summaryrefslogtreecommitdiffstats
path: root/fs/nfs/dir.c
blob: 82d3f5bc35ff9dbd73c05ffc613da95070494a70 (plain)
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
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
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
268
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
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
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
567
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
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
/*
 *  linux/fs/nfs/dir.c
 *
 *  Copyright (C) 1992  Rick Sladkey
 *
 *  nfs directory handling functions
 *
 * 10 Apr 1996	Added silly rename for unlink	--okir
 * 28 Sep 1996	Improved directory cache --okir
 * 23 Aug 1997  Claus Heine claus@momo.math.rwth-aachen.de 
 *              Re-implemented silly rename for unlink, newly implemented
 *              silly rename for nfs_rename() following the suggestions
 *              of Olaf Kirch (okir) found in this file.
 *              Following Linus comments on my original hack, this version
 *              depends only on the dcache stuff and doesn't touch the inode
 *              layer (iput() and friends).
 *  6 Jun 1999	Cache readdir lookups in the page cache. -DaveM
 */

#define NFS_NEED_XDR_TYPES
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/string.h>
#include <linux/kernel.h>
#include <linux/malloc.h>
#include <linux/mm.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs_fs.h>
#include <linux/nfs.h>
#include <linux/pagemap.h>

#include <asm/segment.h>	/* for fs functions */

#define NFS_PARANOIA 1
/* #define NFS_DEBUG_VERBOSE 1 */

static int nfs_safe_remove(struct dentry *);

static ssize_t nfs_dir_read(struct file *, char *, size_t, loff_t *);
static int nfs_readdir(struct file *, void *, filldir_t);
static struct dentry *nfs_lookup(struct inode *, struct dentry *);
static int nfs_create(struct inode *, struct dentry *, int);
static int nfs_mkdir(struct inode *, struct dentry *, int);
static int nfs_rmdir(struct inode *, struct dentry *);
static int nfs_unlink(struct inode *, struct dentry *);
static int nfs_symlink(struct inode *, struct dentry *, const char *);
static int nfs_link(struct dentry *, struct inode *, struct dentry *);
static int nfs_mknod(struct inode *, struct dentry *, int, int);
static int nfs_rename(struct inode *, struct dentry *,
		      struct inode *, struct dentry *);

static struct file_operations nfs_dir_operations = {
	read:		nfs_dir_read,
	readdir:	nfs_readdir,
	open:		nfs_open,
	release:	nfs_release,
};

struct inode_operations nfs_dir_inode_operations = {
	&nfs_dir_operations,	/* default directory file-ops */
	nfs_create,		/* create */
	nfs_lookup,		/* lookup */
	nfs_link,		/* link */
	nfs_unlink,		/* unlink */
	nfs_symlink,		/* symlink */
	nfs_mkdir,		/* mkdir */
	nfs_rmdir,		/* rmdir */
	nfs_mknod,		/* mknod */
	nfs_rename,		/* rename */
	NULL,			/* readlink */
	NULL,			/* follow_link */
	NULL,			/* truncate */
	NULL,			/* permission */
	nfs_revalidate,		/* revalidate */
};

static ssize_t
nfs_dir_read(struct file *filp, char *buf, size_t count, loff_t *ppos)
{
	return -EISDIR;
}

/* Each readdir response is composed of entries which look
 * like the following, as per the NFSv2 RFC:
 *
 *	__u32	not_end			zero if end of response
 *	__u32	file ID			opaque ino_t
 *	__u32	namelen			size of name string
 *	VAR	name string		the string, padded to modulo 4 bytes
 *	__u32	cookie			opaque ID of next entry
 *
 * When you hit not_end being zero, the next __u32 is non-zero if
 * this is the end of the complete set of readdir entires for this
 * directory.  This can be used, for example, to initiate pre-fetch.
 *
 * In order to know what to ask the server for, we only need to know
 * the final cookie of the previous page, and offset zero has cookie
 * zero, so we cache cookie to page offset translations in chunks.
 */
#define COOKIES_PER_CHUNK (8 - ((sizeof(void *) / sizeof(__u32))))
struct nfs_cookie_table {
	struct nfs_cookie_table *next;
	__u32	cookies[COOKIES_PER_CHUNK];
};
static kmem_cache_t *nfs_cookie_cachep;

/* This whole scheme relies on the fact that dirent cookies
 * are monotonically increasing.
 *
 * Another invariant is that once we have a valid non-zero
 * EOF marker cached, we also have the complete set of cookie
 * table entries.
 *
 * We return the page offset assosciated with the page where
 * cookie must be if it exists at all, however if we can not
 * figure that out conclusively, we return < 0.
 */
static long __nfs_readdir_offset(struct inode *inode, __u32 cookie)
{
	struct nfs_cookie_table *p;
	unsigned long ret = 0;

	for(p = NFS_COOKIES(inode); p != NULL; p = p->next) {
		int i;

		for (i = 0; i < COOKIES_PER_CHUNK; i++) {
			__u32 this_cookie = p->cookies[i];

			/* End of known cookies, EOF is our only hope. */
			if (!this_cookie)
				goto check_eof;

			/* Next cookie is larger, must be in previous page. */
			if (this_cookie > cookie)
				return ret;

			ret += 1;

			/* Exact cookie match, it must be in this page :-) */
			if (this_cookie == cookie)
				return ret;
		}
	}
check_eof:
	if (NFS_DIREOF(inode) != 0)
		return ret;

	return -1L;
}

static __inline__ long nfs_readdir_offset(struct inode *inode, __u32 cookie)
{
	/* Cookie zero is always at page offset zero.   Optimize the
	 * other common case since most directories fit entirely
	 * in one page.
	 */
	if (!cookie || (!NFS_COOKIES(inode) && NFS_DIREOF(inode)))
		return 0;
	return __nfs_readdir_offset(inode, cookie);
}

/* Since a cookie of zero is declared special by the NFS
 * protocol, we easily can tell if a cookie in an existing
 * table chunk is valid or not.
 *
 * NOTE: The cookies are indexed off-by-one because zero
 *       need not an entry.
 */
static __inline__ __u32 *find_cookie(struct inode *inode, unsigned long off)
{
	static __u32 cookie_zero = 0;
	struct nfs_cookie_table *p;
	__u32 *ret;

	if (!off)
		return &cookie_zero;
	off -= 1;
	p = NFS_COOKIES(inode);
	while(off >= COOKIES_PER_CHUNK && p) {
		off -= COOKIES_PER_CHUNK;
		p = p->next;
	}
	ret = NULL;
	if (p) {
		ret = &p->cookies[off];
		if (!*ret)
			ret = NULL;
	}
	return ret;
}

#define NFS_NAMELEN_ALIGN(__len) ((((__len)+3)>>2)<<2)
static int create_cookie(__u32 cookie, unsigned long off, struct inode *inode)
{
	struct nfs_cookie_table **cpp;

	cpp = (struct nfs_cookie_table **) &NFS_COOKIES(inode);
	while (off >= COOKIES_PER_CHUNK && *cpp) {
		off -= COOKIES_PER_CHUNK;
		cpp = &(*cpp)->next;
	}
	if (*cpp) {
		(*cpp)->cookies[off] = cookie;
	} else {
		struct nfs_cookie_table *new;
		int i;

		new = kmem_cache_alloc(nfs_cookie_cachep, SLAB_ATOMIC);
		if(!new)
			return -1;
		*cpp = new;
		new->next = NULL;
		for(i = 0; i < COOKIES_PER_CHUNK; i++) {
			if (i == off) {
				new->cookies[i] = cookie;
			} else {
				new->cookies[i] = 0;
			}
		}
	}
	return 0;
}

static struct page *try_to_get_dirent_page(struct file *, __u32, int);

/* Recover from a revalidation flush.  The case here is that
 * the inode for the directory got invalidated somehow, and
 * all of our cached information is lost.  In order to get
 * a correct cookie for the current readdir request from the
 * user, we must (re-)fetch older readdir page cache entries.
 *
 * Returns < 0 if some error occurrs, else it is the page offset
 * to fetch.
 */
static long refetch_to_readdir_cookie(struct file *file, struct inode *inode)
{
	struct page *page;
	u32 goal_cookie = file->f_pos;
	long cur_off, ret = -1L;

again:
	cur_off = 0;
	for (;;) {
		page = find_get_page(&inode->i_data, cur_off);
		if (page) {
			if (!Page_Uptodate(page))
				goto out_error;
		} else {
			__u32 *cp = find_cookie(inode, cur_off);

			if (!cp)
				goto out_error;

			page = try_to_get_dirent_page(file, *cp, 0);
			if (!page) {
				if (!cur_off)
					goto out_error;

				/* Someone touched the dir on us. */
				goto again;
			}
		}
		page_cache_release(page);

		if ((ret = nfs_readdir_offset(inode, goal_cookie)) >= 0)
			goto out;

		cur_off += 1;
	}
out:
	return ret;

out_error:
	if (page)
		page_cache_release(page);
	goto out;
}

/* Now we cache directories properly, by stuffing the dirent
 * data directly in the page cache.
 *
 * Inode invalidation due to refresh etc. takes care of
 * _everything_, no sloppy entry flushing logic, no extraneous
 * copying, network direct to page cache, the way it was meant
 * to be.
 *
 * NOTE: Dirent information verification is done always by the
 *	 page-in of the RPC reply, nowhere else, this simplies
 *	 things substantially.
 */

static int nfs_dir_filler(struct dentry *dentry, struct page *page)
{
	struct nfs_readdirargs rd_args;
	struct nfs_readdirres rd_res;
	struct inode *inode = dentry->d_inode;
	long offset = page->index;
	__u32 *cookiep;
	int err;

	kmap(page);

	err = -EIO;
	cookiep = find_cookie(inode, offset);
	if (!cookiep)
		goto fail;

	rd_args.fh = NFS_FH(dentry);
	rd_res.buffer = (char *)page_address(page);
	rd_res.bufsiz = PAGE_CACHE_SIZE;
	rd_res.cookie = *cookiep;
	do {
		rd_args.buffer = rd_res.buffer;
		rd_args.bufsiz = rd_res.bufsiz;
		rd_args.cookie = rd_res.cookie;
		err = rpc_call(NFS_CLIENT(inode),
			     NFSPROC_READDIR, &rd_args, &rd_res, 0); 
		if (err < 0)
			goto fail;
	} while(rd_res.bufsiz > 0);

	err = -EIO;
	if (rd_res.bufsiz < 0)
		NFS_DIREOF(inode) = rd_res.cookie;
	else if (create_cookie(rd_res.cookie, offset, inode))
		goto fail;

	SetPageUptodate(page);
	kunmap(page);
	UnlockPage(page);
	return 0;
fail:
	SetPageError(page);
	kunmap(page);
	UnlockPage(page);
	return err;
}

static struct page *try_to_get_dirent_page(struct file *file, __u32 cookie, int refetch_ok)
{
	struct dentry *dentry = file->f_dentry;
	struct inode *inode = dentry->d_inode;
	struct page *page;
	long offset;

	if ((offset = nfs_readdir_offset(inode, cookie)) < 0) {
		if (!refetch_ok ||
		    (offset = refetch_to_readdir_cookie(file, inode)) < 0) {
			goto fail;
		}
	}

	page = read_cache_page(&inode->i_data, offset,
				(filler_t *)nfs_dir_filler, dentry);
	if (IS_ERR(page))
		goto fail;
	if (!Page_Uptodate(page))
		goto fail2;
	return page;

fail2:
	page_cache_release(page);
fail:
	return NULL;
}

/* Seek up to dirent assosciated with the passed in cookie,
 * then fill in dirents found.  Return the last cookie
 * actually given to the user, to update the file position.
 */
static __inline__ u32 nfs_do_filldir(__u32 *p, u32 cookie,
				     void *dirent, filldir_t filldir)
{
	u32 end;

	while((end = *p++) != 0) {
		__u32 fileid, len, skip, this_cookie;
		char *name;

		fileid = *p++;
		len = *p++;
		name = (char *) p;
		skip = NFS_NAMELEN_ALIGN(len);
		p += (skip >> 2);
		this_cookie = *p++;

		if (this_cookie < cookie)
			continue;

		cookie = this_cookie;
		if (filldir(dirent, name, len, cookie, fileid) < 0)
			break;
	}

	return cookie;
}

/* The file offset position is represented in pure bytes, to
 * make the page cache interface straight forward.
 *
 * However, some way is needed to make the connection between the
 * opaque NFS directory entry cookies and our offsets, so a per-inode
 * cookie cache table is used.
 */
static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
	struct dentry *dentry = filp->f_dentry;
	struct inode *inode = dentry->d_inode;
	struct page *page;
	long offset;
	int res;

	res = nfs_revalidate_inode(NFS_DSERVER(dentry), dentry);
	if (res < 0)
		return res;

	if (NFS_DIREOF(inode) && filp->f_pos >= NFS_DIREOF(inode))
		return 0;

	if ((offset = nfs_readdir_offset(inode, filp->f_pos)) < 0)
		goto no_dirent_page;

	page = find_get_page(&inode->i_data, offset);
	if (!page)
		goto no_dirent_page;
	if (!Page_Uptodate(page))
		goto dirent_read_error;
success:
	kmap(page);
	filp->f_pos = nfs_do_filldir((__u32 *) page_address(page),
				     filp->f_pos, dirent, filldir);
	kunmap(page);
	page_cache_release(page);
	return 0;

no_dirent_page:
	page = try_to_get_dirent_page(filp, filp->f_pos, 1);
	if (!page)
		goto no_page;

	if (Page_Uptodate(page))
		goto success;
dirent_read_error:
	page_cache_release(page);
no_page:
	return -EIO;
}

/* Flush directory cookie and EOF caches for an inode.
 * So we don't thrash allocating/freeing cookie tables,
 * we keep the cookies around until the inode is
 * deleted/reused.
 */
__inline__ void nfs_flush_dircache(struct inode *inode)
{
	struct nfs_cookie_table *p = NFS_COOKIES(inode);

	while (p != NULL) {
		int i;

		for(i = 0; i < COOKIES_PER_CHUNK; i++)
			p->cookies[i] = 0;

		p = p->next;
	}
	NFS_DIREOF(inode) = 0;
}

/* Free up directory cache state, this happens when
 * nfs_delete_inode is called on an NFS directory.
 */
void nfs_free_dircache(struct inode *inode)
{
	struct nfs_cookie_table *p = NFS_COOKIES(inode);

	while (p != NULL) {
		struct nfs_cookie_table *next = p->next;
		kmem_cache_free(nfs_cookie_cachep, p);
		p = next;
	}
	NFS_COOKIES(inode) = NULL;
	NFS_DIREOF(inode) = 0;
}

/*
 * Whenever an NFS operation succeeds, we know that the dentry
 * is valid, so we update the revalidation timestamp.
 */
static inline void nfs_renew_times(struct dentry * dentry)
{
	dentry->d_time = jiffies;
}

static inline int nfs_dentry_force_reval(struct dentry *dentry, int flags)
{
	struct inode *inode = dentry->d_inode;
	unsigned long timeout = NFS_ATTRTIMEO(inode);

	/*
	 * If it's the last lookup in a series, we use a stricter
	 * cache consistency check by looking at the parent mtime.
	 *
	 * If it's been modified in the last hour, be really strict.
	 * (This still means that we can avoid doing unnecessary
	 * work on directories like /usr/share/bin etc which basically
	 * never change).
	 */
	if (!(flags & LOOKUP_CONTINUE)) {
		long diff = CURRENT_TIME - dentry->d_parent->d_inode->i_mtime;

		if (diff < 15*60)
			timeout = 0;
	}
	
	return time_after(jiffies,dentry->d_time + timeout);
}

/*
 * We judge how long we want to trust negative
 * dentries by looking at the parent inode mtime.
 *
 * If mtime is close to present time, we revalidate
 * more often.
 */
#define NFS_REVALIDATE_NEGATIVE (1 * HZ)
static inline int nfs_neg_need_reval(struct dentry *dentry)
{
	struct inode *dir = dentry->d_parent->d_inode;
	unsigned long timeout = NFS_ATTRTIMEO(dir);
	long diff = CURRENT_TIME - dir->i_mtime;

	if (diff < 5*60 && timeout > NFS_REVALIDATE_NEGATIVE)
		timeout = NFS_REVALIDATE_NEGATIVE;

	return time_after(jiffies, dentry->d_time + timeout);
}

/*
 * This is called every time the dcache has a lookup hit,
 * and we should check whether we can really trust that
 * lookup.
 *
 * NOTE! The hit can be a negative hit too, don't assume
 * we have an inode!
 *
 * If the dentry is older than the revalidation interval, 
 * we do a new lookup and verify that the dentry is still
 * correct.
 */
static int nfs_lookup_revalidate(struct dentry * dentry, int flags)
{
	struct dentry * parent = dentry->d_parent;
	struct inode * inode = dentry->d_inode;
	int error;
	struct nfs_fh fhandle;
	struct nfs_fattr fattr;

	/*
	 * If we don't have an inode, let's look at the parent
	 * directory mtime to get a hint about how often we
	 * should validate things..
	 */
	if (!inode) {
		if (nfs_neg_need_reval(dentry))
			goto out_bad;
		goto out_valid;
	}

	if (is_bad_inode(inode)) {
		dfprintk(VFS, "nfs_lookup_validate: %s/%s has dud inode\n",
			parent->d_name.name, dentry->d_name.name);
		goto out_bad;
	}

	if (!nfs_dentry_force_reval(dentry, flags))
		goto out_valid;

	if (IS_ROOT(dentry)) {
		__nfs_revalidate_inode(NFS_DSERVER(dentry), dentry);
		goto out_valid_renew;
	}

	/*
	 * Do a new lookup and check the dentry attributes.
	 */
	error = nfs_proc_lookup(NFS_DSERVER(parent), NFS_FH(parent),
				dentry->d_name.name, &fhandle, &fattr);
	if (error)
		goto out_bad;

	/* Inode number matches? */
	if (NFS_FSID(inode) != fattr.fsid ||
	    NFS_FILEID(inode) != fattr.fileid)
		goto out_bad;

	/* Filehandle matches? */
	if (memcmp(dentry->d_fsdata, &fhandle, sizeof(struct nfs_fh)))
		goto out_bad;

	/* Ok, remeber that we successfully checked it.. */
	nfs_refresh_inode(inode, &fattr);

 out_valid_renew:
	nfs_renew_times(dentry);
out_valid:
	return 1;
out_bad:
	d_drop(dentry);
	if (!list_empty(&dentry->d_subdirs))
		shrink_dcache_parent(dentry);
	/* Purge readdir caches. */
	if (dentry->d_parent->d_inode) {
		nfs_zap_caches(dentry->d_parent->d_inode);
		NFS_CACHEINV(dentry->d_parent->d_inode);
	}
	return 0;
}

/*
 * This is called from dput() when d_count is going to 0.
 * We use it to clean up silly-renamed files.
 */
static void nfs_dentry_delete(struct dentry *dentry)
{
	dfprintk(VFS, "NFS: dentry_delete(%s/%s, %x)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name,
		dentry->d_flags);

	if (dentry->d_flags & DCACHE_NFSFS_RENAMED) {
		int error;
		
		dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
		/* Unhash it first */
		d_drop(dentry);
		error = nfs_safe_remove(dentry);
		if (error)
			printk("NFS: can't silly-delete %s/%s, error=%d\n",
				dentry->d_parent->d_name.name,
				dentry->d_name.name, error);
	}

}

static kmem_cache_t *nfs_fh_cachep;

__inline__ struct nfs_fh *nfs_fh_alloc(void)
{
	return kmem_cache_alloc(nfs_fh_cachep, SLAB_KERNEL);
}

__inline__ void nfs_fh_free(struct nfs_fh *p)
{
	kmem_cache_free(nfs_fh_cachep, p);
}

/*
 * Called when the dentry is being freed to release private memory.
 */
static void nfs_dentry_release(struct dentry *dentry)
{
	if (dentry->d_fsdata)
		nfs_fh_free(dentry->d_fsdata);
}

struct dentry_operations nfs_dentry_operations = {
	nfs_lookup_revalidate,	/* d_revalidate(struct dentry *, int) */
	NULL,			/* d_hash */
	NULL,			/* d_compare */
	nfs_dentry_delete,	/* d_delete(struct dentry *) */
	nfs_dentry_release,	/* d_release(struct dentry *) */
	NULL			/* d_iput */
};

#if 0 /* dead code */
#ifdef NFS_PARANOIA
/*
 * Display all dentries holding the specified inode.
 */
static void show_dentry(struct list_head * dlist)
{
	struct list_head *tmp = dlist;

	while ((tmp = tmp->next) != dlist) {
		struct dentry * dentry = list_entry(tmp, struct dentry, d_alias);
		const char * unhashed = "";

		if (list_empty(&dentry->d_hash))
			unhashed = "(unhashed)";

		printk("show_dentry: %s/%s, d_count=%d%s\n",
			dentry->d_parent->d_name.name,
			dentry->d_name.name, dentry->d_count,
			unhashed);
	}
}
#endif /* NFS_PARANOIA */
#endif /* 0 */

static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry)
{
	struct inode *inode;
	int error;
	struct nfs_fh fhandle;
	struct nfs_fattr fattr;

	dfprintk(VFS, "NFS: lookup(%s/%s)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name);

	error = -ENAMETOOLONG;
	if (dentry->d_name.len > NFS_MAXNAMLEN)
		goto out;

	error = -ENOMEM;
	if (!dentry->d_fsdata) {
		dentry->d_fsdata = nfs_fh_alloc();
		if (!dentry->d_fsdata)
			goto out;
	}
	dentry->d_op = &nfs_dentry_operations;

	error = nfs_proc_lookup(NFS_SERVER(dir), NFS_FH(dentry->d_parent), 
				dentry->d_name.name, &fhandle, &fattr);
	inode = NULL;
	if (error == -ENOENT)
		goto no_entry;
	if (!error) {
		error = -EACCES;
		inode = nfs_fhget(dentry, &fhandle, &fattr);
		if (inode) {
	    no_entry:
			d_add(dentry, inode);
			nfs_renew_times(dentry);
			error = 0;
		}
	}
out:
	return ERR_PTR(error);
}

/*
 * Code common to create, mkdir, and mknod.
 */
static int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
				struct nfs_fattr *fattr)
{
	struct inode *inode;
	int error = -EACCES;

	inode = nfs_fhget(dentry, fhandle, fattr);
	if (inode) {
		d_instantiate(dentry, inode);
		nfs_renew_times(dentry);
		error = 0;
	}
	return error;
}

/*
 * Following a failed create operation, we drop the dentry rather
 * than retain a negative dentry. This avoids a problem in the event
 * that the operation succeeded on the server, but an error in the
 * reply path made it appear to have failed.
 */
static int nfs_create(struct inode *dir, struct dentry *dentry, int mode)
{
	int error;
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;

	dfprintk(VFS, "NFS: create(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	attr.ia_mode = mode;
	attr.ia_valid = ATTR_MODE;

	/*
	 * Invalidate the dir cache before the operation to avoid a race.
	 */
	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_create(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
			dentry->d_name.name, &attr, &fhandle, &fattr);
	if (!error)
		error = nfs_instantiate(dentry, &fhandle, &fattr);
	if (error)
		d_drop(dentry);
	return error;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int nfs_mknod(struct inode *dir, struct dentry *dentry, int mode, int rdev)
{
	int error;
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;

	dfprintk(VFS, "NFS: mknod(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	attr.ia_mode = mode;
	attr.ia_valid = ATTR_MODE;
	/* FIXME: move this to a special nfs_proc_mknod() */
	if (S_ISCHR(mode) || S_ISBLK(mode)) {
		attr.ia_size = rdev; /* get out your barf bag */
		attr.ia_valid |= ATTR_SIZE;
	}

	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_create(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
				dentry->d_name.name, &attr, &fhandle, &fattr);
	if (!error)
		error = nfs_instantiate(dentry, &fhandle, &fattr);
	if (error)
		d_drop(dentry);
	return error;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int nfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
	int error;
	struct iattr attr;
	struct nfs_fattr fattr;
	struct nfs_fh fhandle;

	dfprintk(VFS, "NFS: mkdir(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	attr.ia_valid = ATTR_MODE;
	attr.ia_mode = mode | S_IFDIR;

	/*
	 * Always drop the dentry, we can't always depend on
	 * the fattr returned by the server (AIX seems to be
	 * broken). We're better off doing another lookup than
	 * depending on potentially bogus information.
	 */
	d_drop(dentry);
	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_mkdir(NFS_DSERVER(dentry), NFS_FH(dentry->d_parent),
				dentry->d_name.name, &attr, &fhandle, &fattr);
	if (!error)
		dir->i_nlink++;
	return error;
}

static int nfs_rmdir(struct inode *dir, struct dentry *dentry)
{
	int error;

	dfprintk(VFS, "NFS: rmdir(%x/%ld, %s\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_rmdir(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
				dentry->d_name.name);

	/* Update i_nlink and invalidate dentry. */
	if (!error) {
		d_drop(dentry);
		if (dir->i_nlink)
			dir->i_nlink--;
	}

	return error;
}


/*  Note: we copy the code from lookup_dentry() here, only: we have to
 *  omit the directory lock. We are already the owner of the lock when
 *  we reach here. And "down(&dir->i_sem)" would make us sleep forever
 *  ('cause WE have the lock)
 * 
 *  VERY IMPORTANT: calculate the hash for this dentry!!!!!!!!
 *  Otherwise the cached lookup DEFINITELY WILL fail. And a new dentry
 *  is created. Without the DCACHE_NFSFS_RENAMED flag. And with d_count
 *  == 1. And trouble.
 *
 *  Concerning my choice of the temp name: it is just nice to have
 *  i_ino part of the temp name, as this offers another check whether
 *  somebody attempts to remove the "silly renamed" dentry itself.
 *  Which is something that I consider evil. Your opinion may vary.
 *  BUT:
 *  Now that I compute the hash value right, it should be possible to simply
 *  check for the DCACHE_NFSFS_RENAMED flag in dentry->d_flag instead of
 *  doing the string compare.
 *  WHICH MEANS:
 *  This offers the opportunity to shorten the temp name. Currently, I use
 *  the hex representation of i_ino + an event counter. This sums up to
 *  as much as 36 characters for a 64 bit machine, and needs 20 chars on 
 *  a 32 bit machine.
 *  QUINTESSENCE
 *  The use of i_ino is simply cosmetic. All we need is a unique temp
 *  file name for the .nfs files. The event counter seemed to be adequate.
 *  And as we retry in case such a file already exists, we are guaranteed
 *  to succeed.
 */

static
struct dentry *nfs_silly_lookup(struct dentry *parent, char *silly, int slen)
{
	struct qstr    sqstr;
	struct dentry *sdentry;
	struct dentry *res;

	sqstr.name = silly;
	sqstr.len  = slen;
	sqstr.hash = full_name_hash(silly, slen);
	sdentry = d_lookup(parent, &sqstr);
	if (!sdentry) {
		sdentry = d_alloc(parent, &sqstr);
		if (sdentry == NULL)
			return ERR_PTR(-ENOMEM);
		res = nfs_lookup(parent->d_inode, sdentry);
		if (res) {
			dput(sdentry);
			return res;
		}
	}
	return sdentry;
}

static int nfs_sillyrename(struct inode *dir, struct dentry *dentry)
{
	static unsigned int sillycounter = 0;
	const int      i_inosize  = sizeof(dir->i_ino)*2;
	const int      countersize = sizeof(sillycounter)*2;
	const int      slen       = strlen(".nfs") + i_inosize + countersize;
	char           silly[slen+1];
	struct dentry *sdentry;
	int            error = -EIO;

	dfprintk(VFS, "NFS: silly-rename(%s/%s, ct=%d)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name, 
		dentry->d_count);

	/*
	 * Note that a silly-renamed file can be deleted once it's
	 * no longer in use -- it's just an ordinary file now.
	 */
	if (dentry->d_count == 1) {
		dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
		goto out;  /* No need to silly rename. */
	}

#ifdef NFS_PARANOIA
if (!dentry->d_inode)
printk("NFS: silly-renaming %s/%s, negative dentry??\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
	/*
	 * We don't allow a dentry to be silly-renamed twice.
	 */
	error = -EBUSY;
	if (dentry->d_flags & DCACHE_NFSFS_RENAMED)
		goto out;

	sprintf(silly, ".nfs%*.*lx",
		i_inosize, i_inosize, dentry->d_inode->i_ino);

	sdentry = NULL;
	do {
		char *suffix = silly + slen - countersize;

		dput(sdentry);
		sillycounter++;
		sprintf(suffix, "%*.*x", countersize, countersize, sillycounter);

		dfprintk(VFS, "trying to rename %s to %s\n",
			 dentry->d_name.name, silly);
		
		sdentry = nfs_silly_lookup(dentry->d_parent, silly, slen);
		/*
		 * N.B. Better to return EBUSY here ... it could be
		 * dangerous to delete the file while it's in use.
		 */
		if (IS_ERR(sdentry))
			goto out;
	} while(sdentry->d_inode != NULL); /* need negative lookup */

	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_rename(NFS_SERVER(dir),
				NFS_FH(dentry->d_parent), dentry->d_name.name,
				NFS_FH(dentry->d_parent), silly);
	if (!error) {
		nfs_renew_times(dentry);
		d_move(dentry, sdentry);
		dentry->d_flags |= DCACHE_NFSFS_RENAMED;
 		/* If we return 0 we don't unlink */
	}
	dput(sdentry);
out:
	return error;
}

/*
 * Remove a file after making sure there are no pending writes,
 * and after checking that the file has only one user. 
 *
 * We update inode->i_nlink and free the inode prior to the operation
 * to avoid possible races if the server reuses the inode.
 */
static int nfs_safe_remove(struct dentry *dentry)
{
	struct inode *dir = dentry->d_parent->d_inode;
	struct inode *inode = dentry->d_inode;
	int error, rehash = 0;
		
	dfprintk(VFS, "NFS: safe_remove(%s/%s, %ld)\n",
		dentry->d_parent->d_name.name, dentry->d_name.name,
		inode->i_ino);

	/* N.B. not needed now that d_delete is done in advance? */
	error = -EBUSY;
	if (!inode) {
#ifdef NFS_PARANOIA
printk("nfs_safe_remove: %s/%s already negative??\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
	}

	if (dentry->d_count > 1) {
#ifdef NFS_PARANOIA
printk("nfs_safe_remove: %s/%s busy, d_count=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name, dentry->d_count);
#endif
		goto out;
	}
	/*
	 * Unhash the dentry while we remove the file ...
	 */
	if (!list_empty(&dentry->d_hash)) {
		d_drop(dentry);
		rehash = 1;
	}
	/*
	 * Update i_nlink and free the inode before unlinking.
	 */
	if (inode) {
		if (inode->i_nlink)
			inode->i_nlink --;
		d_delete(dentry);
	}
	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_remove(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
				dentry->d_name.name);
	/*
	 * Rehash the negative dentry if the operation succeeded.
	 */
	if (!error && rehash)
		d_add(dentry, NULL);
out:
	return error;
}

/*  We do silly rename. In case sillyrename() returns -EBUSY, the inode
 *  belongs to an active ".nfs..." file and we return -EBUSY.
 *
 *  If sillyrename() returns 0, we do nothing, otherwise we unlink.
 */
static int nfs_unlink(struct inode *dir, struct dentry *dentry)
{
	int error;

	dfprintk(VFS, "NFS: unlink(%x/%ld, %s)\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name);

	error = nfs_sillyrename(dir, dentry);
	if (error && error != -EBUSY) {
		error = nfs_safe_remove(dentry);
		if (!error) {
			nfs_renew_times(dentry);
		}
	}
	return error;
}

static int
nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
	struct iattr attr;
	int error;

	dfprintk(VFS, "NFS: symlink(%x/%ld, %s, %s)\n",
		dir->i_dev, dir->i_ino, dentry->d_name.name, symname);

	error = -ENAMETOOLONG;
	if (strlen(symname) > NFS_MAXPATHLEN)
		goto out;

#ifdef NFS_PARANOIA
if (dentry->d_inode)
printk("nfs_proc_symlink: %s/%s not negative!\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
	/*
	 * Fill in the sattr for the call.
 	 * Note: SunOS 4.1.2 crashes if the mode isn't initialized!
	 */
	attr.ia_valid = ATTR_MODE;
	attr.ia_mode = S_IFLNK | S_IRWXUGO;

	/*
	 * Drop the dentry in advance to force a new lookup.
	 * Since nfs_proc_symlink doesn't return a fattr, we
	 * can't instantiate the new inode.
	 */
	d_drop(dentry);
	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_symlink(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
				dentry->d_name.name, symname, &attr);
	if (!error) {
		nfs_renew_times(dentry->d_parent);
	} else if (error == -EEXIST) {
		printk("nfs_proc_symlink: %s/%s already exists??\n",
			dentry->d_parent->d_name.name, dentry->d_name.name);
	}

out:
	return error;
}

static int 
nfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
{
	struct inode *inode = old_dentry->d_inode;
	int error;

	dfprintk(VFS, "NFS: link(%s/%s -> %s/%s)\n",
		old_dentry->d_parent->d_name.name, old_dentry->d_name.name,
		dentry->d_parent->d_name.name, dentry->d_name.name);

	/*
	 * Drop the dentry in advance to force a new lookup.
	 * Since nfs_proc_link doesn't return a file handle,
	 * we can't use the existing dentry.
	 */
	d_drop(dentry);
	invalidate_inode_pages(dir);
	nfs_flush_dircache(dir);
	error = nfs_proc_link(NFS_DSERVER(old_dentry), NFS_FH(old_dentry),
				NFS_FH(dentry->d_parent), dentry->d_name.name);
	if (!error) {
 		/*
		 * Update the link count immediately, as some apps
		 * (e.g. pine) test this after making a link.
		 */
		inode->i_nlink++;
	}
	return error;
}

/*
 * RENAME
 * FIXME: Some nfsds, like the Linux user space nfsd, may generate a
 * different file handle for the same inode after a rename (e.g. when
 * moving to a different directory). A fail-safe method to do so would
 * be to look up old_dir/old_name, create a link to new_dir/new_name and
 * rename the old file using the sillyrename stuff. This way, the original
 * file in old_dir will go away when the last process iput()s the inode.
 *
 * FIXED.
 * 
 * It actually works quite well. One needs to have the possibility for
 * at least one ".nfs..." file in each directory the file ever gets
 * moved or linked to which happens automagically with the new
 * implementation that only depends on the dcache stuff instead of
 * using the inode layer
 *
 * Unfortunately, things are a little more complicated than indicated
 * above. For a cross-directory move, we want to make sure we can get
 * rid of the old inode after the operation.  This means there must be
 * no pending writes (if it's a file), and the use count must be 1.
 * If these conditions are met, we can drop the dentries before doing
 * the rename.
 */
static int nfs_rename(struct inode *old_dir, struct dentry *old_dentry,
		      struct inode *new_dir, struct dentry *new_dentry)
{
	struct inode *old_inode = old_dentry->d_inode;
	struct inode *new_inode = new_dentry->d_inode;
	struct dentry *dentry = NULL;
	int error, rehash = 0;

	dfprintk(VFS, "NFS: rename(%s/%s -> %s/%s, ct=%d)\n",
		 old_dentry->d_parent->d_name.name, old_dentry->d_name.name,
		 new_dentry->d_parent->d_name.name, new_dentry->d_name.name,
		 new_dentry->d_count);

	/*
	 * First check whether the target is busy ... we can't
	 * safely do _any_ rename if the target is in use.
	 *
	 * For files, make a copy of the dentry and then do a 
	 * silly-rename. If the silly-rename succeeds, the
	 * copied dentry is hashed and becomes the new target.
	 *
	 * With directories check is done in VFS.
	 */
	error = -EBUSY;
	if (new_dentry->d_count > 1 && new_inode) {
		int err;
		/* copy the target dentry's name */
		dentry = d_alloc(new_dentry->d_parent,
				 &new_dentry->d_name);
		if (!dentry)
			goto out;

		/* silly-rename the existing target ... */
		err = nfs_sillyrename(new_dir, new_dentry);
		if (!err) {
			new_dentry = dentry;
			new_inode = NULL;
			/* hash the replacement target */
			d_add(new_dentry, NULL);
		}

		/* dentry still busy? */
		if (new_dentry->d_count > 1) {
#ifdef NFS_PARANOIA
			printk("nfs_rename: target %s/%s busy, d_count=%d\n",
			       new_dentry->d_parent->d_name.name,
			       new_dentry->d_name.name,
			       new_dentry->d_count);
#endif
			goto out;
		}
	}

	/*
	 * ... prune child dentries and writebacks if needed.
	 */
	if (old_dentry->d_count > 1) {
		nfs_wb_all(old_inode);
		shrink_dcache_parent(old_dentry);
	}

	if (new_dentry->d_count > 1 && new_inode) {
#ifdef NFS_PARANOIA
		printk("nfs_rename: new dentry %s/%s busy, d_count=%d\n",
		       new_dentry->d_parent->d_name.name,
		       new_dentry->d_name.name,
		       new_dentry->d_count);
#endif
		goto out;
	}

	/*
	 * To prevent any new references to the target during the rename,
	 * we unhash the dentry and free the inode in advance.
	 */
	if (!list_empty(&new_dentry->d_hash)) {
		d_drop(new_dentry);
		rehash = 1;
	}
	if (new_inode)
		d_delete(new_dentry);

	invalidate_inode_pages(new_dir);
	nfs_flush_dircache(new_dir);
	invalidate_inode_pages(old_dir);
	nfs_flush_dircache(old_dir);
	error = nfs_proc_rename(NFS_DSERVER(old_dentry),
			NFS_FH(old_dentry->d_parent), old_dentry->d_name.name,
			NFS_FH(new_dentry->d_parent), new_dentry->d_name.name);

	NFS_CACHEINV(old_dir);
	NFS_CACHEINV(new_dir);
	/* Update the dcache if needed */
	if (rehash)
		d_add(new_dentry, NULL);
	if (!error && !S_ISDIR(old_inode->i_mode))
		d_move(old_dentry, new_dentry);

out:
	/* new dentry created? */
	if (dentry)
		dput(dentry);
	return error;
}

int nfs_init_fhcache(void)
{
	nfs_fh_cachep = kmem_cache_create("nfs_fh",
					  sizeof(struct nfs_fh),
					  0, SLAB_HWCACHE_ALIGN,
					  NULL, NULL);
	if (nfs_fh_cachep == NULL)
		return -ENOMEM;

	nfs_cookie_cachep = kmem_cache_create("nfs_dcookie",
					      sizeof(struct nfs_cookie_table),
					      0, SLAB_HWCACHE_ALIGN,
					      NULL, NULL);
	if (nfs_cookie_cachep == NULL)
		return -ENOMEM;

	return 0;
}

/*
 * Local variables:
 *  version-control: t
 *  kept-new-versions: 5
 * End:
 */