jfs_imap.c 85.8 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
/*
 *   Copyright (C) International Business Machines Corp., 2000-2004
 *
 *   This program is free software;  you can redistribute it and/or modify
 *   it under the terms of the GNU General Public License as published by
 *   the Free Software Foundation; either version 2 of the License, or 
 *   (at your option) any later version.
 * 
 *   This program is distributed in the hope that it will be useful,
 *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
 *   the GNU General Public License for more details.
 *
 *   You should have received a copy of the GNU General Public License
 *   along with this program;  if not, write to the Free Software 
 *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
 */

/*
 *	jfs_imap.c: inode allocation map manager
 *
 * Serialization:
 *   Each AG has a simple lock which is used to control the serialization of
 *	the AG level lists.  This lock should be taken first whenever an AG
 *	level list will be modified or accessed.
 *
 *   Each IAG is locked by obtaining the buffer for the IAG page.
 *
 *   There is also a inode lock for the inode map inode.  A read lock needs to
 *	be taken whenever an IAG is read from the map or the global level
 *	information is read.  A write lock needs to be taken whenever the global
 *	level information is modified or an atomic operation needs to be used.
 *
 *	If more than one IAG is read at one time, the read lock may not
 *	be given up until all of the IAG's are read.  Otherwise, a deadlock
 *	may occur when trying to obtain the read lock while another thread
 *	holding the read lock is waiting on the IAG already being held.
 *
 *   The control page of the inode map is read into memory by diMount().
 *	Thereafter it should only be modified in memory and then it will be
 *	written out when the filesystem is unmounted by diUnmount().
 */

#include <linux/fs.h>
#include <linux/buffer_head.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>

#include "jfs_incore.h"
50
#include "jfs_inode.h"
Linus Torvalds's avatar
Linus Torvalds committed
51
52
53
54
55
56
57
58
#include "jfs_filsys.h"
#include "jfs_dinode.h"
#include "jfs_dmap.h"
#include "jfs_imap.h"
#include "jfs_metapage.h"
#include "jfs_superblock.h"
#include "jfs_debug.h"

59
60
61
62
63
64
/*
 * __mark_inode_dirty expects inodes to be hashed.  Since we don't want
 * special inodes in the fileset inode space, we hash them to a dummy head
 */
static HLIST_HEAD(aggregate_hash);

Linus Torvalds's avatar
Linus Torvalds committed
65
66
67
68
/*
 * imap locks
 */
/* iag free list lock */
69
70
71
#define IAGFREE_LOCK_INIT(imap)		mutex_init(&imap->im_freelock)
#define IAGFREE_LOCK(imap)		mutex_lock(&imap->im_freelock)
#define IAGFREE_UNLOCK(imap)		mutex_unlock(&imap->im_freelock)
Linus Torvalds's avatar
Linus Torvalds committed
72
73

/* per ag iag list locks */
74
75
76
#define AG_LOCK_INIT(imap,index)	mutex_init(&(imap->im_aglock[index]))
#define AG_LOCK(imap,agno)		mutex_lock(&imap->im_aglock[agno])
#define AG_UNLOCK(imap,agno)		mutex_unlock(&imap->im_aglock[agno])
Linus Torvalds's avatar
Linus Torvalds committed
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
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
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267

/*
 * forward references
 */
static int diAllocAG(struct inomap *, int, boolean_t, struct inode *);
static int diAllocAny(struct inomap *, int, boolean_t, struct inode *);
static int diAllocBit(struct inomap *, struct iag *, int);
static int diAllocExt(struct inomap *, int, struct inode *);
static int diAllocIno(struct inomap *, int, struct inode *);
static int diFindFree(u32, int);
static int diNewExt(struct inomap *, struct iag *, int);
static int diNewIAG(struct inomap *, int *, int, struct metapage **);
static void duplicateIXtree(struct super_block *, s64, int, s64 *);

static int diIAGRead(struct inomap * imap, int, struct metapage **);
static int copy_from_dinode(struct dinode *, struct inode *);
static void copy_to_dinode(struct dinode *, struct inode *);

/*
 * NAME:        diMount()
 *
 * FUNCTION:    initialize the incore inode map control structures for
 *		a fileset or aggregate init time.
 *
 *              the inode map's control structure (dinomap) is 
 *              brought in from disk and placed in virtual memory.
 *
 * PARAMETERS:
 *      ipimap  - pointer to inode map inode for the aggregate or fileset.
 *
 * RETURN VALUES:
 *      0       - success
 *      -ENOMEM  - insufficient free virtual memory.
 *      -EIO  	- i/o error.
 */
int diMount(struct inode *ipimap)
{
	struct inomap *imap;
	struct metapage *mp;
	int index;
	struct dinomap_disk *dinom_le;

	/*
	 * allocate/initialize the in-memory inode map control structure
	 */
	/* allocate the in-memory inode map control structure. */
	imap = (struct inomap *) kmalloc(sizeof(struct inomap), GFP_KERNEL);
	if (imap == NULL) {
		jfs_err("diMount: kmalloc returned NULL!");
		return -ENOMEM;
	}

	/* read the on-disk inode map control structure. */

	mp = read_metapage(ipimap,
			   IMAPBLKNO << JFS_SBI(ipimap->i_sb)->l2nbperpage,
			   PSIZE, 0);
	if (mp == NULL) {
		kfree(imap);
		return -EIO;
	}

	/* copy the on-disk version to the in-memory version. */
	dinom_le = (struct dinomap_disk *) mp->data;
	imap->im_freeiag = le32_to_cpu(dinom_le->in_freeiag);
	imap->im_nextiag = le32_to_cpu(dinom_le->in_nextiag);
	atomic_set(&imap->im_numinos, le32_to_cpu(dinom_le->in_numinos));
	atomic_set(&imap->im_numfree, le32_to_cpu(dinom_le->in_numfree));
	imap->im_nbperiext = le32_to_cpu(dinom_le->in_nbperiext);
	imap->im_l2nbperiext = le32_to_cpu(dinom_le->in_l2nbperiext);
	for (index = 0; index < MAXAG; index++) {
		imap->im_agctl[index].inofree =
		    le32_to_cpu(dinom_le->in_agctl[index].inofree);
		imap->im_agctl[index].extfree =
		    le32_to_cpu(dinom_le->in_agctl[index].extfree);
		imap->im_agctl[index].numinos =
		    le32_to_cpu(dinom_le->in_agctl[index].numinos);
		imap->im_agctl[index].numfree =
		    le32_to_cpu(dinom_le->in_agctl[index].numfree);
	}

	/* release the buffer. */
	release_metapage(mp);

	/*
	 * allocate/initialize inode allocation map locks
	 */
	/* allocate and init iag free list lock */
	IAGFREE_LOCK_INIT(imap);

	/* allocate and init ag list locks */
	for (index = 0; index < MAXAG; index++) {
		AG_LOCK_INIT(imap, index);
	}

	/* bind the inode map inode and inode map control structure
	 * to each other.
	 */
	imap->im_ipimap = ipimap;
	JFS_IP(ipimap)->i_imap = imap;

	return (0);
}


/*
 * NAME:        diUnmount()
 *
 * FUNCTION:    write to disk the incore inode map control structures for
 *		a fileset or aggregate at unmount time.
 *
 * PARAMETERS:
 *      ipimap  - pointer to inode map inode for the aggregate or fileset.
 *
 * RETURN VALUES:
 *      0       - success
 *      -ENOMEM  - insufficient free virtual memory.
 *      -EIO  	- i/o error.
 */
int diUnmount(struct inode *ipimap, int mounterror)
{
	struct inomap *imap = JFS_IP(ipimap)->i_imap;

	/*
	 * update the on-disk inode map control structure
	 */

	if (!(mounterror || isReadOnly(ipimap)))
		diSync(ipimap);

	/*
	 * Invalidate the page cache buffers
	 */
	truncate_inode_pages(ipimap->i_mapping, 0);

	/*
	 * free in-memory control structure
	 */
	kfree(imap);

	return (0);
}


/*
 *	diSync()
 */
int diSync(struct inode *ipimap)
{
	struct dinomap_disk *dinom_le;
	struct inomap *imp = JFS_IP(ipimap)->i_imap;
	struct metapage *mp;
	int index;

	/*
	 * write imap global conrol page
	 */
	/* read the on-disk inode map control structure */
	mp = get_metapage(ipimap,
			  IMAPBLKNO << JFS_SBI(ipimap->i_sb)->l2nbperpage,
			  PSIZE, 0);
	if (mp == NULL) {
		jfs_err("diSync: get_metapage failed!");
		return -EIO;
	}

	/* copy the in-memory version to the on-disk version */
	dinom_le = (struct dinomap_disk *) mp->data;
	dinom_le->in_freeiag = cpu_to_le32(imp->im_freeiag);
	dinom_le->in_nextiag = cpu_to_le32(imp->im_nextiag);
	dinom_le->in_numinos = cpu_to_le32(atomic_read(&imp->im_numinos));
	dinom_le->in_numfree = cpu_to_le32(atomic_read(&imp->im_numfree));
	dinom_le->in_nbperiext = cpu_to_le32(imp->im_nbperiext);
	dinom_le->in_l2nbperiext = cpu_to_le32(imp->im_l2nbperiext);
	for (index = 0; index < MAXAG; index++) {
		dinom_le->in_agctl[index].inofree =
		    cpu_to_le32(imp->im_agctl[index].inofree);
		dinom_le->in_agctl[index].extfree =
		    cpu_to_le32(imp->im_agctl[index].extfree);
		dinom_le->in_agctl[index].numinos =
		    cpu_to_le32(imp->im_agctl[index].numinos);
		dinom_le->in_agctl[index].numfree =
		    cpu_to_le32(imp->im_agctl[index].numfree);
	}

	/* write out the control structure */
	write_metapage(mp);

	/*
	 * write out dirty pages of imap
	 */
268
	filemap_write_and_wait(ipimap->i_mapping);
Linus Torvalds's avatar
Linus Torvalds committed
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
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

	diWriteSpecial(ipimap, 0);

	return (0);
}


/*
 * NAME:        diRead()
 *
 * FUNCTION:    initialize an incore inode from disk.
 *
 *		on entry, the specifed incore inode should itself
 *		specify the disk inode number corresponding to the
 *		incore inode (i.e. i_number should be initialized).
 *		
 *		this routine handles incore inode initialization for
 *		both "special" and "regular" inodes.  special inodes
 *		are those required early in the mount process and
 *	        require special handling since much of the file system
 *		is not yet initialized.  these "special" inodes are
 *		identified by a NULL inode map inode pointer and are
 *		actually initialized by a call to diReadSpecial().
 *		
 *		for regular inodes, the iag describing the disk inode
 *		is read from disk to determine the inode extent address
 *		for the disk inode.  with the inode extent address in
 *		hand, the page of the extent that contains the disk
 *		inode is read and the disk inode is copied to the
 *		incore inode.
 *
 * PARAMETERS:
 *      ip  -  pointer to incore inode to be initialized from disk.
 *
 * RETURN VALUES:
 *      0       - success
 *      -EIO  	- i/o error.
 *      -ENOMEM	- insufficient memory
 *      
 */
int diRead(struct inode *ip)
{
	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
	int iagno, ino, extno, rc;
	struct inode *ipimap;
	struct dinode *dp;
	struct iag *iagp;
	struct metapage *mp;
	s64 blkno, agstart;
	struct inomap *imap;
	int block_offset;
	int inodes_left;
	uint pageno;
	int rel_inode;

	jfs_info("diRead: ino = %ld", ip->i_ino);

	ipimap = sbi->ipimap;
	JFS_IP(ip)->ipimap = ipimap;

	/* determine the iag number for this inode (number) */
	iagno = INOTOIAG(ip->i_ino);

	/* read the iag */
	imap = JFS_IP(ipimap)->i_imap;
	IREAD_LOCK(ipimap);
	rc = diIAGRead(imap, iagno, &mp);
	IREAD_UNLOCK(ipimap);
	if (rc) {
		jfs_err("diRead: diIAGRead returned %d", rc);
		return (rc);
	}

	iagp = (struct iag *) mp->data;

	/* determine inode extent that holds the disk inode */
	ino = ip->i_ino & (INOSPERIAG - 1);
	extno = ino >> L2INOSPEREXT;

	if ((lengthPXD(&iagp->inoext[extno]) != imap->im_nbperiext) ||
	    (addressPXD(&iagp->inoext[extno]) == 0)) {
		release_metapage(mp);
		return -ESTALE;
	}

	/* get disk block number of the page within the inode extent
	 * that holds the disk inode.
	 */
	blkno = INOPBLK(&iagp->inoext[extno], ino, sbi->l2nbperpage);

	/* get the ag for the iag */
	agstart = le64_to_cpu(iagp->agstart);

	release_metapage(mp);

	rel_inode = (ino & (INOSPERPAGE - 1));
	pageno = blkno >> sbi->l2nbperpage;

	if ((block_offset = ((u32) blkno & (sbi->nbperpage - 1)))) {
		/*
		 * OS/2 didn't always align inode extents on page boundaries
		 */
		inodes_left =
		     (sbi->nbperpage - block_offset) << sbi->l2niperblk;

		if (rel_inode < inodes_left)
			rel_inode += block_offset << sbi->l2niperblk;
		else {
			pageno += 1;
			rel_inode -= inodes_left;
		}
	}

	/* read the page of disk inode */
	mp = read_metapage(ipimap, pageno << sbi->l2nbperpage, PSIZE, 1);
	if (mp == 0) {
		jfs_err("diRead: read_metapage failed");
		return -EIO;
	}

	/* locate the the disk inode requested */
	dp = (struct dinode *) mp->data;
	dp += rel_inode;

	if (ip->i_ino != le32_to_cpu(dp->di_number)) {
		jfs_error(ip->i_sb, "diRead: i_ino != di_number");
		rc = -EIO;
	} else if (le32_to_cpu(dp->di_nlink) == 0)
		rc = -ESTALE;
	else
		/* copy the disk inode to the in-memory inode */
		rc = copy_from_dinode(dp, ip);

	release_metapage(mp);

	/* set the ag for the inode */
	JFS_IP(ip)->agno = BLKTOAG(agstart, sbi);
	JFS_IP(ip)->active_ag = -1;

	return (rc);
}


/*
 * NAME:        diReadSpecial()
 *
 * FUNCTION:    initialize a 'special' inode from disk.
 *
 *		this routines handles aggregate level inodes.  The
 *		inode cache cannot differentiate between the
 *		aggregate inodes and the filesystem inodes, so we
 *		handle these here.  We don't actually use the aggregate
 *	        inode map, since these inodes are at a fixed location
 *		and in some cases the aggregate inode map isn't initialized
 *		yet.
 *
 * PARAMETERS:
 *      sb - filesystem superblock
 *	inum - aggregate inode number
 *	secondary - 1 if secondary aggregate inode table
 *
 * RETURN VALUES:
 *      new inode	- success
 *      NULL		- i/o error.
 */
struct inode *diReadSpecial(struct super_block *sb, ino_t inum, int secondary)
{
	struct jfs_sb_info *sbi = JFS_SBI(sb);
	uint address;
	struct dinode *dp;
	struct inode *ip;
	struct metapage *mp;

	ip = new_inode(sb);
	if (ip == NULL) {
		jfs_err("diReadSpecial: new_inode returned NULL!");
		return ip;
	}

	if (secondary) {
		address = addressPXD(&sbi->ait2) >> sbi->l2nbperpage;
		JFS_IP(ip)->ipimap = sbi->ipaimap2;
	} else {
		address = AITBL_OFF >> L2PSIZE;
		JFS_IP(ip)->ipimap = sbi->ipaimap;
	}

	ASSERT(inum < INOSPEREXT);

	ip->i_ino = inum;

	address += inum >> 3;	/* 8 inodes per 4K page */

	/* read the page of fixed disk inode (AIT) in raw mode */
	mp = read_metapage(ip, address << sbi->l2nbperpage, PSIZE, 1);
	if (mp == NULL) {
		ip->i_nlink = 1;	/* Don't want iput() deleting it */
		iput(ip);
		return (NULL);
	}

	/* get the pointer to the disk inode of interest */
	dp = (struct dinode *) (mp->data);
	dp += inum % 8;		/* 8 inodes per 4K page */

	/* copy on-disk inode to in-memory inode */
	if ((copy_from_dinode(dp, ip)) != 0) {
		/* handle bad return by returning NULL for ip */
		ip->i_nlink = 1;	/* Don't want iput() deleting it */
		iput(ip);
		/* release the page */
		release_metapage(mp);
		return (NULL);

	}

485
	ip->i_mapping->a_ops = &jfs_metapage_aops;
Linus Torvalds's avatar
Linus Torvalds committed
486
487
488
489
490
491
492
493
494
495
496
497
498
	mapping_set_gfp_mask(ip->i_mapping, GFP_NOFS);

	/* Allocations to metadata inodes should not affect quotas */
	ip->i_flags |= S_NOQUOTA;

	if ((inum == FILESYSTEM_I) && (JFS_IP(ip)->ipimap == sbi->ipaimap)) {
		sbi->gengen = le32_to_cpu(dp->di_gengen);
		sbi->inostamp = le32_to_cpu(dp->di_inostamp);
	}

	/* release the page */
	release_metapage(mp);

499
500
	hlist_add_head(&ip->i_hash, &aggregate_hash);

Linus Torvalds's avatar
Linus Torvalds committed
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
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
	return (ip);
}

/*
 * NAME:        diWriteSpecial()
 *
 * FUNCTION:    Write the special inode to disk
 *
 * PARAMETERS:
 *      ip - special inode
 *	secondary - 1 if secondary aggregate inode table
 *
 * RETURN VALUES: none
 */

void diWriteSpecial(struct inode *ip, int secondary)
{
	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
	uint address;
	struct dinode *dp;
	ino_t inum = ip->i_ino;
	struct metapage *mp;

	if (secondary)
		address = addressPXD(&sbi->ait2) >> sbi->l2nbperpage;
	else
		address = AITBL_OFF >> L2PSIZE;

	ASSERT(inum < INOSPEREXT);

	address += inum >> 3;	/* 8 inodes per 4K page */

	/* read the page of fixed disk inode (AIT) in raw mode */
	mp = read_metapage(ip, address << sbi->l2nbperpage, PSIZE, 1);
	if (mp == NULL) {
		jfs_err("diWriteSpecial: failed to read aggregate inode "
			"extent!");
		return;
	}

	/* get the pointer to the disk inode of interest */
	dp = (struct dinode *) (mp->data);
	dp += inum % 8;		/* 8 inodes per 4K page */

	/* copy on-disk inode to in-memory inode */
	copy_to_dinode(dp, ip);
	memcpy(&dp->di_xtroot, &JFS_IP(ip)->i_xtroot, 288);

	if (inum == FILESYSTEM_I)
		dp->di_gengen = cpu_to_le32(sbi->gengen);

	/* write the page */
	write_metapage(mp);
}

/*
 * NAME:        diFreeSpecial()
 *
 * FUNCTION:    Free allocated space for special inode
 */
void diFreeSpecial(struct inode *ip)
{
	if (ip == NULL) {
		jfs_err("diFreeSpecial called with NULL ip!");
		return;
	}
567
	filemap_write_and_wait(ip->i_mapping);
Linus Torvalds's avatar
Linus Torvalds committed
568
569
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
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
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
	truncate_inode_pages(ip->i_mapping, 0);
	iput(ip);
}



/*
 * NAME:        diWrite()
 *
 * FUNCTION:    write the on-disk inode portion of the in-memory inode
 *		to its corresponding on-disk inode.
 *
 *		on entry, the specifed incore inode should itself
 *		specify the disk inode number corresponding to the
 *		incore inode (i.e. i_number should be initialized).
 *
 *		the inode contains the inode extent address for the disk
 *		inode.  with the inode extent address in hand, the
 *		page of the extent that contains the disk inode is
 *		read and the disk inode portion of the incore inode
 *		is copied to the disk inode.
 *		
 * PARAMETERS:
 *	tid -  transacation id
 *      ip  -  pointer to incore inode to be written to the inode extent.
 *
 * RETURN VALUES:
 *      0       - success
 *      -EIO  	- i/o error.
 */
int diWrite(tid_t tid, struct inode *ip)
{
	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
	struct jfs_inode_info *jfs_ip = JFS_IP(ip);
	int rc = 0;
	s32 ino;
	struct dinode *dp;
	s64 blkno;
	int block_offset;
	int inodes_left;
	struct metapage *mp;
	uint pageno;
	int rel_inode;
	int dioffset;
	struct inode *ipimap;
	uint type;
	lid_t lid;
	struct tlock *ditlck, *tlck;
	struct linelock *dilinelock, *ilinelock;
	struct lv *lv;
	int n;

	ipimap = jfs_ip->ipimap;

	ino = ip->i_ino & (INOSPERIAG - 1);

	if (!addressPXD(&(jfs_ip->ixpxd)) ||
	    (lengthPXD(&(jfs_ip->ixpxd)) !=
	     JFS_IP(ipimap)->i_imap->im_nbperiext)) {
		jfs_error(ip->i_sb, "diWrite: ixpxd invalid");
		return -EIO;
	}

	/*
	 * read the page of disk inode containing the specified inode:
	 */
	/* compute the block address of the page */
	blkno = INOPBLK(&(jfs_ip->ixpxd), ino, sbi->l2nbperpage);

	rel_inode = (ino & (INOSPERPAGE - 1));
	pageno = blkno >> sbi->l2nbperpage;

	if ((block_offset = ((u32) blkno & (sbi->nbperpage - 1)))) {
		/*
		 * OS/2 didn't always align inode extents on page boundaries
		 */
		inodes_left =
		    (sbi->nbperpage - block_offset) << sbi->l2niperblk;

		if (rel_inode < inodes_left)
			rel_inode += block_offset << sbi->l2niperblk;
		else {
			pageno += 1;
			rel_inode -= inodes_left;
		}
	}
	/* read the page of disk inode */
      retry:
	mp = read_metapage(ipimap, pageno << sbi->l2nbperpage, PSIZE, 1);
	if (mp == 0)
		return -EIO;

	/* get the pointer to the disk inode */
	dp = (struct dinode *) mp->data;
	dp += rel_inode;

	dioffset = (ino & (INOSPERPAGE - 1)) << L2DISIZE;

	/*
	 * acquire transaction lock on the on-disk inode;
	 * N.B. tlock is acquired on ipimap not ip;
	 */
	if ((ditlck =
	     txLock(tid, ipimap, mp, tlckINODE | tlckENTRY)) == NULL)
		goto retry;
	dilinelock = (struct linelock *) & ditlck->lock;

	/*
	 * copy btree root from in-memory inode to on-disk inode
	 *
	 * (tlock is taken from inline B+-tree root in in-memory
	 * inode when the B+-tree root is updated, which is pointed 
	 * by jfs_ip->blid as well as being on tx tlock list)
	 *
	 * further processing of btree root is based on the copy 
	 * in in-memory inode, where txLog() will log from, and, 
	 * for xtree root, txUpdateMap() will update map and reset
	 * XAD_NEW bit;
	 */

	if (S_ISDIR(ip->i_mode) && (lid = jfs_ip->xtlid)) {
		/*
		 * This is the special xtree inside the directory for storing
		 * the directory table
		 */
		xtpage_t *p, *xp;
		xad_t *xad;

		jfs_ip->xtlid = 0;
		tlck = lid_to_tlock(lid);
		assert(tlck->type & tlckXTREE);
		tlck->type |= tlckBTROOT;
		tlck->mp = mp;
		ilinelock = (struct linelock *) & tlck->lock;

		/*
		 * copy xtree root from inode to dinode:
		 */
		p = &jfs_ip->i_xtroot;
		xp = (xtpage_t *) &dp->di_dirtable;
		lv = ilinelock->lv;
		for (n = 0; n < ilinelock->index; n++, lv++) {
			memcpy(&xp->xad[lv->offset], &p->xad[lv->offset],
			       lv->length << L2XTSLOTSIZE);
		}

		/* reset on-disk (metadata page) xtree XAD_NEW bit */
		xad = &xp->xad[XTENTRYSTART];
		for (n = XTENTRYSTART;
		     n < le16_to_cpu(xp->header.nextindex); n++, xad++)
			if (xad->flag & (XAD_NEW | XAD_EXTENDED))
				xad->flag &= ~(XAD_NEW | XAD_EXTENDED);
	}

	if ((lid = jfs_ip->blid) == 0)
		goto inlineData;
	jfs_ip->blid = 0;

	tlck = lid_to_tlock(lid);
	type = tlck->type;
	tlck->type |= tlckBTROOT;
	tlck->mp = mp;
	ilinelock = (struct linelock *) & tlck->lock;

	/*
	 *      regular file: 16 byte (XAD slot) granularity
	 */
	if (type & tlckXTREE) {
		xtpage_t *p, *xp;
		xad_t *xad;

		/*
		 * copy xtree root from inode to dinode:
		 */
		p = &jfs_ip->i_xtroot;
		xp = &dp->di_xtroot;
		lv = ilinelock->lv;
		for (n = 0; n < ilinelock->index; n++, lv++) {
			memcpy(&xp->xad[lv->offset], &p->xad[lv->offset],
			       lv->length << L2XTSLOTSIZE);
		}

		/* reset on-disk (metadata page) xtree XAD_NEW bit */
		xad = &xp->xad[XTENTRYSTART];
		for (n = XTENTRYSTART;
		     n < le16_to_cpu(xp->header.nextindex); n++, xad++)
			if (xad->flag & (XAD_NEW | XAD_EXTENDED))
				xad->flag &= ~(XAD_NEW | XAD_EXTENDED);
	}
	/*
	 *      directory: 32 byte (directory entry slot) granularity
	 */
	else if (type & tlckDTREE) {
		dtpage_t *p, *xp;

		/*
		 * copy dtree root from inode to dinode:
		 */
		p = (dtpage_t *) &jfs_ip->i_dtroot;
		xp = (dtpage_t *) & dp->di_dtroot;
		lv = ilinelock->lv;
		for (n = 0; n < ilinelock->index; n++, lv++) {
			memcpy(&xp->slot[lv->offset], &p->slot[lv->offset],
			       lv->length << L2DTSLOTSIZE);
		}
	} else {
		jfs_err("diWrite: UFO tlock");
	}

      inlineData:
	/*
	 * copy inline symlink from in-memory inode to on-disk inode
	 */
	if (S_ISLNK(ip->i_mode) && ip->i_size < IDATASIZE) {
		lv = & dilinelock->lv[dilinelock->index];
		lv->offset = (dioffset + 2 * 128) >> L2INODESLOTSIZE;
		lv->length = 2;
		memcpy(&dp->di_fastsymlink, jfs_ip->i_inline, IDATASIZE);
		dilinelock->index++;
	}
	/*
	 * copy inline data from in-memory inode to on-disk inode:
	 * 128 byte slot granularity
	 */
	if (test_cflag(COMMIT_Inlineea, ip)) {
		lv = & dilinelock->lv[dilinelock->index];
		lv->offset = (dioffset + 3 * 128) >> L2INODESLOTSIZE;
		lv->length = 1;
		memcpy(&dp->di_inlineea, jfs_ip->i_inline_ea, INODESLOTSIZE);
		dilinelock->index++;

		clear_cflag(COMMIT_Inlineea, ip);
	}

	/*
	 *      lock/copy inode base: 128 byte slot granularity
	 */
// baseDinode:
	lv = & dilinelock->lv[dilinelock->index];
	lv->offset = dioffset >> L2INODESLOTSIZE;
	copy_to_dinode(dp, ip);
	if (test_and_clear_cflag(COMMIT_Dirtable, ip)) {
		lv->length = 2;
		memcpy(&dp->di_dirtable, &jfs_ip->i_dirtable, 96);
	} else
		lv->length = 1;
	dilinelock->index++;

#ifdef _JFS_FASTDASD
	/*
	 * We aren't logging changes to the DASD used in directory inodes,
	 * but we need to write them to disk.  If we don't unmount cleanly,
	 * mount will recalculate the DASD used.
	 */
	if (S_ISDIR(ip->i_mode)
	    && (ip->i_ipmnt->i_mntflag & JFS_DASD_ENABLED))
		memcpy(&dp->di_DASD, &ip->i_DASD, sizeof(struct dasd));
#endif				/*  _JFS_FASTDASD */

	/* release the buffer holding the updated on-disk inode. 
	 * the buffer will be later written by commit processing.
	 */
	write_metapage(mp);

	return (rc);
}


/*
 * NAME:        diFree(ip)
 *
 * FUNCTION:    free a specified inode from the inode working map
 *		for a fileset or aggregate.
 *
 *		if the inode to be freed represents the first (only)
 *		free inode within the iag, the iag will be placed on
 *		the ag free inode list.
 *	
 *		freeing the inode will cause the inode extent to be
 *		freed if the inode is the only allocated inode within
 *		the extent.  in this case all the disk resource backing
 *		up the inode extent will be freed. in addition, the iag
 *		will be placed on the ag extent free list if the extent
 *		is the first free extent in the iag.  if freeing the
 *		extent also means that no free inodes will exist for
 *		the iag, the iag will also be removed from the ag free
 *		inode list.
 *
 *		the iag describing the inode will be freed if the extent
 *		is to be freed and it is the only backed extent within
 *		the iag.  in this case, the iag will be removed from the
 *		ag free extent list and ag free inode list and placed on
 *		the inode map's free iag list.
 *
 *		a careful update approach is used to provide consistency
 *		in the face of updates to multiple buffers.  under this
 *		approach, all required buffers are obtained before making
 *		any updates and are held until all updates are complete.
 *
 * PARAMETERS:
 *      ip  	- inode to be freed.
 *
 * RETURN VALUES:
 *      0       - success
 *      -EIO  	- i/o error.
 */
int diFree(struct inode *ip)
{
	int rc;
	ino_t inum = ip->i_ino;
	struct iag *iagp, *aiagp, *biagp, *ciagp, *diagp;
	struct metapage *mp, *amp, *bmp, *cmp, *dmp;
	int iagno, ino, extno, bitno, sword, agno;
	int back, fwd;
	u32 bitmap, mask;
	struct inode *ipimap = JFS_SBI(ip->i_sb)->ipimap;
	struct inomap *imap = JFS_IP(ipimap)->i_imap;
	pxd_t freepxd;
	tid_t tid;
	struct inode *iplist[3];
	struct tlock *tlck;
	struct pxd_lock *pxdlock;

	/*
	 * This is just to suppress compiler warnings.  The same logic that
	 * references these variables is used to initialize them.
	 */
	aiagp = biagp = ciagp = diagp = NULL;

	/* get the iag number containing the inode.
	 */
	iagno = INOTOIAG(inum);

	/* make sure that the iag is contained within 
	 * the map.
	 */
	if (iagno >= imap->im_nextiag) {
		dump_mem("imap", imap, 32);
		jfs_error(ip->i_sb,
			  "diFree: inum = %d, iagno = %d, nextiag = %d",
			  (uint) inum, iagno, imap->im_nextiag);
		return -EIO;
	}

	/* get the allocation group for this ino.
	 */
	agno = JFS_IP(ip)->agno;

	/* Lock the AG specific inode map information
	 */
	AG_LOCK(imap, agno);

	/* Obtain read lock in imap inode.  Don't release it until we have
	 * read all of the IAG's that we are going to.
	 */
	IREAD_LOCK(ipimap);

	/* read the iag.
	 */
	if ((rc = diIAGRead(imap, iagno, &mp))) {
		IREAD_UNLOCK(ipimap);
		AG_UNLOCK(imap, agno);
		return (rc);
	}
	iagp = (struct iag *) mp->data;

	/* get the inode number and extent number of the inode within
	 * the iag and the inode number within the extent.
	 */
	ino = inum & (INOSPERIAG - 1);
	extno = ino >> L2INOSPEREXT;
	bitno = ino & (INOSPEREXT - 1);
	mask = HIGHORDER >> bitno;

	if (!(le32_to_cpu(iagp->wmap[extno]) & mask)) {
		jfs_error(ip->i_sb,
			  "diFree: wmap shows inode already free");
	}

	if (!addressPXD(&iagp->inoext[extno])) {
		release_metapage(mp);
		IREAD_UNLOCK(ipimap);
		AG_UNLOCK(imap, agno);
		jfs_error(ip->i_sb, "diFree: invalid inoext");
		return -EIO;
	}

	/* compute the bitmap for the extent reflecting the freed inode.
	 */
	bitmap = le32_to_cpu(iagp->wmap[extno]) & ~mask;

	if (imap->im_agctl[agno].numfree > imap->im_agctl[agno].numinos) {
		release_metapage(mp);
		IREAD_UNLOCK(ipimap);
		AG_UNLOCK(imap, agno);
		jfs_error(ip->i_sb, "diFree: numfree > numinos");
		return -EIO;
	}
	/*
	 *      inode extent still has some inodes or below low water mark:
	 *      keep the inode extent;
	 */
	if (bitmap ||
	    imap->im_agctl[agno].numfree < 96 ||
	    (imap->im_agctl[agno].numfree < 288 &&
	     (((imap->im_agctl[agno].numfree * 100) /
	       imap->im_agctl[agno].numinos) <= 25))) {
		/* if the iag currently has no free inodes (i.e.,
		 * the inode being freed is the first free inode of iag),
		 * insert the iag at head of the inode free list for the ag.
		 */
		if (iagp->nfreeinos == 0) {
			/* check if there are any iags on the ag inode
			 * free list.  if so, read the first one so that
			 * we can link the current iag onto the list at
			 * the head.
			 */
			if ((fwd = imap->im_agctl[agno].inofree) >= 0) {
				/* read the iag that currently is the head
				 * of the list.
				 */
				if ((rc = diIAGRead(imap, fwd, &amp))) {
					IREAD_UNLOCK(ipimap);
					AG_UNLOCK(imap, agno);
					release_metapage(mp);
					return (rc);
				}
				aiagp = (struct iag *) amp->data;

				/* make current head point back to the iag.
				 */
				aiagp->inofreeback = cpu_to_le32(iagno);

				write_metapage(amp);
			}

			/* iag points forward to current head and iag
			 * becomes the new head of the list.
			 */
			iagp->inofreefwd =
			    cpu_to_le32(imap->im_agctl[agno].inofree);
			iagp->inofreeback = cpu_to_le32(-1);
			imap->im_agctl[agno].inofree = iagno;
		}
		IREAD_UNLOCK(ipimap);

		/* update the free inode summary map for the extent if
		 * freeing the inode means the extent will now have free
		 * inodes (i.e., the inode being freed is the first free 
		 * inode of extent),
		 */
		if (iagp->wmap[extno] == cpu_to_le32(ONES)) {
			sword = extno >> L2EXTSPERSUM;
			bitno = extno & (EXTSPERSUM - 1);
			iagp->inosmap[sword] &=
			    cpu_to_le32(~(HIGHORDER >> bitno));
		}

		/* update the bitmap.
		 */
		iagp->wmap[extno] = cpu_to_le32(bitmap);

		/* update the free inode counts at the iag, ag and
		 * map level.
		 */
		iagp->nfreeinos =
		    cpu_to_le32(le32_to_cpu(iagp->nfreeinos) + 1);
		imap->im_agctl[agno].numfree += 1;
		atomic_inc(&imap->im_numfree);

		/* release the AG inode map lock
		 */
		AG_UNLOCK(imap, agno);

		/* write the iag */
		write_metapage(mp);

		return (0);
	}


	/*
	 *      inode extent has become free and above low water mark:
	 *      free the inode extent;
	 */

	/*
	 *      prepare to update iag list(s) (careful update step 1)
	 */
	amp = bmp = cmp = dmp = NULL;
	fwd = back = -1;

	/* check if the iag currently has no free extents.  if so,
	 * it will be placed on the head of the ag extent free list.
	 */
	if (iagp->nfreeexts == 0) {
		/* check if the ag extent free list has any iags.
		 * if so, read the iag at the head of the list now.
		 * this (head) iag will be updated later to reflect
		 * the addition of the current iag at the head of
		 * the list.
		 */
		if ((fwd = imap->im_agctl[agno].extfree) >= 0) {
			if ((rc = diIAGRead(imap, fwd, &amp)))
				goto error_out;
			aiagp = (struct iag *) amp->data;
		}
	} else {
		/* iag has free extents. check if the addition of a free
		 * extent will cause all extents to be free within this
		 * iag.  if so, the iag will be removed from the ag extent
		 * free list and placed on the inode map's free iag list.
		 */
		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG - 1)) {
			/* in preparation for removing the iag from the
			 * ag extent free list, read the iags preceeding
			 * and following the iag on the ag extent free
			 * list.
			 */
			if ((fwd = le32_to_cpu(iagp->extfreefwd)) >= 0) {
				if ((rc = diIAGRead(imap, fwd, &amp)))
					goto error_out;
				aiagp = (struct iag *) amp->data;
			}

			if ((back = le32_to_cpu(iagp->extfreeback)) >= 0) {
				if ((rc = diIAGRead(imap, back, &bmp)))
					goto error_out;
				biagp = (struct iag *) bmp->data;
			}
		}
	}

	/* remove the iag from the ag inode free list if freeing
	 * this extent cause the iag to have no free inodes.
	 */
	if (iagp->nfreeinos == cpu_to_le32(INOSPEREXT - 1)) {
		int inofreeback = le32_to_cpu(iagp->inofreeback);
		int inofreefwd = le32_to_cpu(iagp->inofreefwd);

		/* in preparation for removing the iag from the
		 * ag inode free list, read the iags preceeding
		 * and following the iag on the ag inode free
		 * list.  before reading these iags, we must make
		 * sure that we already don't have them in hand
		 * from up above, since re-reading an iag (buffer)
		 * we are currently holding would cause a deadlock.
		 */
		if (inofreefwd >= 0) {

			if (inofreefwd == fwd)
				ciagp = (struct iag *) amp->data;
			else if (inofreefwd == back)
				ciagp = (struct iag *) bmp->data;
			else {
				if ((rc =
				     diIAGRead(imap, inofreefwd, &cmp)))
					goto error_out;
				ciagp = (struct iag *) cmp->data;
			}
			assert(ciagp != NULL);
		}

		if (inofreeback >= 0) {
			if (inofreeback == fwd)
				diagp = (struct iag *) amp->data;
			else if (inofreeback == back)
				diagp = (struct iag *) bmp->data;
			else {
				if ((rc =
				     diIAGRead(imap, inofreeback, &dmp)))
					goto error_out;
				diagp = (struct iag *) dmp->data;
			}
			assert(diagp != NULL);
		}
	}

	IREAD_UNLOCK(ipimap);

	/*
	 * invalidate any page of the inode extent freed from buffer cache;
	 */
	freepxd = iagp->inoext[extno];
	invalidate_pxd_metapages(ip, freepxd);

	/*
	 *      update iag list(s) (careful update step 2)
	 */
	/* add the iag to the ag extent free list if this is the
	 * first free extent for the iag.
	 */
	if (iagp->nfreeexts == 0) {
		if (fwd >= 0)
			aiagp->extfreeback = cpu_to_le32(iagno);

		iagp->extfreefwd =
		    cpu_to_le32(imap->im_agctl[agno].extfree);
		iagp->extfreeback = cpu_to_le32(-1);
		imap->im_agctl[agno].extfree = iagno;
	} else {
		/* remove the iag from the ag extent list if all extents
		 * are now free and place it on the inode map iag free list.
		 */
		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG - 1)) {
			if (fwd >= 0)
				aiagp->extfreeback = iagp->extfreeback;

			if (back >= 0)
				biagp->extfreefwd = iagp->extfreefwd;
			else
				imap->im_agctl[agno].extfree =
				    le32_to_cpu(iagp->extfreefwd);

			iagp->extfreefwd = iagp->extfreeback = cpu_to_le32(-1);

			IAGFREE_LOCK(imap);
			iagp->iagfree = cpu_to_le32(imap->im_freeiag);
			imap->im_freeiag = iagno;
			IAGFREE_UNLOCK(imap);
		}
	}

	/* remove the iag from the ag inode free list if freeing
	 * this extent causes the iag to have no free inodes.
	 */
	if (iagp->nfreeinos == cpu_to_le32(INOSPEREXT - 1)) {
		if ((int) le32_to_cpu(iagp->inofreefwd) >= 0)
			ciagp->inofreeback = iagp->inofreeback;

		if ((int) le32_to_cpu(iagp->inofreeback) >= 0)
			diagp->inofreefwd = iagp->inofreefwd;
		else
			imap->im_agctl[agno].inofree =
			    le32_to_cpu(iagp->inofreefwd);

		iagp->inofreefwd = iagp->inofreeback = cpu_to_le32(-1);
	}

	/* update the inode extent address and working map 
	 * to reflect the free extent.
	 * the permanent map should have been updated already 
	 * for the inode being freed.
	 */
	if (iagp->pmap[extno] != 0) {
		jfs_error(ip->i_sb, "diFree: the pmap does not show inode free");
	}
	iagp->wmap[extno] = 0;
	PXDlength(&iagp->inoext[extno], 0);
	PXDaddress(&iagp->inoext[extno], 0);

	/* update the free extent and free inode summary maps
	 * to reflect the freed extent.
	 * the inode summary map is marked to indicate no inodes 
	 * available for the freed extent.
	 */
	sword = extno >> L2EXTSPERSUM;
	bitno = extno & (EXTSPERSUM - 1);
	mask = HIGHORDER >> bitno;
	iagp->inosmap[sword] |= cpu_to_le32(mask);
	iagp->extsmap[sword] &= cpu_to_le32(~mask);

	/* update the number of free inodes and number of free extents
	 * for the iag.
	 */
	iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) -
				      (INOSPEREXT - 1));
	iagp->nfreeexts = cpu_to_le32(le32_to_cpu(iagp->nfreeexts) + 1);

	/* update the number of free inodes and backed inodes
	 * at the ag and inode map level.
	 */
	imap->im_agctl[agno].numfree -= (INOSPEREXT - 1);
	imap->im_agctl[agno].numinos -= INOSPEREXT;
	atomic_sub(INOSPEREXT - 1, &imap->im_numfree);
	atomic_sub(INOSPEREXT, &imap->im_numinos);

	if (amp)
		write_metapage(amp);
	if (bmp)
		write_metapage(bmp);
	if (cmp)
		write_metapage(cmp);
	if (dmp)
		write_metapage(dmp);

	/*
	 * start transaction to update block allocation map
	 * for the inode extent freed;
	 *
	 * N.B. AG_LOCK is released and iag will be released below, and 
	 * other thread may allocate inode from/reusing the ixad freed
	 * BUT with new/different backing inode extent from the extent 
	 * to be freed by the transaction;  
	 */
	tid = txBegin(ipimap->i_sb, COMMIT_FORCE);
1264
	mutex_lock(&JFS_IP(ipimap)->commit_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
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

	/* acquire tlock of the iag page of the freed ixad 
	 * to force the page NOHOMEOK (even though no data is
	 * logged from the iag page) until NOREDOPAGE|FREEXTENT log 
	 * for the free of the extent is committed;
	 * write FREEXTENT|NOREDOPAGE log record
	 * N.B. linelock is overlaid as freed extent descriptor;
	 */
	tlck = txLock(tid, ipimap, mp, tlckINODE | tlckFREE);
	pxdlock = (struct pxd_lock *) & tlck->lock;
	pxdlock->flag = mlckFREEPXD;
	pxdlock->pxd = freepxd;
	pxdlock->index = 1;

	write_metapage(mp);

	iplist[0] = ipimap;

	/*
	 * logredo needs the IAG number and IAG extent index in order
	 * to ensure that the IMap is consistent.  The least disruptive
	 * way to pass these values through  to the transaction manager
	 * is in the iplist array.  
	 * 
	 * It's not pretty, but it works.
	 */
	iplist[1] = (struct inode *) (size_t)iagno;
	iplist[2] = (struct inode *) (size_t)extno;

	rc = txCommit(tid, 1, &iplist[0], COMMIT_FORCE);

	txEnd(tid);
1297
	mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
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

	/* unlock the AG inode map information */
	AG_UNLOCK(imap, agno);

	return (0);

      error_out:
	IREAD_UNLOCK(ipimap);

	if (amp)
		release_metapage(amp);
	if (bmp)
		release_metapage(bmp);
	if (cmp)
		release_metapage(cmp);
	if (dmp)
		release_metapage(dmp);

	AG_UNLOCK(imap, agno);

	release_metapage(mp);

	return (rc);
}

/*
 * There are several places in the diAlloc* routines where we initialize
 * the inode.
 */
static inline void
diInitInode(struct inode *ip, int iagno, int ino, int extno, struct iag * iagp)
{
	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
	struct jfs_inode_info *jfs_ip = JFS_IP(ip);

	ip->i_ino = (iagno << L2INOSPERIAG) + ino;
	jfs_ip->ixpxd = iagp->inoext[extno];
	jfs_ip->agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
	jfs_ip->active_ag = -1;
}


/*
 * NAME:        diAlloc(pip,dir,ip)
 *
 * FUNCTION:    allocate a disk inode from the inode working map 
 *		for a fileset or aggregate.
 *
 * PARAMETERS:
 *      pip  	- pointer to incore inode for the parent inode.
 *      dir  	- TRUE if the new disk inode is for a directory.
 *      ip  	- pointer to a new inode
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
int diAlloc(struct inode *pip, boolean_t dir, struct inode *ip)
{
	int rc, ino, iagno, addext, extno, bitno, sword;
	int nwords, rem, i, agno;
	u32 mask, inosmap, extsmap;
	struct inode *ipimap;
	struct metapage *mp;
	ino_t inum;
	struct iag *iagp;
	struct inomap *imap;

	/* get the pointers to the inode map inode and the
	 * corresponding imap control structure.
	 */
	ipimap = JFS_SBI(pip->i_sb)->ipimap;
	imap = JFS_IP(ipimap)->i_imap;
	JFS_IP(ip)->ipimap = ipimap;
	JFS_IP(ip)->fileset = FILESYSTEM_I;

	/* for a directory, the allocation policy is to start 
	 * at the ag level using the preferred ag.
	 */
	if (dir == TRUE) {
		agno = dbNextAG(JFS_SBI(pip->i_sb)->ipbmap);
		AG_LOCK(imap, agno);
		goto tryag;
	}

	/* for files, the policy starts off by trying to allocate from
	 * the same iag containing the parent disk inode:
	 * try to allocate the new disk inode close to the parent disk
	 * inode, using parent disk inode number + 1 as the allocation
	 * hint.  (we use a left-to-right policy to attempt to avoid
	 * moving backward on the disk.)  compute the hint within the
	 * file system and the iag.
	 */

	/* get the ag number of this iag */
	agno = JFS_IP(pip)->agno;

	if (atomic_read(&JFS_SBI(pip->i_sb)->bmap->db_active[agno])) {
		/*
		 * There is an open file actively growing.  We want to
		 * allocate new inodes from a different ag to avoid
		 * fragmentation problems.
		 */
		agno = dbNextAG(JFS_SBI(pip->i_sb)->ipbmap);
		AG_LOCK(imap, agno);
		goto tryag;
	}

	inum = pip->i_ino + 1;
	ino = inum & (INOSPERIAG - 1);

	/* back off the the hint if it is outside of the iag */
	if (ino == 0)
		inum = pip->i_ino;

	/* lock the AG inode map information */
	AG_LOCK(imap, agno);

	/* Get read lock on imap inode */
	IREAD_LOCK(ipimap);

	/* get the iag number and read the iag */
	iagno = INOTOIAG(inum);
	if ((rc = diIAGRead(imap, iagno, &mp))) {
		IREAD_UNLOCK(ipimap);
		AG_UNLOCK(imap, agno);
		return (rc);
	}
	iagp = (struct iag *) mp->data;

	/* determine if new inode extent is allowed to be added to the iag.
	 * new inode extent can be added to the iag if the ag
	 * has less than 32 free disk inodes and the iag has free extents.
	 */
	addext = (imap->im_agctl[agno].numfree < 32 && iagp->nfreeexts);

	/*
	 *      try to allocate from the IAG
	 */
	/* check if the inode may be allocated from the iag 
	 * (i.e. the inode has free inodes or new extent can be added).
	 */
	if (iagp->nfreeinos || addext) {
		/* determine the extent number of the hint.
		 */
		extno = ino >> L2INOSPEREXT;

		/* check if the extent containing the hint has backed
		 * inodes.  if so, try to allocate within this extent.
		 */
		if (addressPXD(&iagp->inoext[extno])) {
			bitno = ino & (INOSPEREXT - 1);
			if ((bitno =
			     diFindFree(le32_to_cpu(iagp->wmap[extno]),
					bitno))
			    < INOSPEREXT) {
				ino = (extno << L2INOSPEREXT) + bitno;

				/* a free inode (bit) was found within this
				 * extent, so allocate it.
				 */
				rc = diAllocBit(imap, iagp, ino);
				IREAD_UNLOCK(ipimap);
				if (rc) {
					assert(rc == -EIO);
				} else {
					/* set the results of the allocation
					 * and write the iag.
					 */
					diInitInode(ip, iagno, ino, extno,
						    iagp);
					mark_metapage_dirty(mp);
				}
				release_metapage(mp);

				/* free the AG lock and return.
				 */
				AG_UNLOCK(imap, agno);
				return (rc);
			}

			if (!addext)
				extno =
				    (extno ==
				     EXTSPERIAG - 1) ? 0 : extno + 1;
		}

		/*
		 * no free inodes within the extent containing the hint.
		 *
		 * try to allocate from the backed extents following
		 * hint or, if appropriate (i.e. addext is true), allocate
		 * an extent of free inodes at or following the extent
		 * containing the hint.
		 * 
		 * the free inode and free extent summary maps are used
		 * here, so determine the starting summary map position
		 * and the number of words we'll have to examine.  again,
		 * the approach is to allocate following the hint, so we
		 * might have to initially ignore prior bits of the summary
		 * map that represent extents prior to the extent containing
		 * the hint and later revisit these bits.
		 */
		bitno = extno & (EXTSPERSUM - 1);
		nwords = (bitno == 0) ? SMAPSZ : SMAPSZ + 1;
		sword = extno >> L2EXTSPERSUM;

		/* mask any prior bits for the starting words of the
		 * summary map.
		 */
		mask = ONES << (EXTSPERSUM - bitno);
		inosmap = le32_to_cpu(iagp->inosmap[sword]) | mask;
		extsmap = le32_to_cpu(iagp->extsmap[sword]) | mask;

		/* scan the free inode and free extent summary maps for
		 * free resources.
		 */
		for (i = 0; i < nwords; i++) {
			/* check if this word of the free inode summary
			 * map describes an extent with free inodes.
			 */
			if (~inosmap) {
				/* an extent with free inodes has been
				 * found. determine the extent number
				 * and the inode number within the extent.
				 */
				rem = diFindFree(inosmap, 0);
				extno = (sword << L2EXTSPERSUM) + rem;
				rem = diFindFree(le32_to_cpu(iagp->wmap[extno]),
						 0);
				if (rem >= INOSPEREXT) {
					IREAD_UNLOCK(ipimap);
					release_metapage(mp);
					AG_UNLOCK(imap, agno);
					jfs_error(ip->i_sb,
						  "diAlloc: can't find free bit "
						  "in wmap");
					return EIO;
				}

				/* determine the inode number within the
				 * iag and allocate the inode from the
				 * map.
				 */
				ino = (extno << L2INOSPEREXT) + rem;
				rc = diAllocBit(imap, iagp, ino);
				IREAD_UNLOCK(ipimap);
				if (rc)
					assert(rc == -EIO);
				else {
					/* set the results of the allocation
					 * and write the iag.
					 */
					diInitInode(ip, iagno, ino, extno,
						    iagp);
					mark_metapage_dirty(mp);
				}
				release_metapage(mp);

				/* free the AG lock and return.
				 */
				AG_UNLOCK(imap, agno);
				return (rc);

			}

			/* check if we may allocate an extent of free
			 * inodes and whether this word of the free
			 * extents summary map describes a free extent.
			 */
			if (addext && ~extsmap) {
				/* a free extent has been found.  determine
				 * the extent number.
				 */
				rem = diFindFree(extsmap, 0);
				extno = (sword << L2EXTSPERSUM) + rem;

				/* allocate an extent of free inodes.
				 */
				if ((rc = diNewExt(imap, iagp, extno))) {
					/* if there is no disk space for a
					 * new extent, try to allocate the
					 * disk inode from somewhere else.
					 */
					if (rc == -ENOSPC)
						break;

					assert(rc == -EIO);
				} else {
					/* set the results of the allocation
					 * and write the iag.
					 */
					diInitInode(ip, iagno,
						    extno << L2INOSPEREXT,
						    extno, iagp);
					mark_metapage_dirty(mp);
				}
				release_metapage(mp);
				/* free the imap inode & the AG lock & return.
				 */
				IREAD_UNLOCK(ipimap);
				AG_UNLOCK(imap, agno);
				return (rc);
			}

			/* move on to the next set of summary map words.
			 */
			sword = (sword == SMAPSZ - 1) ? 0 : sword + 1;
			inosmap = le32_to_cpu(iagp->inosmap[sword]);
			extsmap = le32_to_cpu(iagp->extsmap[sword]);
		}
	}
	/* unlock imap inode */
	IREAD_UNLOCK(ipimap);

	/* nothing doing in this iag, so release it. */
	release_metapage(mp);

      tryag:
	/*
	 * try to allocate anywhere within the same AG as the parent inode.
	 */
	rc = diAllocAG(imap, agno, dir, ip);

	AG_UNLOCK(imap, agno);

	if (rc != -ENOSPC)
		return (rc);

	/*
	 * try to allocate in any AG.
	 */
	return (diAllocAny(imap, agno, dir, ip));
}


/*
 * NAME:        diAllocAG(imap,agno,dir,ip)
 *
 * FUNCTION:    allocate a disk inode from the allocation group.
 *
 *		this routine first determines if a new extent of free
 *		inodes should be added for the allocation group, with
 *		the current request satisfied from this extent. if this
 *		is the case, an attempt will be made to do just that.  if
 *		this attempt fails or it has been determined that a new 
 *		extent should not be added, an attempt is made to satisfy
 *		the request by allocating an existing (backed) free inode
 *		from the allocation group.
 *
 * PRE CONDITION: Already have the AG lock for this AG.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      agno  	- allocation group to allocate from.
 *      dir  	- TRUE if the new disk inode is for a directory.
 *      ip  	- pointer to the new inode to be filled in on successful return
 *		  with the disk inode number allocated, its extent address
 *		  and the start of the ag.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int
diAllocAG(struct inomap * imap, int agno, boolean_t dir, struct inode *ip)
{
	int rc, addext, numfree, numinos;

	/* get the number of free and the number of backed disk 
	 * inodes currently within the ag.
	 */
	numfree = imap->im_agctl[agno].numfree;
	numinos = imap->im_agctl[agno].numinos;

	if (numfree > numinos) {
		jfs_error(ip->i_sb, "diAllocAG: numfree > numinos");
		return -EIO;
	}

	/* determine if we should allocate a new extent of free inodes
	 * within the ag: for directory inodes, add a new extent
	 * if there are a small number of free inodes or number of free
	 * inodes is a small percentage of the number of backed inodes.
	 */
	if (dir == TRUE)
		addext = (numfree < 64 ||
			  (numfree < 256
			   && ((numfree * 100) / numinos) <= 20));
	else
		addext = (numfree == 0);

	/*
	 * try to allocate a new extent of free inodes.
	 */
	if (addext) {
		/* if free space is not avaliable for this new extent, try
		 * below to allocate a free and existing (already backed)
		 * inode from the ag.
		 */
		if ((rc = diAllocExt(imap, agno, ip)) != -ENOSPC)
			return (rc);
	}

	/*
	 * try to allocate an existing free inode from the ag.
	 */
	return (diAllocIno(imap, agno, ip));
}


/*
 * NAME:        diAllocAny(imap,agno,dir,iap)
 *
 * FUNCTION:    allocate a disk inode from any other allocation group.
 *
 *		this routine is called when an allocation attempt within
 *		the primary allocation group has failed. if attempts to
 *		allocate an inode from any allocation group other than the
 *		specified primary group.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      agno  	- primary allocation group (to avoid).
 *      dir  	- TRUE if the new disk inode is for a directory.
 *      ip  	- pointer to a new inode to be filled in on successful return
 *		  with the disk inode number allocated, its extent address
 *		  and the start of the ag.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int
diAllocAny(struct inomap * imap, int agno, boolean_t dir, struct inode *ip)
{
	int ag, rc;
	int maxag = JFS_SBI(imap->im_ipimap->i_sb)->bmap->db_maxag;


	/* try to allocate from the ags following agno up to 
	 * the maximum ag number.
	 */
	for (ag = agno + 1; ag <= maxag; ag++) {
		AG_LOCK(imap, ag);

		rc = diAllocAG(imap, ag, dir, ip);

		AG_UNLOCK(imap, ag);

		if (rc != -ENOSPC)
			return (rc);
	}

	/* try to allocate from the ags in front of agno.
	 */
	for (ag = 0; ag < agno; ag++) {
		AG_LOCK(imap, ag);

		rc = diAllocAG(imap, ag, dir, ip);

		AG_UNLOCK(imap, ag);

		if (rc != -ENOSPC)
			return (rc);
	}

	/* no free disk inodes.
	 */
	return -ENOSPC;
}


/*
 * NAME:        diAllocIno(imap,agno,ip)
 *
 * FUNCTION:    allocate a disk inode from the allocation group's free
 *		inode list, returning an error if this free list is
 *		empty (i.e. no iags on the list).
 *
 *		allocation occurs from the first iag on the list using
 *		the iag's free inode summary map to find the leftmost
 *		free inode in the iag. 
 *		
 * PRE CONDITION: Already have AG lock for this AG.
 *		
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      agno  	- allocation group.
 *      ip  	- pointer to new inode to be filled in on successful return
 *		  with the disk inode number allocated, its extent address
 *		  and the start of the ag.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int diAllocIno(struct inomap * imap, int agno, struct inode *ip)
{
	int iagno, ino, rc, rem, extno, sword;
	struct metapage *mp;
	struct iag *iagp;

	/* check if there are iags on the ag's free inode list.
	 */
	if ((iagno = imap->im_agctl[agno].inofree) < 0)
		return -ENOSPC;

	/* obtain read lock on imap inode */
	IREAD_LOCK(imap->im_ipimap);

	/* read the iag at the head of the list.
	 */
	if ((rc = diIAGRead(imap, iagno, &mp))) {
		IREAD_UNLOCK(imap->im_ipimap);
		return (rc);
	}
	iagp = (struct iag *) mp->data;

	/* better be free inodes in this iag if it is on the
	 * list.
	 */
	if (!iagp->nfreeinos) {
		IREAD_UNLOCK(imap->im_ipimap);
		release_metapage(mp);
		jfs_error(ip->i_sb,
			  "diAllocIno: nfreeinos = 0, but iag on freelist");
		return -EIO;
	}

	/* scan the free inode summary map to find an extent
	 * with free inodes.
	 */
	for (sword = 0;; sword++) {
		if (sword >= SMAPSZ) {
			IREAD_UNLOCK(imap->im_ipimap);
			release_metapage(mp);
			jfs_error(ip->i_sb,
				  "diAllocIno: free inode not found in summary map");
			return -EIO;
		}

		if (~iagp->inosmap[sword])
			break;
	}

	/* found a extent with free inodes. determine
	 * the extent number.
	 */
	rem = diFindFree(le32_to_cpu(iagp->inosmap[sword]), 0);
	if (rem >= EXTSPERSUM) {
		IREAD_UNLOCK(imap->im_ipimap);
		release_metapage(mp);
		jfs_error(ip->i_sb, "diAllocIno: no free extent found");
		return -EIO;
	}
	extno = (sword << L2EXTSPERSUM) + rem;

	/* find the first free inode in the extent.
	 */
	rem = diFindFree(le32_to_cpu(iagp->wmap[extno]), 0);
	if (rem >= INOSPEREXT) {
		IREAD_UNLOCK(imap->im_ipimap);
		release_metapage(mp);
		jfs_error(ip->i_sb, "diAllocIno: free inode not found");
		return -EIO;
	}

	/* compute the inode number within the iag. 
	 */
	ino = (extno << L2INOSPEREXT) + rem;

	/* allocate the inode.
	 */
	rc = diAllocBit(imap, iagp, ino);
	IREAD_UNLOCK(imap->im_ipimap);
	if (rc) {
		release_metapage(mp);
		return (rc);
	}

	/* set the results of the allocation and write the iag.
	 */
	diInitInode(ip, iagno, ino, extno, iagp);
	write_metapage(mp);

	return (0);
}


/*
 * NAME:        diAllocExt(imap,agno,ip)
 *
 * FUNCTION:   	add a new extent of free inodes to an iag, allocating
 *	       	an inode from this extent to satisfy the current allocation
 *	       	request.
 *		
 *		this routine first tries to find an existing iag with free
 *		extents through the ag free extent list.  if list is not
 *		empty, the head of the list will be selected as the home
 *		of the new extent of free inodes.  otherwise (the list is
 *		empty), a new iag will be allocated for the ag to contain
 *		the extent.
 *		
 *		once an iag has been selected, the free extent summary map
 *		is used to locate a free extent within the iag and diNewExt()
 *		is called to initialize the extent, with initialization
 *		including the allocation of the first inode of the extent
 *		for the purpose of satisfying this request.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      agno  	- allocation group number.
 *      ip  	- pointer to new inode to be filled in on successful return
 *		  with the disk inode number allocated, its extent address
 *		  and the start of the ag.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int diAllocExt(struct inomap * imap, int agno, struct inode *ip)
{
	int rem, iagno, sword, extno, rc;
	struct metapage *mp;
	struct iag *iagp;

	/* check if the ag has any iags with free extents.  if not,
	 * allocate a new iag for the ag.
	 */
	if ((iagno = imap->im_agctl[agno].extfree) < 0) {
		/* If successful, diNewIAG will obtain the read lock on the
		 * imap inode.
		 */
		if ((rc = diNewIAG(imap, &iagno, agno, &mp))) {
			return (rc);
		}
		iagp = (struct iag *) mp->data;

		/* set the ag number if this a brand new iag
		 */
		iagp->agstart =
		    cpu_to_le64(AGTOBLK(agno, imap->im_ipimap));
	} else {
		/* read the iag.
		 */
		IREAD_LOCK(imap->im_ipimap);
		if ((rc = diIAGRead(imap, iagno, &mp))) {
			IREAD_UNLOCK(imap->im_ipimap);
			jfs_error(ip->i_sb, "diAllocExt: error reading iag");
			return rc;
		}
		iagp = (struct iag *) mp->data;
	}

	/* using the free extent summary map, find a free extent.
	 */
	for (sword = 0;; sword++) {
		if (sword >= SMAPSZ) {
			release_metapage(mp);
			IREAD_UNLOCK(imap->im_ipimap);
			jfs_error(ip->i_sb,
				  "diAllocExt: free ext summary map not found");
			return -EIO;
		}
		if (~iagp->extsmap[sword])
			break;
	}

	/* determine the extent number of the free extent.
	 */
	rem = diFindFree(le32_to_cpu(iagp->extsmap[sword]), 0);
	if (rem >= EXTSPERSUM) {
		release_metapage(mp);
		IREAD_UNLOCK(imap->im_ipimap);
		jfs_error(ip->i_sb, "diAllocExt: free extent not found");
		return -EIO;
	}
	extno = (sword << L2EXTSPERSUM) + rem;

	/* initialize the new extent.
	 */
	rc = diNewExt(imap, iagp, extno);
	IREAD_UNLOCK(imap->im_ipimap);
	if (rc) {
		/* something bad happened.  if a new iag was allocated,
		 * place it back on the inode map's iag free list, and
		 * clear the ag number information.
		 */
		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG)) {
			IAGFREE_LOCK(imap);
			iagp->iagfree = cpu_to_le32(imap->im_freeiag);
			imap->im_freeiag = iagno;
			IAGFREE_UNLOCK(imap);
		}
		write_metapage(mp);
		return (rc);
	}

	/* set the results of the allocation and write the iag.
	 */
	diInitInode(ip, iagno, extno << L2INOSPEREXT, extno, iagp);

	write_metapage(mp);

	return (0);
}


/*
 * NAME:        diAllocBit(imap,iagp,ino)
 *
 * FUNCTION:   	allocate a backed inode from an iag.
 *
 *		this routine performs the mechanics of allocating a
 *		specified inode from a backed extent.
 *
 *		if the inode to be allocated represents the last free
 *		inode within the iag, the iag will be removed from the
 *		ag free inode list.
 *
 *		a careful update approach is used to provide consistency
 *		in the face of updates to multiple buffers.  under this
 *		approach, all required buffers are obtained before making
 *		any updates and are held all are updates are complete.
 *		
 * PRE CONDITION: Already have buffer lock on iagp.  Already have AG lock on
 *	this AG.  Must have read lock on imap inode.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      iagp  	- pointer to iag. 
 *      ino   	- inode number to be allocated within the iag.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int diAllocBit(struct inomap * imap, struct iag * iagp, int ino)
{
	int extno, bitno, agno, sword, rc;
	struct metapage *amp = NULL, *bmp = NULL;
	struct iag *aiagp = NULL, *biagp = NULL;
	u32 mask;

	/* check if this is the last free inode within the iag.
	 * if so, it will have to be removed from the ag free
	 * inode list, so get the iags preceeding and following
	 * it on the list.
	 */
	if (iagp->nfreeinos == cpu_to_le32(1)) {
		if ((int) le32_to_cpu(iagp->inofreefwd) >= 0) {
			if ((rc =
			     diIAGRead(imap, le32_to_cpu(iagp->inofreefwd),
				       &amp)))
				return (rc);
			aiagp = (struct iag *) amp->data;
		}

		if ((int) le32_to_cpu(iagp->inofreeback) >= 0) {
			if ((rc =
			     diIAGRead(imap,
				       le32_to_cpu(iagp->inofreeback),
				       &bmp))) {
				if (amp)
					release_metapage(amp);
				return (rc);
			}
			biagp = (struct iag *) bmp->data;
		}
	}

	/* get the ag number, extent number, inode number within
	 * the extent.
	 */
	agno = BLKTOAG(le64_to_cpu(iagp->agstart), JFS_SBI(imap->im_ipimap->i_sb));
	extno = ino >> L2INOSPEREXT;
	bitno = ino & (INOSPEREXT - 1);

	/* compute the mask for setting the map.
	 */
	mask = HIGHORDER >> bitno;

	/* the inode should be free and backed.
	 */
	if (((le32_to_cpu(iagp->pmap[extno]) & mask) != 0) ||
	    ((le32_to_cpu(iagp->wmap[extno]) & mask) != 0) ||
	    (addressPXD(&iagp->inoext[extno]) == 0)) {
		if (amp)
			release_metapage(amp);
		if (bmp)
			release_metapage(bmp);

		jfs_error(imap->im_ipimap->i_sb,
			  "diAllocBit: iag inconsistent");
		return -EIO;
	}

	/* mark the inode as allocated in the working map.
	 */
	iagp->wmap[extno] |= cpu_to_le32(mask);

	/* check if all inodes within the extent are now
	 * allocated.  if so, update the free inode summary
	 * map to reflect this.
	 */
	if (iagp->wmap[extno] == cpu_to_le32(ONES)) {
		sword = extno >> L2EXTSPERSUM;
		bitno = extno & (EXTSPERSUM - 1);
		iagp->inosmap[sword] |= cpu_to_le32(HIGHORDER >> bitno);
	}

	/* if this was the last free inode in the iag, remove the
	 * iag from the ag free inode list.
	 */
	if (iagp->nfreeinos == cpu_to_le32(1)) {
		if (amp) {
			aiagp->inofreeback = iagp->inofreeback;
			write_metapage(amp);
		}

		if (bmp) {
			biagp->inofreefwd = iagp->inofreefwd;
			write_metapage(bmp);
		} else {
			imap->im_agctl[agno].inofree =
			    le32_to_cpu(iagp->inofreefwd);
		}
		iagp->inofreefwd = iagp->inofreeback = cpu_to_le32(-1);
	}

	/* update the free inode count at the iag, ag, inode
	 * map levels.
	 */
	iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) - 1);
	imap->im_agctl[agno].numfree -= 1;
	atomic_dec(&imap->im_numfree);

	return (0);
}


/*
 * NAME:        diNewExt(imap,iagp,extno)
 *
 * FUNCTION:    initialize a new extent of inodes for an iag, allocating
 *	        the first inode of the extent for use for the current
 *	        allocation request.
 *
 *		disk resources are allocated for the new extent of inodes
 *		and the inodes themselves are initialized to reflect their
 *		existence within the extent (i.e. their inode numbers and
 *		inode extent addresses are set) and their initial state
 *		(mode and link count are set to zero).
 *
 *		if the iag is new, it is not yet on an ag extent free list
 *		but will now be placed on this list.
 *
 *		if the allocation of the new extent causes the iag to
 *		have no free extent, the iag will be removed from the
 *		ag extent free list.
 *
 *		if the iag has no free backed inodes, it will be placed
 *		on the ag free inode list, since the addition of the new
 *		extent will now cause it to have free inodes.
 *
 *		a careful update approach is used to provide consistency
 *		(i.e. list consistency) in the face of updates to multiple
 *		buffers.  under this approach, all required buffers are
 *		obtained before making any updates and are held until all
 *		updates are complete.
 *		
 * PRE CONDITION: Already have buffer lock on iagp.  Already have AG lock on
 *	this AG.  Must have read lock on imap inode.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      iagp  	- pointer to iag. 
 *      extno  	- extent number.
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 */
static int diNewExt(struct inomap * imap, struct iag * iagp, int extno)
{
	int agno, iagno, fwd, back, freei = 0, sword, rc;
	struct iag *aiagp = NULL, *biagp = NULL, *ciagp = NULL;
	struct metapage *amp, *bmp, *cmp, *dmp;
	struct inode *ipimap;
	s64 blkno, hint;
	int i, j;
	u32 mask;
	ino_t ino;
	struct dinode *dp;
	struct jfs_sb_info *sbi;

	/* better have free extents.
	 */
	if (!iagp->nfreeexts) {
		jfs_error(imap->im_ipimap->i_sb, "diNewExt: no free extents");
		return -EIO;
	}

	/* get the inode map inode.
	 */
	ipimap = imap->im_ipimap;
	sbi = JFS_SBI(ipimap->i_sb);

	amp = bmp = cmp = NULL;

	/* get the ag and iag numbers for this iag.
	 */
	agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
	iagno = le32_to_cpu(iagp->iagnum);

	/* check if this is the last free extent within the
	 * iag.  if so, the iag must be removed from the ag
	 * free extent list, so get the iags preceeding and
	 * following the iag on this list.
	 */
	if (iagp->nfreeexts == cpu_to_le32(1)) {
		if ((fwd = le32_to_cpu(iagp->extfreefwd)) >= 0) {
			if ((rc = diIAGRead(imap, fwd, &amp)))
				return (rc);
			aiagp = (struct iag *) amp->data;
		}

		if ((back = le32_to_cpu(iagp->extfreeback)) >= 0) {
			if ((rc = diIAGRead(imap, back, &bmp)))
				goto error_out;
			biagp = (struct iag *) bmp->data;
		}
	} else {
		/* the iag has free extents.  if all extents are free
		 * (as is the case for a newly allocated iag), the iag
		 * must be added to the ag free extent list, so get
		 * the iag at the head of the list in preparation for
		 * adding this iag to this list.
		 */
		fwd = back = -1;
		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG)) {
			if ((fwd = imap->im_agctl[agno].extfree) >= 0) {
				if ((rc = diIAGRead(imap, fwd, &amp)))
					goto error_out;
				aiagp = (struct iag *) amp->data;
			}
		}
	}

	/* check if the iag has no free inodes.  if so, the iag
	 * will have to be added to the ag free inode list, so get
	 * the iag at the head of the list in preparation for
	 * adding this iag to this list.  in doing this, we must
	 * check if we already have the iag at the head of
	 * the list in hand.
	 */
	if (iagp->nfreeinos == 0) {
		freei = imap->im_agctl[agno].inofree;

		if (freei >= 0) {
			if (freei == fwd) {
				ciagp = aiagp;
			} else if (freei == back) {
				ciagp = biagp;
			} else {
				if ((rc = diIAGRead(imap, freei, &cmp)))
					goto error_out;
				ciagp = (struct iag *) cmp->data;
			}
			if (ciagp == NULL) {
				jfs_error(imap->im_ipimap->i_sb,
					  "diNewExt: ciagp == NULL");
				rc = -EIO;
				goto error_out;
			}
		}
	}

	/* allocate disk space for the inode extent.
	 */
	if ((extno == 0) || (addressPXD(&iagp->inoext[extno - 1]) == 0))
		hint = ((s64) agno << sbi->bmap->db_agl2size) - 1;
	else
		hint = addressPXD(&iagp->inoext[extno - 1]) +
		    lengthPXD(&iagp->inoext[extno - 1]) - 1;

	if ((rc = dbAlloc(ipimap, hint, (s64) imap->im_nbperiext, &blkno)))
		goto error_out;

	/* compute the inode number of the first inode within the
	 * extent.
	 */
	ino = (iagno << L2INOSPERIAG) + (extno << L2INOSPEREXT);

	/* initialize the inodes within the newly allocated extent a
	 * page at a time.
	 */
	for (i = 0; i < imap->im_nbperiext; i += sbi->nbperpage) {
		/* get a buffer for this page of disk inodes.
		 */
		dmp = get_metapage(ipimap, blkno + i, PSIZE, 1);
		if (dmp == NULL) {
			rc = -EIO;
			goto error_out;
		}
		dp = (struct dinode *) dmp->data;

		/* initialize the inode number, mode, link count and
		 * inode extent address.
		 */
		for (j = 0; j < INOSPERPAGE; j++, dp++, ino++) {
			dp->di_inostamp = cpu_to_le32(sbi->inostamp);
			dp->di_number = cpu_to_le32(ino);
			dp->di_fileset = cpu_to_le32(FILESYSTEM_I);
			dp->di_mode = 0;
			dp->di_nlink = 0;
			PXDaddress(&(dp->di_ixpxd), blkno);
			PXDlength(&(dp->di_ixpxd), imap->im_nbperiext);
		}
		write_metapage(dmp);
	}

	/* if this is the last free extent within the iag, remove the
	 * iag from the ag free extent list.
	 */
	if (iagp->nfreeexts == cpu_to_le32(1)) {
		if (fwd >= 0)
			aiagp->extfreeback = iagp->extfreeback;

		if (back >= 0)
			biagp->extfreefwd = iagp->extfreefwd;
		else
			imap->im_agctl[agno].extfree =
			    le32_to_cpu(iagp->extfreefwd);

		iagp->extfreefwd = iagp->extfreeback = cpu_to_le32(-1);
	} else {
		/* if the iag has all free extents (newly allocated iag),
		 * add the iag to the ag free extent list.
		 */
		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG)) {
			if (fwd >= 0)
				aiagp->extfreeback = cpu_to_le32(iagno);

			iagp->extfreefwd = cpu_to_le32(fwd);
			iagp->extfreeback = cpu_to_le32(-1);
			imap->im_agctl[agno].extfree = iagno;
		}
	}

	/* if the iag has no free inodes, add the iag to the
	 * ag free inode list.
	 */
	if (iagp->nfreeinos == 0) {
		if (freei >= 0)
			ciagp->inofreeback = cpu_to_le32(iagno);

		iagp->inofreefwd =
		    cpu_to_le32(imap->im_agctl[agno].inofree);
		iagp->inofreeback = cpu_to_le32(-1);
		imap->im_agctl[agno].inofree = iagno;
	}

	/* initialize the extent descriptor of the extent. */
	PXDlength(&iagp->inoext[extno], imap->im_nbperiext);
	PXDaddress(&iagp->inoext[extno], blkno);

	/* initialize the working and persistent map of the extent.
	 * the working map will be initialized such that
	 * it indicates the first inode of the extent is allocated.
	 */
	iagp->wmap[extno] = cpu_to_le32(HIGHORDER);
	iagp->pmap[extno] = 0;

	/* update the free inode and free extent summary maps
	 * for the extent to indicate the extent has free inodes
	 * and no longer represents a free extent.
	 */
	sword = extno >> L2EXTSPERSUM;
	mask = HIGHORDER >> (extno & (EXTSPERSUM - 1));
	iagp->extsmap[sword] |= cpu_to_le32(mask);
	iagp->inosmap[sword] &= cpu_to_le32(~mask);

	/* update the free inode and free extent counts for the
	 * iag.
	 */
	iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) +
				      (INOSPEREXT - 1));
	iagp->nfreeexts = cpu_to_le32(le32_to_cpu(iagp->nfreeexts) - 1);

	/* update the free and backed inode counts for the ag.
	 */
	imap->im_agctl[agno].numfree += (INOSPEREXT - 1);
	imap->im_agctl[agno].numinos += INOSPEREXT;

	/* update the free and backed inode counts for the inode map.
	 */
	atomic_add(INOSPEREXT - 1, &imap->im_numfree);
	atomic_add(INOSPEREXT, &imap->im_numinos);

	/* write the iags.
	 */
	if (amp)
		write_metapage(amp);
	if (bmp)
		write_metapage(bmp);
	if (cmp)
		write_metapage(cmp);

	return (0);

      error_out:

	/* release the iags.
	 */
	if (amp)
		release_metapage(amp);
	if (bmp)
		release_metapage(bmp);
	if (cmp)
		release_metapage(cmp);

	return (rc);
}


/*
 * NAME:        diNewIAG(imap,iagnop,agno)
 *
 * FUNCTION:   	allocate a new iag for an allocation group.
 *		
 *		first tries to allocate the iag from the inode map 
 *		iagfree list:  
 *		if the list has free iags, the head of the list is removed 
 *		and returned to satisfy the request.
 *		if the inode map's iag free list is empty, the inode map
 *		is extended to hold a new iag. this new iag is initialized
 *		and returned to satisfy the request.
 *
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      iagnop 	- pointer to an iag number set with the number of the
 *		  newly allocated iag upon successful return.
 *      agno  	- allocation group number.
 *	bpp	- Buffer pointer to be filled in with new IAG's buffer
 *
 * RETURN VALUES:
 *      0       - success.
 *      -ENOSPC	- insufficient disk resources.
 *      -EIO  	- i/o error.
 *
 * serialization: 
 *	AG lock held on entry/exit;
 *	write lock on the map is held inside;
 *	read lock on the map is held on successful completion;
 *
 * note: new iag transaction: 
 * . synchronously write iag;
 * . write log of xtree and inode  of imap;
 * . commit;
 * . synchronous write of xtree (right to left, bottom to top);
 * . at start of logredo(): init in-memory imap with one additional iag page;
 * . at end of logredo(): re-read imap inode to determine
 *   new imap size;
 */
static int
diNewIAG(struct inomap * imap, int *iagnop, int agno, struct metapage ** mpp)
{
	int rc;
	int iagno, i, xlen;
	struct inode *ipimap;
	struct super_block *sb;
	struct jfs_sb_info *sbi;
	struct metapage *mp;
	struct iag *iagp;
	s64 xaddr = 0;
	s64 blkno;
	tid_t tid;
#ifdef _STILL_TO_PORT
	xad_t xad;
#endif				/*  _STILL_TO_PORT */
	struct inode *iplist[1];

	/* pick up pointers to the inode map and mount inodes */
	ipimap = imap->im_ipimap;
	sb = ipimap->i_sb;
	sbi = JFS_SBI(sb);

	/* acquire the free iag lock */
	IAGFREE_LOCK(imap);

	/* if there are any iags on the inode map free iag list, 
	 * allocate the iag from the head of the list.
	 */
	if (imap->im_freeiag >= 0) {
		/* pick up the iag number at the head of the list */
		iagno = imap->im_freeiag;

		/* determine the logical block number of the iag */
		blkno = IAGTOLBLK(iagno, sbi->l2nbperpage);
	} else {
		/* no free iags. the inode map will have to be extented
		 * to include a new iag.
		 */

		/* acquire inode map lock */
		IWRITE_LOCK(ipimap);

		if (ipimap->i_size >> L2PSIZE != imap->im_nextiag + 1) {
			IWRITE_UNLOCK(ipimap);
			IAGFREE_UNLOCK(imap);
			jfs_error(imap->im_ipimap->i_sb,
				  "diNewIAG: ipimap->i_size is wrong");
			return -EIO;
		}


		/* get the next avaliable iag number */
		iagno = imap->im_nextiag;

		/* make sure that we have not exceeded the maximum inode
		 * number limit.
		 */
		if (iagno > (MAXIAGS - 1)) {
			/* release the inode map lock */
			IWRITE_UNLOCK(ipimap);

			rc = -ENOSPC;
			goto out;
		}

		/*
		 * synchronously append new iag page.
		 */
		/* determine the logical address of iag page to append */
		blkno = IAGTOLBLK(iagno, sbi->l2nbperpage);

		/* Allocate extent for new iag page */
		xlen = sbi->nbperpage;
		if ((rc = dbAlloc(ipimap, 0, (s64) xlen, &xaddr))) {
			/* release the inode map lock */
			IWRITE_UNLOCK(ipimap);

			goto out;
		}

2552
2553
2554
2555
2556
		/*
		 * start transaction of update of the inode map
		 * addressing structure pointing to the new iag page;
		 */
		tid = txBegin(sb, COMMIT_FORCE);
2557
		mutex_lock(&JFS_IP(ipimap)->commit_mutex);
2558
2559
2560
2561
2562

		/* update the inode map addressing structure to point to it */
		if ((rc =
		     xtInsert(tid, ipimap, 0, blkno, xlen, &xaddr, 0))) {
			txEnd(tid);
2563
			mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
2564
2565
2566
2567
2568
2569
2570
2571
			/* Free the blocks allocated for the iag since it was
			 * not successfully added to the inode map
			 */
			dbFree(ipimap, xaddr, (s64) xlen);

			/* release the inode map lock */
			IWRITE_UNLOCK(ipimap);

2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
			goto out;
		}

		/* update the inode map's inode to reflect the extension */
		ipimap->i_size += PSIZE;
		inode_add_bytes(ipimap, PSIZE);

		/* assign a buffer for the page */
		mp = get_metapage(ipimap, blkno, PSIZE, 0);
		if (!mp) {
			/*
			 * This is very unlikely since we just created the
			 * extent, but let's try to handle it correctly
			 */
			xtTruncate(tid, ipimap, ipimap->i_size - PSIZE,
				   COMMIT_PWMAP);

			txAbort(tid, 0);
			txEnd(tid);

			/* release the inode map lock */
			IWRITE_UNLOCK(ipimap);

Linus Torvalds's avatar
Linus Torvalds committed
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
			rc = -EIO;
			goto out;
		}
		iagp = (struct iag *) mp->data;

		/* init the iag */
		memset(iagp, 0, sizeof(struct iag));
		iagp->iagnum = cpu_to_le32(iagno);
		iagp->inofreefwd = iagp->inofreeback = cpu_to_le32(-1);
		iagp->extfreefwd = iagp->extfreeback = cpu_to_le32(-1);
		iagp->iagfree = cpu_to_le32(-1);
		iagp->nfreeinos = 0;
		iagp->nfreeexts = cpu_to_le32(EXTSPERIAG);

		/* initialize the free inode summary map (free extent
		 * summary map initialization handled by bzero).
		 */
		for (i = 0; i < SMAPSZ; i++)
			iagp->inosmap[i] = cpu_to_le32(ONES);

		/*
2616
		 * Write and sync the metapage
Linus Torvalds's avatar
Linus Torvalds committed
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
		 */
		flush_metapage(mp);

		/*
		 * txCommit(COMMIT_FORCE) will synchronously write address 
		 * index pages and inode after commit in careful update order 
		 * of address index pages (right to left, bottom up);
		 */
		iplist[0] = ipimap;
		rc = txCommit(tid, 1, &iplist[0], COMMIT_FORCE);

		txEnd(tid);
2629
		mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769

		duplicateIXtree(sb, blkno, xlen, &xaddr);

		/* update the next avaliable iag number */
		imap->im_nextiag += 1;

		/* Add the iag to the iag free list so we don't lose the iag
		 * if a failure happens now.
		 */
		imap->im_freeiag = iagno;

		/* Until we have logredo working, we want the imap inode &
		 * control page to be up to date.
		 */
		diSync(ipimap);

		/* release the inode map lock */
		IWRITE_UNLOCK(ipimap);
	}

	/* obtain read lock on map */
	IREAD_LOCK(ipimap);

	/* read the iag */
	if ((rc = diIAGRead(imap, iagno, &mp))) {
		IREAD_UNLOCK(ipimap);
		rc = -EIO;
		goto out;
	}
	iagp = (struct iag *) mp->data;

	/* remove the iag from the iag free list */
	imap->im_freeiag = le32_to_cpu(iagp->iagfree);
	iagp->iagfree = cpu_to_le32(-1);

	/* set the return iag number and buffer pointer */
	*iagnop = iagno;
	*mpp = mp;

      out:
	/* release the iag free lock */
	IAGFREE_UNLOCK(imap);

	return (rc);
}

/*
 * NAME:        diIAGRead()
 *
 * FUNCTION:    get the buffer for the specified iag within a fileset
 *		or aggregate inode map.
 *		
 * PARAMETERS:
 *      imap  	- pointer to inode map control structure.
 *      iagno  	- iag number.
 *      bpp  	- point to buffer pointer to be filled in on successful
 *		  exit.
 *
 * SERIALIZATION:
 *	must have read lock on imap inode
 *	(When called by diExtendFS, the filesystem is quiesced, therefore
 *	 the read lock is unnecessary.)
 *
 * RETURN VALUES:
 *      0       - success.
 *      -EIO  	- i/o error.
 */
static int diIAGRead(struct inomap * imap, int iagno, struct metapage ** mpp)
{
	struct inode *ipimap = imap->im_ipimap;
	s64 blkno;

	/* compute the logical block number of the iag. */
	blkno = IAGTOLBLK(iagno, JFS_SBI(ipimap->i_sb)->l2nbperpage);

	/* read the iag. */
	*mpp = read_metapage(ipimap, blkno, PSIZE, 0);
	if (*mpp == NULL) {
		return -EIO;
	}

	return (0);
}

/*
 * NAME:        diFindFree()
 *
 * FUNCTION:    find the first free bit in a word starting at
 *		the specified bit position.
 *
 * PARAMETERS:
 *      word  	- word to be examined.
 *      start  	- starting bit position.
 *
 * RETURN VALUES:
 *      bit position of first free bit in the word or 32 if
 *	no free bits were found.
 */
static int diFindFree(u32 word, int start)
{
	int bitno;
	assert(start < 32);
	/* scan the word for the first free bit. */
	for (word <<= start, bitno = start; bitno < 32;
	     bitno++, word <<= 1) {
		if ((word & HIGHORDER) == 0)
			break;
	}
	return (bitno);
}

/*
 * NAME:	diUpdatePMap()
 *                                                                    
 * FUNCTION: Update the persistent map in an IAG for the allocation or 
 *	freeing of the specified inode.
 *                                                                    
 * PRE CONDITIONS: Working map has already been updated for allocate.
 *
 * PARAMETERS:
 *	ipimap	- Incore inode map inode
 *	inum	- Number of inode to mark in permanent map
 *	is_free	- If TRUE indicates inode should be marked freed, otherwise
 *		  indicates inode should be marked allocated.
 *
 * RETURN VALUES: 
 *		0 for success
 */
int
diUpdatePMap(struct inode *ipimap,
	     unsigned long inum, boolean_t is_free, struct tblock * tblk)
{
	int rc;
	struct iag *iagp;
	struct metapage *mp;
	int iagno, ino, extno, bitno;
	struct inomap *imap;
	u32 mask;
	struct jfs_log *log;
	int lsn, difft, diffp;
2770
	unsigned long flags;
Linus Torvalds's avatar
Linus Torvalds committed
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786

	imap = JFS_IP(ipimap)->i_imap;
	/* get the iag number containing the inode */
	iagno = INOTOIAG(inum);
	/* make sure that the iag is contained within the map */
	if (iagno >= imap->im_nextiag) {
		jfs_error(ipimap->i_sb,
			  "diUpdatePMap: the iag is outside the map");
		return -EIO;
	}
	/* read the iag */
	IREAD_LOCK(ipimap);
	rc = diIAGRead(imap, iagno, &mp);
	IREAD_UNLOCK(ipimap);
	if (rc)
		return (rc);
2787
	metapage_wait_for_io(mp);
Linus Torvalds's avatar
Linus Torvalds committed
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
	iagp = (struct iag *) mp->data;
	/* get the inode number and extent number of the inode within
	 * the iag and the inode number within the extent.
	 */
	ino = inum & (INOSPERIAG - 1);
	extno = ino >> L2INOSPEREXT;
	bitno = ino & (INOSPEREXT - 1);
	mask = HIGHORDER >> bitno;
	/* 
	 * mark the inode free in persistent map:
	 */
	if (is_free == TRUE) {
		/* The inode should have been allocated both in working
		 * map and in persistent map;
		 * the inode will be freed from working map at the release
		 * of last reference release;
		 */
		if (!(le32_to_cpu(iagp->wmap[extno]) & mask)) {
			jfs_error(ipimap->i_sb, 
				  "diUpdatePMap: inode %ld not marked as "
				  "allocated in wmap!", inum);
		}
		if (!(le32_to_cpu(iagp->pmap[extno]) & mask)) {
			jfs_error(ipimap->i_sb,
				  "diUpdatePMap: inode %ld not marked as "
				  "allocated in pmap!", inum);
		}
		/* update the bitmap for the extent of the freed inode */
		iagp->pmap[extno] &= cpu_to_le32(~mask);
	}
	/*
	 * mark the inode allocated in persistent map:
	 */
	else {
		/* The inode should be already allocated in the working map
		 * and should be free in persistent map;
		 */
		if (!(le32_to_cpu(iagp->wmap[extno]) & mask)) {
			release_metapage(mp);
			jfs_error(ipimap->i_sb,
				  "diUpdatePMap: the inode is not allocated in "
				  "the working map");
			return -EIO;
		}
		if ((le32_to_cpu(iagp->pmap[extno]) & mask) != 0) {
			release_metapage(mp);
			jfs_error(ipimap->i_sb,
				  "diUpdatePMap: the inode is not free in the "
				  "persistent map");
			return -EIO;
		}
		/* update the bitmap for the extent of the allocated inode */
		iagp->pmap[extno] |= cpu_to_le32(mask);
	}
	/*
	 * update iag lsn
	 */
	lsn = tblk->lsn;
	log = JFS_SBI(tblk->sb)->log;
2847
	LOGSYNC_LOCK(log, flags);
Linus Torvalds's avatar
Linus Torvalds committed
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
	if (mp->lsn != 0) {
		/* inherit older/smaller lsn */
		logdiff(difft, lsn, log);
		logdiff(diffp, mp->lsn, log);
		if (difft < diffp) {
			mp->lsn = lsn;
			/* move mp after tblock in logsync list */
			list_move(&mp->synclist, &tblk->synclist);
		}
		/* inherit younger/larger clsn */
		assert(mp->clsn);
		logdiff(difft, tblk->clsn, log);
		logdiff(diffp, mp->clsn, log);
		if (difft > diffp)
			mp->clsn = tblk->clsn;
	} else {
		mp->log = log;
		mp->lsn = lsn;
		/* insert mp after tblock in logsync list */
		log->count++;
		list_add(&mp->synclist, &tblk->synclist);
		mp->clsn = tblk->clsn;
	}
2871
	LOGSYNC_UNLOCK(log, flags);
Linus Torvalds's avatar
Linus Torvalds committed
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885