, i/o type %lu\n", 
 			block_offset, space_id, space->name, byte_offset, len,
			type);
 		ut_a(0);
	}

2957 2958 2959 2960 2961 2962
	/* Now we have made the changes in the data structures of system */
	mutex_exit(&(system->mutex));

	/* Calculate the low 32 bits and the high 32 bits of the file offset */

	offset_high = (block_offset >> (32 - UNIV_PAGE_SIZE_SHIFT));
unknown's avatar
unknown committed
2963
	offset_low  = ((block_offset << UNIV_PAGE_SIZE_SHIFT) & 0xFFFFFFFFUL)
2964 2965 2966 2967 2968 2969 2970
			+ byte_offset;

	ut_a(node->size - block_offset >=
 		(byte_offset + len + (UNIV_PAGE_SIZE - 1)) / UNIV_PAGE_SIZE);

	/* Do aio */

unknown's avatar
unknown committed
2971 2972
	ut_anp(byte_offset % OS_FILE_LOG_BLOCK_SIZE == 0);
	ut_anp((len % OS_FILE_LOG_BLOCK_SIZE) == 0);
2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984

	/* Queue the aio request */
	ret = os_aio(type, mode | wake_later, node->name, node->handle, buf,
				offset_low, offset_high, len, node, message);
	ut_a(ret);

	if (mode == OS_AIO_SYNC) {
		/* The i/o operation is already completed when we return from
		os_aio: */
		
		mutex_enter(&(system->mutex));

2985
		fil_node_complete_io(node, system, type);
2986 2987 2988 2989 2990

		mutex_exit(&(system->mutex));

		ut_ad(fil_validate());
	}
unknown's avatar
unknown committed
2991 2992

	return(DB_SUCCESS);
2993 2994 2995 2996 2997 2998 2999
}

/************************************************************************
Reads data from a space to a buffer. Remember that the possible incomplete
blocks at the end of file are ignored: they are not taken into account when
calculating the byte offset within a space. */

unknown's avatar
unknown committed
3000
ulint
3001 3002
fil_read(
/*=====*/
unknown's avatar
unknown committed
3003 3004 3005
				/* out: DB_SUCCESS, or DB_TABLESPACE_DELETED
				if we are trying to do i/o on a tablespace
				which does not exist */
3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018
	ibool	sync,		/* in: TRUE if synchronous aio is desired */
	ulint	space_id,	/* in: space id */
	ulint	block_offset,	/* in: offset in number of blocks */
	ulint	byte_offset,	/* in: remainder of offset in bytes; in aio
				this must be divisible by the OS block size */
	ulint	len,		/* in: how many bytes to read; this must not
				cross a file boundary; in aio this must be a
				block size multiple */
	void*	buf,		/* in/out: buffer where to store data read;
				in aio this must be appropriately aligned */
	void*	message)	/* in: message for aio handler if non-sync
				aio used, else ignored */
{
unknown's avatar
unknown committed
3019 3020
	return(fil_io(OS_FILE_READ, sync, space_id, block_offset,
					  byte_offset, len, buf, message));
3021 3022 3023 3024 3025 3026 3027
}

/************************************************************************
Writes data to a space from a buffer. Remember that the possible incomplete
blocks at the end of file are ignored: they are not taken into account when
calculating the byte offset within a space. */

unknown's avatar
unknown committed
3028
ulint
3029 3030
fil_write(
/*======*/
unknown's avatar
unknown committed
3031 3032 3033
				/* out: DB_SUCCESS, or DB_TABLESPACE_DELETED
				if we are trying to do i/o on a tablespace
				which does not exist */
3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046
	ibool	sync,		/* in: TRUE if synchronous aio is desired */
	ulint	space_id,	/* in: space id */
	ulint	block_offset,	/* in: offset in number of blocks */
	ulint	byte_offset,	/* in: remainder of offset in bytes; in aio
				this must be divisible by the OS block size */
	ulint	len,		/* in: how many bytes to write; this must
				not cross a file boundary; in aio this must
				be a block size multiple */
	void*	buf,		/* in: buffer from which to write; in aio
				this must be appropriately aligned */
	void*	message)	/* in: message for aio handler if non-sync
				aio used, else ignored */
{
unknown's avatar
unknown committed
3047 3048
	return(fil_io(OS_FILE_WRITE, sync, space_id, block_offset,
					   byte_offset, len, buf, message));
3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062
}

/**************************************************************************
Waits for an aio operation to complete. This function is used to write the
handler for completed requests. The aio array of pending requests is divided
into segments (see os0file.c for more info). The thread specifies which
segment it wants to wait for. */

void
fil_aio_wait(
/*=========*/
	ulint	segment)	/* in: the number of the segment in the aio
				array to wait for */ 
{
unknown's avatar
unknown committed
3063
	fil_system_t*	system		= fil_system;
3064 3065 3066
	ibool		ret;		
	fil_node_t*	fil_node;
	void*		message;
3067
	ulint		type;
3068 3069 3070 3071
	
	ut_ad(fil_validate());

	if (os_aio_use_native_aio) {
unknown's avatar
unknown committed
3072
		srv_io_thread_op_info[segment] = (char *) "handle native aio";
3073
#ifdef WIN_ASYNC_IO
3074 3075
		ret = os_aio_windows_handle(segment, 0, &fil_node, &message,
								&type);
3076 3077 3078
#elif defined(POSIX_ASYNC_IO)
		ret = os_aio_posix_handle(segment, &fil_node, &message);
#else
3079
		ret = 0; /* Eliminate compiler warning */
3080 3081 3082
		ut_a(0);
#endif
	} else {
unknown's avatar
unknown committed
3083
		srv_io_thread_op_info[segment] =(char *)"handle simulated aio";
3084

3085
		ret = os_aio_simulated_handle(segment, (void**) &fil_node,
3086
	                                               &message, &type);
3087 3088 3089
	}
	
	ut_a(ret);
3090

unknown's avatar
unknown committed
3091
	srv_io_thread_op_info[segment] = (char *) "complete io for fil node";
3092

3093 3094
	mutex_enter(&(system->mutex));

3095
	fil_node_complete_io(fil_node, fil_system, type);
3096 3097 3098 3099 3100 3101

	mutex_exit(&(system->mutex));

	ut_ad(fil_validate());

	/* Do the i/o handling */
unknown's avatar
unknown committed
3102 3103 3104 3105
	/* IMPORTANT: since i/o handling for reads will read also the insert
	buffer in tablespace 0, you have to be very careful not to introduce
	deadlocks in the i/o system. We keep tablespace 0 data files always
	open, and use a special i/o thread to serve insert buffer requests. */
3106 3107

	if (buf_pool_is_block(message)) {
unknown's avatar
unknown committed
3108 3109
		srv_io_thread_op_info[segment] =
		  (char *) "complete io for buf page";
3110 3111
		buf_page_io_complete(message);
	} else {
unknown's avatar
unknown committed
3112
		srv_io_thread_op_info[segment] =(char *) "complete io for log";
3113 3114 3115 3116 3117
		log_io_complete(message);
	}
}

/**************************************************************************
unknown's avatar
unknown committed
3118 3119
Flushes to disk possible writes cached by the OS. If the space does not exist
or is being dropped, does not do anything. */
3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130

void
fil_flush(
/*======*/
	ulint	space_id)	/* in: file space id (this can be a group of
				log files or a tablespace of the database) */
{
	fil_system_t*	system	= fil_system;
	fil_space_t*	space;
	fil_node_t*	node;
	os_file_t	file;
unknown's avatar
unknown committed
3131
	ib_longlong	old_mod_counter;
3132 3133 3134 3135

	mutex_enter(&(system->mutex));
	
	HASH_SEARCH(hash, system->spaces, space_id, space,
unknown's avatar
unknown committed
3136 3137 3138 3139 3140 3141
							space->id == space_id);
	if (!space || space->is_being_deleted) {
		mutex_exit(&(system->mutex));

		return;
	}
3142

unknown's avatar
unknown committed
3143 3144
	space->n_pending_flushes++;	/* prevent dropping of the space while
					we are flushing */
3145 3146 3147
	node = UT_LIST_GET_FIRST(space->chain);

	while (node) {
unknown's avatar
unknown committed
3148 3149 3150 3151 3152 3153
		if (node->modification_counter > node->flush_counter) {
			ut_a(node->open);

			/* We want to flush the changes at least up to
			old_mod_counter */
			old_mod_counter = node->modification_counter;
3154

3155 3156 3157 3158 3159
			if (space->purpose == FIL_TABLESPACE) {
				fil_n_pending_tablespace_flushes++;
			} else {
				fil_n_pending_log_flushes++;
			}
unknown's avatar
unknown committed
3160 3161
#ifdef __WIN__
			if (node->is_raw_disk) {
3162

unknown's avatar
unknown committed
3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182
				goto skip_flush;
			}
#endif
retry:			
			if (node->n_pending_flushes > 0) {
				/* We want to avoid calling os_file_flush() on
				the file twice at the same time, because we do
				not know what bugs OS's may contain in file
				i/o; sleep for a while */

				mutex_exit(&(system->mutex));

				os_thread_sleep(20000);

				mutex_enter(&(system->mutex));

				if (node->flush_counter >= old_mod_counter) {

					goto skip_flush;
				}
3183

unknown's avatar
unknown committed
3184 3185 3186 3187 3188 3189 3190 3191
				goto retry;
			}

			ut_a(node->open);
			file = node->handle;
			node->n_pending_flushes++;

			mutex_exit(&(system->mutex));
3192 3193

			/* printf("Flushing to file %s\n", node->name); */
unknown's avatar
unknown committed
3194 3195
			os_file_flush(file);		

3196
			mutex_enter(&(system->mutex));
3197

unknown's avatar
unknown committed
3198 3199 3200 3201 3202 3203
			node->n_pending_flushes--;
skip_flush:
			if (node->flush_counter < old_mod_counter) {
				node->flush_counter = old_mod_counter;
			}

3204 3205 3206 3207 3208
			if (space->purpose == FIL_TABLESPACE) {
				fil_n_pending_tablespace_flushes--;
			} else {
				fil_n_pending_log_flushes--;
			}
3209 3210 3211 3212 3213
		}

		node = UT_LIST_GET_NEXT(chain, node);
	}		

unknown's avatar
unknown committed
3214 3215
	space->n_pending_flushes--;

3216 3217 3218 3219
	mutex_exit(&(system->mutex));
}

/**************************************************************************
unknown's avatar
unknown committed
3220
Flushes to disk the writes in file spaces of the given type possibly cached by
3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236
the OS. */

void
fil_flush_file_spaces(
/*==================*/
	ulint	purpose)	/* in: FIL_TABLESPACE, FIL_LOG */
{
	fil_system_t*	system	= fil_system;
	fil_space_t*	space;

	mutex_enter(&(system->mutex));

	space = UT_LIST_GET_FIRST(system->space_list);

	while (space) {
		if (space->purpose == purpose) {
unknown's avatar
unknown committed
3237 3238 3239
			space->n_pending_flushes++; /* prevent dropping of the
						    space while we are
						    flushing */
3240 3241 3242 3243 3244 3245
			mutex_exit(&(system->mutex));

			fil_flush(space->id);

			mutex_enter(&(system->mutex));

unknown's avatar
unknown committed
3246 3247
			space->n_pending_flushes--;
		}
3248 3249 3250 3251 3252 3253 3254
		space = UT_LIST_GET_NEXT(space_list, space);
	}
	
	mutex_exit(&(system->mutex));
}

/**********************************************************************
unknown's avatar
unknown committed
3255
Checks the consistency of the tablespace cache. */
3256 3257 3258 3259 3260 3261

ibool
fil_validate(void)
/*==============*/
			/* out: TRUE if ok */
{	
unknown's avatar
unknown committed
3262
	fil_system_t*	system		= fil_system;
3263 3264
	fil_space_t*	space;
	fil_node_t*	fil_node;
unknown's avatar
unknown committed
3265
	ulint		n_open		= 0;
3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285
	ulint		i;
	
	mutex_enter(&(system->mutex));

	/* Look for spaces in the hash table */

	for (i = 0; i < hash_get_n_cells(system->spaces); i++) {

		space = HASH_GET_FIRST(system->spaces, i);
	
		while (space != NULL) {
			UT_LIST_VALIDATE(chain, fil_node_t, space->chain); 

			fil_node = UT_LIST_GET_FIRST(space->chain);

			while (fil_node != NULL) {
				if (fil_node->n_pending > 0) {
					ut_a(fil_node->open);
				}

unknown's avatar
unknown committed
3286 3287 3288
				if (fil_node->open) {
					n_open++;
				}
3289 3290 3291 3292 3293 3294
				fil_node = UT_LIST_GET_NEXT(chain, fil_node);
			}
			space = HASH_GET_NEXT(hash, space);
		}
	}

unknown's avatar
unknown committed
3295
	ut_a(system->n_open == n_open);
3296 3297 3298 3299 3300 3301 3302 3303

	UT_LIST_VALIDATE(LRU, fil_node_t, system->LRU);

	fil_node = UT_LIST_GET_FIRST(system->LRU);

	while (fil_node != NULL) {
		ut_a(fil_node->n_pending == 0);
		ut_a(fil_node->open);
unknown's avatar
unknown committed
3304 3305
		ut_a(fil_node->space->purpose == FIL_TABLESPACE);
		ut_a(fil_node->space->id != 0);
3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372

		fil_node = UT_LIST_GET_NEXT(LRU, fil_node);
	}
	
	mutex_exit(&(system->mutex));

	return(TRUE);
}

/************************************************************************
Returns TRUE if file address is undefined. */
ibool
fil_addr_is_null(
/*=============*/
				/* out: TRUE if undefined */
	fil_addr_t	addr)	/* in: address */
{
	if (addr.page == FIL_NULL) {

		return(TRUE);
	}

	return(FALSE);
}

/************************************************************************
Accessor functions for a file page */

ulint
fil_page_get_prev(byte*	page)
{
	return(mach_read_from_4(page + FIL_PAGE_PREV));
}

ulint
fil_page_get_next(byte*	page)
{
	return(mach_read_from_4(page + FIL_PAGE_NEXT));
}

/*************************************************************************
Sets the file page type. */

void
fil_page_set_type(
/*==============*/
	byte* 	page,	/* in: file page */
	ulint	type)	/* in: type */
{
	ut_ad(page);

	mach_write_to_2(page + FIL_PAGE_TYPE, type);
}	

/*************************************************************************
Gets the file page type. */

ulint
fil_page_get_type(
/*==============*/
			/* out: type; NOTE that if the type has not been
			written to page, the return value not defined */
	byte* 	page)	/* in: file page */
{
	ut_ad(page);

	return(mach_read_from_2(page + FIL_PAGE_TYPE));
unknown's avatar
unknown committed
3373
}