xfs_super.c 46.9 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
/*
2
 * Copyright (c) 2000-2006 Silicon Graphics, Inc.
3
 * All Rights Reserved.
Linus Torvalds's avatar
Linus Torvalds committed
4
 *
5
6
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
Linus Torvalds's avatar
Linus Torvalds committed
7
8
 * published by the Free Software Foundation.
 *
9
10
11
12
 * This program is distributed in the hope that it would 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.
Linus Torvalds's avatar
Linus Torvalds committed
13
 *
14
15
16
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
Linus Torvalds's avatar
Linus Torvalds committed
17
 */
Christoph Hellwig's avatar
Christoph Hellwig committed
18

Linus Torvalds's avatar
Linus Torvalds committed
19
#include "xfs.h"
20
#include "xfs_bit.h"
Linus Torvalds's avatar
Linus Torvalds committed
21
#include "xfs_log.h"
22
#include "xfs_inum.h"
Linus Torvalds's avatar
Linus Torvalds committed
23
24
#include "xfs_trans.h"
#include "xfs_sb.h"
25
#include "xfs_ag.h"
Linus Torvalds's avatar
Linus Torvalds committed
26
27
28
29
30
31
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
32
#include "xfs_alloc_btree.h"
Linus Torvalds's avatar
Linus Torvalds committed
33
34
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
35
#include "xfs_attr_sf.h"
Linus Torvalds's avatar
Linus Torvalds committed
36
37
#include "xfs_dinode.h"
#include "xfs_inode.h"
38
#include "xfs_btree.h"
39
#include "xfs_btree_trace.h"
40
#include "xfs_ialloc.h"
Linus Torvalds's avatar
Linus Torvalds committed
41
42
43
44
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
Christoph Hellwig's avatar
Christoph Hellwig committed
45
#include "xfs_fsops.h"
Linus Torvalds's avatar
Linus Torvalds committed
46
47
48
49
#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
50
#include "xfs_vnodeops.h"
Linus Torvalds's avatar
Linus Torvalds committed
51
#include "xfs_version.h"
52
#include "xfs_log_priv.h"
53
#include "xfs_trans_priv.h"
54
#include "xfs_filestream.h"
55
56
57
58
#include "xfs_da_btree.h"
#include "xfs_extfree_item.h"
#include "xfs_mru_cache.h"
#include "xfs_inode_item.h"
59
#include "xfs_sync.h"
Christoph Hellwig's avatar
Christoph Hellwig committed
60
#include "xfs_trace.h"
Linus Torvalds's avatar
Linus Torvalds committed
61
62
63
64

#include <linux/namei.h>
#include <linux/init.h>
#include <linux/mount.h>
65
#include <linux/mempool.h>
Linus Torvalds's avatar
Linus Torvalds committed
66
#include <linux/writeback.h>
67
#include <linux/kthread.h>
68
#include <linux/freezer.h>
69
#include <linux/parser.h>
Linus Torvalds's avatar
Linus Torvalds committed
70

71
static const struct super_operations xfs_super_operations;
72
static kmem_zone_t *xfs_ioend_zone;
73
mempool_t *xfs_ioend_pool;
Linus Torvalds's avatar
Linus Torvalds committed
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
#define MNTOPT_LOGBUFS	"logbufs"	/* number of XFS log buffers */
#define MNTOPT_LOGBSIZE	"logbsize"	/* size of XFS log buffers */
#define MNTOPT_LOGDEV	"logdev"	/* log device */
#define MNTOPT_RTDEV	"rtdev"		/* realtime I/O device */
#define MNTOPT_BIOSIZE	"biosize"	/* log2 of preferred buffered io size */
#define MNTOPT_WSYNC	"wsync"		/* safe-mode nfs compatible mount */
#define MNTOPT_NOALIGN	"noalign"	/* turn off stripe alignment */
#define MNTOPT_SWALLOC	"swalloc"	/* turn on stripe width allocation */
#define MNTOPT_SUNIT	"sunit"		/* data volume stripe unit */
#define MNTOPT_SWIDTH	"swidth"	/* data volume stripe width */
#define MNTOPT_NOUUID	"nouuid"	/* ignore filesystem UUID */
#define MNTOPT_MTPT	"mtpt"		/* filesystem mount point */
#define MNTOPT_GRPID	"grpid"		/* group-ID from parent directory */
#define MNTOPT_NOGRPID	"nogrpid"	/* group-ID from current process */
#define MNTOPT_BSDGROUPS    "bsdgroups"    /* group-ID from parent directory */
#define MNTOPT_SYSVGROUPS   "sysvgroups"   /* group-ID from current process */
#define MNTOPT_ALLOCSIZE    "allocsize"    /* preferred allocation size */
#define MNTOPT_NORECOVERY   "norecovery"   /* don't run XFS recovery */
#define MNTOPT_BARRIER	"barrier"	/* use writer barriers for log write and
					 * unwritten extent conversion */
#define MNTOPT_NOBARRIER "nobarrier"	/* .. disable */
#define MNTOPT_OSYNCISOSYNC "osyncisosync" /* o_sync is REALLY o_sync */
#define MNTOPT_64BITINODE   "inode64"	/* inodes can be allocated anywhere */
#define MNTOPT_IKEEP	"ikeep"		/* do not free empty inode clusters */
#define MNTOPT_NOIKEEP	"noikeep"	/* free empty inode clusters */
#define MNTOPT_LARGEIO	   "largeio"	/* report large I/O sizes in stat() */
#define MNTOPT_NOLARGEIO   "nolargeio"	/* do not report large I/O sizes
					 * in stat(). */
#define MNTOPT_ATTR2	"attr2"		/* do use attr2 attribute format */
#define MNTOPT_NOATTR2	"noattr2"	/* do not use attr2 attribute format */
#define MNTOPT_FILESTREAM  "filestreams" /* use filestreams allocator */
#define MNTOPT_QUOTA	"quota"		/* disk quotas (user) */
#define MNTOPT_NOQUOTA	"noquota"	/* no quotas */
#define MNTOPT_USRQUOTA	"usrquota"	/* user quota enabled */
#define MNTOPT_GRPQUOTA	"grpquota"	/* group quota enabled */
#define MNTOPT_PRJQUOTA	"prjquota"	/* project quota enabled */
#define MNTOPT_UQUOTA	"uquota"	/* user quota (IRIX variant) */
#define MNTOPT_GQUOTA	"gquota"	/* group quota (IRIX variant) */
#define MNTOPT_PQUOTA	"pquota"	/* project quota (IRIX variant) */
#define MNTOPT_UQUOTANOENF "uqnoenforce"/* user quota limit enforcement */
#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
#define MNTOPT_QUOTANOENF  "qnoenforce"	/* same as uqnoenforce */
#define MNTOPT_DMAPI	"dmapi"		/* DMI enabled (DMAPI / XDSM) */
#define MNTOPT_XDSM	"xdsm"		/* DMI enabled (DMAPI / XDSM) */
#define MNTOPT_DMI	"dmi"		/* DMI enabled (DMAPI / XDSM) */

122
123
124
125
126
127
128
129
130
131
/*
 * Table driven mount option parser.
 *
 * Currently only used for remount, but it will be used for mount
 * in the future, too.
 */
enum {
	Opt_barrier, Opt_nobarrier, Opt_err
};

132
static const match_table_t tokens = {
133
134
135
136
137
138
	{Opt_barrier, "barrier"},
	{Opt_nobarrier, "nobarrier"},
	{Opt_err, NULL}
};


139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
STATIC unsigned long
suffix_strtoul(char *s, char **endp, unsigned int base)
{
	int	last, shift_left_factor = 0;
	char	*value = s;

	last = strlen(value) - 1;
	if (value[last] == 'K' || value[last] == 'k') {
		shift_left_factor = 10;
		value[last] = '\0';
	}
	if (value[last] == 'M' || value[last] == 'm') {
		shift_left_factor = 20;
		value[last] = '\0';
	}
	if (value[last] == 'G' || value[last] == 'g') {
		shift_left_factor = 30;
		value[last] = '\0';
	}

	return simple_strtoul((const char *)s, endp, base) << shift_left_factor;
}

162
163
164
165
166
167
168
/*
 * This function fills in xfs_mount_t fields based on mount args.
 * Note: the superblock has _not_ yet been read in.
 *
 * Note that this function leaks the various device name allocations on
 * failure.  The caller takes care of them.
 */
169
170
171
172
STATIC int
xfs_parseargs(
	struct xfs_mount	*mp,
	char			*options,
173
	char			**mtpt)
174
{
175
	struct super_block	*sb = mp->m_super;
176
	char			*this_char, *value, *eov;
177
178
179
	int			dsunit = 0;
	int			dswidth = 0;
	int			iosize = 0;
180
	int			dmapi_implies_ikeep = 1;
181
	__uint8_t		iosizelog = 0;
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199

	/*
	 * Copy binary VFS mount flags we are interested in.
	 */
	if (sb->s_flags & MS_RDONLY)
		mp->m_flags |= XFS_MOUNT_RDONLY;
	if (sb->s_flags & MS_DIRSYNC)
		mp->m_flags |= XFS_MOUNT_DIRSYNC;
	if (sb->s_flags & MS_SYNCHRONOUS)
		mp->m_flags |= XFS_MOUNT_WSYNC;

	/*
	 * Set some default flags that could be cleared by the mount option
	 * parsing.
	 */
	mp->m_flags |= XFS_MOUNT_BARRIER;
	mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
	mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
200

201
202
203
204
205
	/*
	 * These can be overridden by the mount option parsing.
	 */
	mp->m_logbufs = -1;
	mp->m_logbsize = -1;
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222

	if (!options)
		goto done;

	while ((this_char = strsep(&options, ",")) != NULL) {
		if (!*this_char)
			continue;
		if ((value = strchr(this_char, '=')) != NULL)
			*value++ = 0;

		if (!strcmp(this_char, MNTOPT_LOGBUFS)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
223
			mp->m_logbufs = simple_strtoul(value, &eov, 10);
224
225
226
227
228
229
230
		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
231
			mp->m_logbsize = suffix_strtoul(value, &eov, 10);
232
233
234
235
236
237
238
		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
239
240
241
			mp->m_logname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_logname)
				return ENOMEM;
242
243
244
245
246
247
248
		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
249
250
251
			*mtpt = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!*mtpt)
				return ENOMEM;
252
253
254
255
256
257
258
		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
259
260
261
			mp->m_rtname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_rtname)
				return ENOMEM;
262
263
264
265
266
267
268
269
		} else if (!strcmp(this_char, MNTOPT_BIOSIZE)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			iosize = simple_strtoul(value, &eov, 10);
Christoph Hellwig's avatar
Christoph Hellwig committed
270
			iosizelog = ffs(iosize) - 1;
271
272
273
274
275
276
277
278
		} else if (!strcmp(this_char, MNTOPT_ALLOCSIZE)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			iosize = suffix_strtoul(value, &eov, 10);
279
			iosizelog = ffs(iosize) - 1;
280
281
282
283
284
285
286
		} else if (!strcmp(this_char, MNTOPT_GRPID) ||
			   !strcmp(this_char, MNTOPT_BSDGROUPS)) {
			mp->m_flags |= XFS_MOUNT_GRPID;
		} else if (!strcmp(this_char, MNTOPT_NOGRPID) ||
			   !strcmp(this_char, MNTOPT_SYSVGROUPS)) {
			mp->m_flags &= ~XFS_MOUNT_GRPID;
		} else if (!strcmp(this_char, MNTOPT_WSYNC)) {
287
			mp->m_flags |= XFS_MOUNT_WSYNC;
288
		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
289
			mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
290
		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
291
			mp->m_flags |= XFS_MOUNT_NORECOVERY;
292
		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
293
			mp->m_flags |= XFS_MOUNT_NOALIGN;
294
		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
295
			mp->m_flags |= XFS_MOUNT_SWALLOC;
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
		} else if (!strcmp(this_char, MNTOPT_SUNIT)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			dsunit = simple_strtoul(value, &eov, 10);
		} else if (!strcmp(this_char, MNTOPT_SWIDTH)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			dswidth = simple_strtoul(value, &eov, 10);
		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
313
			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
314
315
316
317
318
319
320
#if !XFS_BIG_INUMS
			cmn_err(CE_WARN,
				"XFS: %s option not allowed on this system",
				this_char);
			return EINVAL;
#endif
		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
321
			mp->m_flags |= XFS_MOUNT_NOUUID;
322
		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
323
			mp->m_flags |= XFS_MOUNT_BARRIER;
324
		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
325
			mp->m_flags &= ~XFS_MOUNT_BARRIER;
326
		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
327
			mp->m_flags |= XFS_MOUNT_IKEEP;
328
		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
329
			dmapi_implies_ikeep = 0;
330
			mp->m_flags &= ~XFS_MOUNT_IKEEP;
331
		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
332
			mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
333
		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
334
			mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
335
		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
336
			mp->m_flags |= XFS_MOUNT_ATTR2;
337
		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
338
339
			mp->m_flags &= ~XFS_MOUNT_ATTR2;
			mp->m_flags |= XFS_MOUNT_NOATTR2;
340
		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
341
			mp->m_flags |= XFS_MOUNT_FILESTREAMS;
342
		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
343
344
			mp->m_qflags &= ~(XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
					  XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
345
					  XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
346
					  XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD);
347
348
349
		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
			   !strcmp(this_char, MNTOPT_UQUOTA) ||
			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
350
351
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
					 XFS_UQUOTA_ENFD);
352
353
		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
354
355
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_UQUOTA_ENFD;
356
357
		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
358
359
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
360
		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
361
362
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
363
364
		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
365
366
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
367
		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
368
369
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
370
		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
371
			mp->m_flags |= XFS_MOUNT_DMAPI;
372
		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
373
			mp->m_flags |= XFS_MOUNT_DMAPI;
374
		} else if (!strcmp(this_char, MNTOPT_DMI)) {
375
			mp->m_flags |= XFS_MOUNT_DMAPI;
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
		} else if (!strcmp(this_char, "ihashsize")) {
			cmn_err(CE_WARN,
	"XFS: ihashsize no longer used, option is deprecated.");
		} else if (!strcmp(this_char, "osyncisdsync")) {
			/* no-op, this is now the default */
			cmn_err(CE_WARN,
	"XFS: osyncisdsync is now the default, option is deprecated.");
		} else if (!strcmp(this_char, "irixsgid")) {
			cmn_err(CE_WARN,
	"XFS: irixsgid is now a sysctl(2) variable, option is deprecated.");
		} else {
			cmn_err(CE_WARN,
				"XFS: unknown mount option [%s].", this_char);
			return EINVAL;
		}
	}

393
394
395
396
397
398
399
	/*
	 * no recovery flag requires a read-only mount
	 */
	if ((mp->m_flags & XFS_MOUNT_NORECOVERY) &&
	    !(mp->m_flags & XFS_MOUNT_RDONLY)) {
		cmn_err(CE_WARN, "XFS: no-recovery mounts must be read-only.");
		return EINVAL;
400
401
	}

402
	if ((mp->m_flags & XFS_MOUNT_NOALIGN) && (dsunit || dswidth)) {
403
404
405
406
407
		cmn_err(CE_WARN,
	"XFS: sunit and swidth options incompatible with the noalign option");
		return EINVAL;
	}

Christoph Hellwig's avatar
Christoph Hellwig committed
408
409
410
411
412
413
414
415
#ifndef CONFIG_XFS_QUOTA
	if (XFS_IS_QUOTA_RUNNING(mp)) {
		cmn_err(CE_WARN,
			"XFS: quota support not available in this kernel.");
		return EINVAL;
	}
#endif

416
417
	if ((mp->m_qflags & (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE)) &&
	    (mp->m_qflags & (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE))) {
418
419
420
421
422
		cmn_err(CE_WARN,
			"XFS: cannot mount with both project and group quota");
		return EINVAL;
	}

423
	if ((mp->m_flags & XFS_MOUNT_DMAPI) && (!*mtpt || *mtpt[0] == '\0')) {
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
		printk("XFS: %s option needs the mount point option as well\n",
			MNTOPT_DMAPI);
		return EINVAL;
	}

	if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
		cmn_err(CE_WARN,
			"XFS: sunit and swidth must be specified together");
		return EINVAL;
	}

	if (dsunit && (dswidth % dsunit != 0)) {
		cmn_err(CE_WARN,
	"XFS: stripe width (%d) must be a multiple of the stripe unit (%d)",
			dswidth, dsunit);
		return EINVAL;
	}

	/*
	 * Applications using DMI filesystems often expect the
	 * inode generation number to be monotonically increasing.
	 * If we delete inode chunks we break this assumption, so
	 * keep unused inode chunks on disk for DMI filesystems
	 * until we come up with a better solution.
	 * Note that if "ikeep" or "noikeep" mount options are
	 * supplied, then they are honored.
	 */
451
452
	if ((mp->m_flags & XFS_MOUNT_DMAPI) && dmapi_implies_ikeep)
		mp->m_flags |= XFS_MOUNT_IKEEP;
453

454
455
456
457
458
459
460
461
done:
	if (!(mp->m_flags & XFS_MOUNT_NOALIGN)) {
		/*
		 * At this point the superblock has not been read
		 * in, therefore we do not know the block size.
		 * Before the mount call ends we will convert
		 * these to FSBs.
		 */
462
		if (dsunit) {
463
464
			mp->m_dalign = dsunit;
			mp->m_flags |= XFS_MOUNT_RETERR;
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

		if (dswidth)
			mp->m_swidth = dswidth;
	}

	if (mp->m_logbufs != -1 &&
	    mp->m_logbufs != 0 &&
	    (mp->m_logbufs < XLOG_MIN_ICLOGS ||
	     mp->m_logbufs > XLOG_MAX_ICLOGS)) {
		cmn_err(CE_WARN,
			"XFS: invalid logbufs value: %d [not %d-%d]",
			mp->m_logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
		return XFS_ERROR(EINVAL);
	}
	if (mp->m_logbsize != -1 &&
	    mp->m_logbsize !=  0 &&
	    (mp->m_logbsize < XLOG_MIN_RECORD_BSIZE ||
	     mp->m_logbsize > XLOG_MAX_RECORD_BSIZE ||
	     !is_power_of_2(mp->m_logbsize))) {
		cmn_err(CE_WARN,
	"XFS: invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
			mp->m_logbsize);
		return XFS_ERROR(EINVAL);
	}

	mp->m_fsname = kstrndup(sb->s_id, MAXNAMELEN, GFP_KERNEL);
	if (!mp->m_fsname)
		return ENOMEM;
	mp->m_fsname_len = strlen(mp->m_fsname) + 1;

	if (iosizelog) {
		if (iosizelog > XFS_MAX_IO_LOG ||
		    iosizelog < XFS_MIN_IO_LOG) {
			cmn_err(CE_WARN,
		"XFS: invalid log iosize: %d [not %d-%d]",
				iosizelog, XFS_MIN_IO_LOG,
				XFS_MAX_IO_LOG);
			return XFS_ERROR(EINVAL);
		}

		mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
		mp->m_readio_log = iosizelog;
		mp->m_writeio_log = iosizelog;
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
	}

	return 0;
}

struct proc_xfs_info {
	int	flag;
	char	*str;
};

STATIC int
xfs_showargs(
	struct xfs_mount	*mp,
	struct seq_file		*m)
{
	static struct proc_xfs_info xfs_info_set[] = {
		/* the few simple ones we can get from the mount struct */
526
		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
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
		{ XFS_MOUNT_WSYNC,		"," MNTOPT_WSYNC },
		{ XFS_MOUNT_NOALIGN,		"," MNTOPT_NOALIGN },
		{ XFS_MOUNT_SWALLOC,		"," MNTOPT_SWALLOC },
		{ XFS_MOUNT_NOUUID,		"," MNTOPT_NOUUID },
		{ XFS_MOUNT_NORECOVERY,		"," MNTOPT_NORECOVERY },
		{ XFS_MOUNT_OSYNCISOSYNC,	"," MNTOPT_OSYNCISOSYNC },
		{ XFS_MOUNT_ATTR2,		"," MNTOPT_ATTR2 },
		{ XFS_MOUNT_FILESTREAMS,	"," MNTOPT_FILESTREAM },
		{ XFS_MOUNT_DMAPI,		"," MNTOPT_DMAPI },
		{ XFS_MOUNT_GRPID,		"," MNTOPT_GRPID },
		{ 0, NULL }
	};
	static struct proc_xfs_info xfs_info_unset[] = {
		/* the few simple ones we can get from the mount struct */
		{ XFS_MOUNT_COMPAT_IOSIZE,	"," MNTOPT_LARGEIO },
		{ XFS_MOUNT_BARRIER,		"," MNTOPT_NOBARRIER },
		{ XFS_MOUNT_SMALL_INUMS,	"," MNTOPT_64BITINODE },
		{ 0, NULL }
	};
	struct proc_xfs_info	*xfs_infop;

	for (xfs_infop = xfs_info_set; xfs_infop->flag; xfs_infop++) {
		if (mp->m_flags & xfs_infop->flag)
			seq_puts(m, xfs_infop->str);
	}
	for (xfs_infop = xfs_info_unset; xfs_infop->flag; xfs_infop++) {
		if (!(mp->m_flags & xfs_infop->flag))
			seq_puts(m, xfs_infop->str);
	}

	if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE)
		seq_printf(m, "," MNTOPT_ALLOCSIZE "=%dk",
				(int)(1 << mp->m_writeio_log) >> 10);

	if (mp->m_logbufs > 0)
		seq_printf(m, "," MNTOPT_LOGBUFS "=%d", mp->m_logbufs);
	if (mp->m_logbsize > 0)
		seq_printf(m, "," MNTOPT_LOGBSIZE "=%dk", mp->m_logbsize >> 10);

	if (mp->m_logname)
		seq_printf(m, "," MNTOPT_LOGDEV "=%s", mp->m_logname);
	if (mp->m_rtname)
		seq_printf(m, "," MNTOPT_RTDEV "=%s", mp->m_rtname);

	if (mp->m_dalign > 0)
		seq_printf(m, "," MNTOPT_SUNIT "=%d",
				(int)XFS_FSB_TO_BB(mp, mp->m_dalign));
	if (mp->m_swidth > 0)
		seq_printf(m, "," MNTOPT_SWIDTH "=%d",
				(int)XFS_FSB_TO_BB(mp, mp->m_swidth));

	if (mp->m_qflags & (XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD))
		seq_puts(m, "," MNTOPT_USRQUOTA);
	else if (mp->m_qflags & XFS_UQUOTA_ACCT)
		seq_puts(m, "," MNTOPT_UQUOTANOENF);

583
584
585
586
587
588
589
590
591
592
593
594
595
	/* Either project or group quotas can be active, not both */

	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
		if (mp->m_qflags & XFS_OQUOTA_ENFD)
			seq_puts(m, "," MNTOPT_PRJQUOTA);
		else
			seq_puts(m, "," MNTOPT_PQUOTANOENF);
	} else if (mp->m_qflags & XFS_GQUOTA_ACCT) {
		if (mp->m_qflags & XFS_OQUOTA_ENFD)
			seq_puts(m, "," MNTOPT_GRPQUOTA);
		else
			seq_puts(m, "," MNTOPT_GQUOTANOENF);
	}
596
597
598
599
600
601

	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
		seq_puts(m, "," MNTOPT_NOQUOTA);

	return 0;
}
Linus Torvalds's avatar
Linus Torvalds committed
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
__uint64_t
xfs_max_file_offset(
	unsigned int		blockshift)
{
	unsigned int		pagefactor = 1;
	unsigned int		bitshift = BITS_PER_LONG - 1;

	/* Figure out maximum filesize, on Linux this can depend on
	 * the filesystem blocksize (on 32 bit platforms).
	 * __block_prepare_write does this in an [unsigned] long...
	 *      page->index << (PAGE_CACHE_SHIFT - bbits)
	 * So, for page sized blocks (4K on 32 bit platforms),
	 * this wraps at around 8Tb (hence MAX_LFS_FILESIZE which is
	 *      (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
	 * but for smaller blocksizes it is less (bbits = log2 bsize).
	 * Note1: get_block_t takes a long (implicit cast from above)
	 * Note2: The Large Block Device (LBD and HAVE_SECTOR_T) patch
	 * can optionally convert the [unsigned] long from above into
	 * an [unsigned] long long.
	 */

#if BITS_PER_LONG == 32
624
# if defined(CONFIG_LBDAF)
Linus Torvalds's avatar
Linus Torvalds committed
625
626
627
628
629
630
631
632
633
634
635
	ASSERT(sizeof(sector_t) == 8);
	pagefactor = PAGE_CACHE_SIZE;
	bitshift = BITS_PER_LONG;
# else
	pagefactor = PAGE_CACHE_SIZE >> (PAGE_CACHE_SHIFT - blockshift);
# endif
#endif

	return (((__uint64_t)pagefactor) << bitshift) - 1;
}

Hannes Eder's avatar
Hannes Eder committed
636
STATIC int
Linus Torvalds's avatar
Linus Torvalds committed
637
638
639
640
641
642
643
xfs_blkdev_get(
	xfs_mount_t		*mp,
	const char		*name,
	struct block_device	**bdevp)
{
	int			error = 0;

644
	*bdevp = open_bdev_exclusive(name, FMODE_READ|FMODE_WRITE, mp);
Linus Torvalds's avatar
Linus Torvalds committed
645
646
647
648
649
650
651
652
	if (IS_ERR(*bdevp)) {
		error = PTR_ERR(*bdevp);
		printk("XFS: Invalid device [%s], error=%d\n", name, error);
	}

	return -error;
}

Hannes Eder's avatar
Hannes Eder committed
653
STATIC void
Linus Torvalds's avatar
Linus Torvalds committed
654
655
656
657
xfs_blkdev_put(
	struct block_device	*bdev)
{
	if (bdev)
658
		close_bdev_exclusive(bdev, FMODE_READ|FMODE_WRITE);
Linus Torvalds's avatar
Linus Torvalds committed
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
/*
 * Try to write out the superblock using barriers.
 */
STATIC int
xfs_barrier_test(
	xfs_mount_t	*mp)
{
	xfs_buf_t	*sbp = xfs_getsb(mp, 0);
	int		error;

	XFS_BUF_UNDONE(sbp);
	XFS_BUF_UNREAD(sbp);
	XFS_BUF_UNDELAYWRITE(sbp);
	XFS_BUF_WRITE(sbp);
	XFS_BUF_UNASYNC(sbp);
	XFS_BUF_ORDERED(sbp);

	xfsbdstrat(mp, sbp);
	error = xfs_iowait(sbp);

	/*
	 * Clear all the flags we set and possible error state in the
	 * buffer.  We only did the write to try out whether barriers
	 * worked and shouldn't leave any traces in the superblock
	 * buffer.
	 */
	XFS_BUF_DONE(sbp);
	XFS_BUF_ERROR(sbp, 0);
	XFS_BUF_UNORDERED(sbp);

	xfs_buf_relse(sbp);
	return error;
}

695
STATIC void
696
697
698
699
700
701
702
703
xfs_mountfs_check_barriers(xfs_mount_t *mp)
{
	int error;

	if (mp->m_logdev_targp != mp->m_ddev_targp) {
		xfs_fs_cmn_err(CE_NOTE, mp,
		  "Disabling barriers, not supported with external log device");
		mp->m_flags &= ~XFS_MOUNT_BARRIER;
704
		return;
705
706
	}

707
708
709
710
711
712
713
	if (xfs_readonly_buftarg(mp->m_ddev_targp)) {
		xfs_fs_cmn_err(CE_NOTE, mp,
		  "Disabling barriers, underlying device is readonly");
		mp->m_flags &= ~XFS_MOUNT_BARRIER;
		return;
	}

714
715
716
717
718
	error = xfs_barrier_test(mp);
	if (error) {
		xfs_fs_cmn_err(CE_NOTE, mp,
		  "Disabling barriers, trial barrier write failed");
		mp->m_flags &= ~XFS_MOUNT_BARRIER;
719
		return;
720
721
722
723
724
725
726
	}
}

void
xfs_blkdev_issue_flush(
	xfs_buftarg_t		*buftarg)
{
727
	blkdev_issue_flush(buftarg->bt_bdev, NULL);
728
}
Linus Torvalds's avatar
Linus Torvalds committed
729

730
731
732
733
734
STATIC void
xfs_close_devices(
	struct xfs_mount	*mp)
{
	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
735
		struct block_device *logdev = mp->m_logdev_targp->bt_bdev;
736
		xfs_free_buftarg(mp, mp->m_logdev_targp);
737
		xfs_blkdev_put(logdev);
738
739
	}
	if (mp->m_rtdev_targp) {
740
		struct block_device *rtdev = mp->m_rtdev_targp->bt_bdev;
741
		xfs_free_buftarg(mp, mp->m_rtdev_targp);
742
		xfs_blkdev_put(rtdev);
743
	}
744
	xfs_free_buftarg(mp, mp->m_ddev_targp);
745
746
747
748
749
750
751
752
753
754
755
756
757
758
}

/*
 * The file system configurations are:
 *	(1) device (partition) with data and internal log
 *	(2) logical volume with data and log subvolumes.
 *	(3) logical volume with data, log, and realtime subvolumes.
 *
 * We only have to handle opening the log and realtime volumes here if
 * they are present.  The data subvolume has already been opened by
 * get_sb_bdev() and is stored in sb->s_bdev.
 */
STATIC int
xfs_open_devices(
759
	struct xfs_mount	*mp)
760
761
762
763
764
765
766
767
{
	struct block_device	*ddev = mp->m_super->s_bdev;
	struct block_device	*logdev = NULL, *rtdev = NULL;
	int			error;

	/*
	 * Open real time and log devices - order is important.
	 */
768
769
	if (mp->m_logname) {
		error = xfs_blkdev_get(mp, mp->m_logname, &logdev);
770
771
772
773
		if (error)
			goto out;
	}

774
775
	if (mp->m_rtname) {
		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev);
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
		if (error)
			goto out_close_logdev;

		if (rtdev == ddev || rtdev == logdev) {
			cmn_err(CE_WARN,
	"XFS: Cannot mount filesystem with identical rtdev and ddev/logdev.");
			error = EINVAL;
			goto out_close_rtdev;
		}
	}

	/*
	 * Setup xfs_mount buffer target pointers
	 */
	error = ENOMEM;
	mp->m_ddev_targp = xfs_alloc_buftarg(ddev, 0);
	if (!mp->m_ddev_targp)
		goto out_close_rtdev;

	if (rtdev) {
		mp->m_rtdev_targp = xfs_alloc_buftarg(rtdev, 1);
		if (!mp->m_rtdev_targp)
			goto out_free_ddev_targ;
	}

	if (logdev && logdev != ddev) {
		mp->m_logdev_targp = xfs_alloc_buftarg(logdev, 1);
		if (!mp->m_logdev_targp)
			goto out_free_rtdev_targ;
	} else {
		mp->m_logdev_targp = mp->m_ddev_targp;
	}

	return 0;

 out_free_rtdev_targ:
	if (mp->m_rtdev_targp)
813
		xfs_free_buftarg(mp, mp->m_rtdev_targp);
814
 out_free_ddev_targ:
815
	xfs_free_buftarg(mp, mp->m_ddev_targp);
816
817
818
819
820
821
822
823
824
825
 out_close_rtdev:
	if (rtdev)
		xfs_blkdev_put(rtdev);
 out_close_logdev:
	if (logdev && logdev != ddev)
		xfs_blkdev_put(logdev);
 out:
	return error;
}

826
827
828
829
830
831
832
833
/*
 * Setup xfs_mount buffer target pointers based on superblock
 */
STATIC int
xfs_setup_devices(
	struct xfs_mount	*mp)
{
	int			error;
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
	error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_blocksize,
				    mp->m_sb.sb_sectsize);
	if (error)
		return error;

	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
		unsigned int	log_sector_size = BBSIZE;

		if (xfs_sb_version_hassector(&mp->m_sb))
			log_sector_size = mp->m_sb.sb_logsectsize;
		error = xfs_setsize_buftarg(mp->m_logdev_targp,
					    mp->m_sb.sb_blocksize,
					    log_sector_size);
		if (error)
			return error;
	}
	if (mp->m_rtdev_targp) {
		error = xfs_setsize_buftarg(mp->m_rtdev_targp,
					    mp->m_sb.sb_blocksize,
					    mp->m_sb.sb_sectsize);
		if (error)
			return error;
	}

	return 0;
}
861

862
863
864
865
866
/*
 * XFS AIL push thread support
 */
void
xfsaild_wakeup(
867
	struct xfs_ail		*ailp,
868
869
	xfs_lsn_t		threshold_lsn)
{
870
871
	ailp->xa_target = threshold_lsn;
	wake_up_process(ailp->xa_task);
872
873
}

Hannes Eder's avatar
Hannes Eder committed
874
STATIC int
875
876
877
xfsaild(
	void	*data)
{
878
	struct xfs_ail	*ailp = data;
879
880
881
882
883
884
885
886
887
888
889
	xfs_lsn_t	last_pushed_lsn = 0;
	long		tout = 0;

	while (!kthread_should_stop()) {
		if (tout)
			schedule_timeout_interruptible(msecs_to_jiffies(tout));
		tout = 1000;

		/* swsusp */
		try_to_freeze();

890
891
		ASSERT(ailp->xa_mount->m_log);
		if (XFS_FORCED_SHUTDOWN(ailp->xa_mount))
892
893
			continue;

894
		tout = xfsaild_push(ailp, &last_pushed_lsn);
895
896
897
898
899
900
901
	}

	return 0;
}	/* xfsaild */

int
xfsaild_start(
902
	struct xfs_ail	*ailp)
903
{
904
905
906
907
	ailp->xa_target = 0;
	ailp->xa_task = kthread_run(xfsaild, ailp, "xfsaild");
	if (IS_ERR(ailp->xa_task))
		return -PTR_ERR(ailp->xa_task);
908
909
910
911
912
	return 0;
}

void
xfsaild_stop(
913
	struct xfs_ail	*ailp)
914
{
915
	kthread_stop(ailp->xa_task);
916
917
918
}


919
/* Catch misguided souls that try to use this interface on XFS */
Linus Torvalds's avatar
Linus Torvalds committed
920
STATIC struct inode *
921
xfs_fs_alloc_inode(
Linus Torvalds's avatar
Linus Torvalds committed
922
923
	struct super_block	*sb)
{
924
	BUG();
925
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
926
927
}

928
/*
929
930
 * Now that the generic code is guaranteed not to be accessing
 * the linux inode, we can reclaim the inode.
931
 */
Linus Torvalds's avatar
Linus Torvalds committed
932
STATIC void
933
xfs_fs_destroy_inode(
934
	struct inode		*inode)
Linus Torvalds's avatar
Linus Torvalds committed
935
{
936
937
938
	struct xfs_inode	*ip = XFS_I(inode);

	xfs_itrace_entry(ip);
939
940

	XFS_STATS_INC(vn_reclaim);
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956

	/* bad inode, get out here ASAP */
	if (is_bad_inode(inode))
		goto out_reclaim;

	xfs_ioend_wait(ip);

	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);

	/*
	 * We should never get here with one of the reclaim flags already set.
	 */
	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIMABLE));
	ASSERT_ALWAYS(!xfs_iflags_test(ip, XFS_IRECLAIM));

	/*
957
958
959
960
961
	 * We always use background reclaim here because even if the
	 * inode is clean, it still may be under IO and hence we have
	 * to take the flush lock. The background reclaim path handles
	 * this more efficiently than we can here, so simply let background
	 * reclaim tear down all inodes.
962
963
	 */
out_reclaim:
964
	xfs_inode_set_reclaim_tag(ip);
Linus Torvalds's avatar
Linus Torvalds committed
965
966
}

967
968
969
970
971
972
973
974
/*
 * Slab object creation initialisation for the XFS inode.
 * This covers only the idempotent fields in the XFS inode;
 * all other fields need to be initialised on allocation
 * from the slab. This avoids the need to repeatedly intialise
 * fields in the xfs inode that left in the initialise state
 * when freeing the inode.
 */
975
976
STATIC void
xfs_fs_inode_init_once(
977
978
979
980
981
	void			*inode)
{
	struct xfs_inode	*ip = inode;

	memset(ip, 0, sizeof(struct xfs_inode));
982
983
984
985
986

	/* vfs inode */
	inode_init_once(VFS_I(ip));

	/* xfs inode */
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
	atomic_set(&ip->i_iocount, 0);
	atomic_set(&ip->i_pincount, 0);
	spin_lock_init(&ip->i_flags_lock);
	init_waitqueue_head(&ip->i_ipin_wait);
	/*
	 * Because we want to use a counting completion, complete
	 * the flush completion once to allow a single access to
	 * the flush completion without blocking.
	 */
	init_completion(&ip->i_flush);
	complete(&ip->i_flush);

	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
		     "xfsino", ip->i_ino);
For faster browsing, not all history is shown. View entire blame