xfs_super.c 47.3 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 * Copyright (c) 2000-2006 Silicon Graphics, Inc.
3
 * All Rights Reserved.
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
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.
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
17
 */
18

L
Linus Torvalds 已提交
19 20
#include "xfs.h"
#include "xfs_log.h"
21
#include "xfs_inum.h"
L
Linus Torvalds 已提交
22 23
#include "xfs_trans.h"
#include "xfs_sb.h"
24
#include "xfs_ag.h"
L
Linus Torvalds 已提交
25 26 27 28 29
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
30
#include "xfs_alloc_btree.h"
L
Linus Torvalds 已提交
31 32 33
#include "xfs_ialloc_btree.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
34 35
#include "xfs_btree.h"
#include "xfs_ialloc.h"
L
Linus Torvalds 已提交
36 37 38 39
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
40
#include "xfs_fsops.h"
L
Linus Torvalds 已提交
41 42 43
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
44
#include "xfs_vnodeops.h"
45
#include "xfs_log_priv.h"
46
#include "xfs_trans_priv.h"
47
#include "xfs_filestream.h"
48 49 50 51
#include "xfs_da_btree.h"
#include "xfs_extfree_item.h"
#include "xfs_mru_cache.h"
#include "xfs_inode_item.h"
52
#include "xfs_icache.h"
53
#include "xfs_trace.h"
L
Linus Torvalds 已提交
54 55 56

#include <linux/namei.h>
#include <linux/init.h>
57
#include <linux/slab.h>
L
Linus Torvalds 已提交
58
#include <linux/mount.h>
59
#include <linux/mempool.h>
L
Linus Torvalds 已提交
60
#include <linux/writeback.h>
61
#include <linux/kthread.h>
62
#include <linux/freezer.h>
63
#include <linux/parser.h>
L
Linus Torvalds 已提交
64

65
static const struct super_operations xfs_super_operations;
66
static kmem_zone_t *xfs_ioend_zone;
67
mempool_t *xfs_ioend_pool;
L
Linus Torvalds 已提交
68

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
#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_64BITINODE   "inode64"	/* inodes can be allocated anywhere */
91 92
#define MNTOPT_32BITINODE   "inode32"	/* inode allocation limited to
					 * XFS_MAXINUMBER_32 */
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
#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 */
113 114 115 116
#define MNTOPT_DELAYLOG    "delaylog"	/* Delayed logging enabled */
#define MNTOPT_NODELAYLOG  "nodelaylog"	/* Delayed logging disabled */
#define MNTOPT_DISCARD	   "discard"	/* Discard unused blocks */
#define MNTOPT_NODISCARD   "nodiscard"	/* Do not discard unused blocks */
117

118 119 120 121 122 123 124
/*
 * Table driven mount option parser.
 *
 * Currently only used for remount, but it will be used for mount
 * in the future, too.
 */
enum {
125 126 127 128 129
	Opt_barrier,
	Opt_nobarrier,
	Opt_inode64,
	Opt_inode32,
	Opt_err
130 131
};

132
static const match_table_t tokens = {
133 134
	{Opt_barrier, "barrier"},
	{Opt_nobarrier, "nobarrier"},
135
	{Opt_inode64, "inode64"},
136
	{Opt_inode32, "inode32"},
137 138 139 140
	{Opt_err, NULL}
};


141
STATIC unsigned long
142
suffix_kstrtoint(char *s, unsigned int base, int *res)
143
{
144
	int	last, shift_left_factor = 0, _res;
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160
	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';
	}

161 162 163 164
	if (kstrtoint(s, base, &_res))
		return -EINVAL;
	*res = _res << shift_left_factor;
	return 0;
165 166
}

167 168 169 170 171 172 173
/*
 * 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.
 */
174 175 176
STATIC int
xfs_parseargs(
	struct xfs_mount	*mp,
C
Christoph Hellwig 已提交
177
	char			*options)
178
{
179
	struct super_block	*sb = mp->m_super;
180
	char			*this_char, *value;
181 182 183
	int			dsunit = 0;
	int			dswidth = 0;
	int			iosize = 0;
184
	__uint8_t		iosizelog = 0;
185

186 187 188 189 190 191 192 193 194
	/*
	 * set up the mount name first so all the errors will refer to the
	 * correct device.
	 */
	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;

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
	/*
	 * 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;
211
#if !XFS_BIG_INUMS
212
	mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
213
#endif
214

215 216 217 218 219
	/*
	 * These can be overridden by the mount option parsing.
	 */
	mp->m_logbufs = -1;
	mp->m_logbsize = -1;
220 221 222 223 224 225 226 227 228 229 230 231

	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) {
232
				xfs_warn(mp, "%s option requires an argument",
233 234 235
					this_char);
				return EINVAL;
			}
236 237
			if (kstrtoint(value, 10, &mp->m_logbufs))
				return EINVAL;
238 239
		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
			if (!value || !*value) {
240
				xfs_warn(mp, "%s option requires an argument",
241 242 243
					this_char);
				return EINVAL;
			}
244 245
			if (suffix_kstrtoint(value, 10, &mp->m_logbsize))
				return EINVAL;
246 247
		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
			if (!value || !*value) {
248
				xfs_warn(mp, "%s option requires an argument",
249 250 251
					this_char);
				return EINVAL;
			}
252 253 254
			mp->m_logname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_logname)
				return ENOMEM;
255
		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
256
			xfs_warn(mp, "%s option not allowed on this system",
C
Christoph Hellwig 已提交
257 258
				this_char);
			return EINVAL;
259 260
		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
			if (!value || !*value) {
261
				xfs_warn(mp, "%s option requires an argument",
262 263 264
					this_char);
				return EINVAL;
			}
265 266 267
			mp->m_rtname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_rtname)
				return ENOMEM;
268 269
		} else if (!strcmp(this_char, MNTOPT_BIOSIZE)) {
			if (!value || !*value) {
270
				xfs_warn(mp, "%s option requires an argument",
271 272 273
					this_char);
				return EINVAL;
			}
274 275
			if (kstrtoint(value, 10, &iosize))
				return EINVAL;
276
			iosizelog = ffs(iosize) - 1;
277 278
		} else if (!strcmp(this_char, MNTOPT_ALLOCSIZE)) {
			if (!value || !*value) {
279
				xfs_warn(mp, "%s option requires an argument",
280 281 282
					this_char);
				return EINVAL;
			}
283 284
			if (suffix_kstrtoint(value, 10, &iosize))
				return EINVAL;
285
			iosizelog = ffs(iosize) - 1;
286 287 288 289 290 291 292
		} 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)) {
293
			mp->m_flags |= XFS_MOUNT_WSYNC;
294
		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
295
			mp->m_flags |= XFS_MOUNT_NORECOVERY;
296
		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
297
			mp->m_flags |= XFS_MOUNT_NOALIGN;
298
		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
299
			mp->m_flags |= XFS_MOUNT_SWALLOC;
300 301
		} else if (!strcmp(this_char, MNTOPT_SUNIT)) {
			if (!value || !*value) {
302
				xfs_warn(mp, "%s option requires an argument",
303 304 305
					this_char);
				return EINVAL;
			}
306 307
			if (kstrtoint(value, 10, &dsunit))
				return EINVAL;
308 309
		} else if (!strcmp(this_char, MNTOPT_SWIDTH)) {
			if (!value || !*value) {
310
				xfs_warn(mp, "%s option requires an argument",
311 312 313
					this_char);
				return EINVAL;
			}
314 315
			if (kstrtoint(value, 10, &dswidth))
				return EINVAL;
316 317
		} else if (!strcmp(this_char, MNTOPT_32BITINODE)) {
			mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
318
		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
319
			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
320
#if !XFS_BIG_INUMS
321
			xfs_warn(mp, "%s option not allowed on this system",
322 323 324 325
				this_char);
			return EINVAL;
#endif
		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
326
			mp->m_flags |= XFS_MOUNT_NOUUID;
327
		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
328
			mp->m_flags |= XFS_MOUNT_BARRIER;
329
		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
330
			mp->m_flags &= ~XFS_MOUNT_BARRIER;
331
		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
332
			mp->m_flags |= XFS_MOUNT_IKEEP;
333
		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
334
			mp->m_flags &= ~XFS_MOUNT_IKEEP;
335
		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
336
			mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
337
		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
338
			mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
339
		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
340
			mp->m_flags |= XFS_MOUNT_ATTR2;
341
		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
342 343
			mp->m_flags &= ~XFS_MOUNT_ATTR2;
			mp->m_flags |= XFS_MOUNT_NOATTR2;
344
		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
345
			mp->m_flags |= XFS_MOUNT_FILESTREAMS;
346
		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
347 348 349
			mp->m_qflags &= ~XFS_ALL_QUOTA_ACCT;
			mp->m_qflags &= ~XFS_ALL_QUOTA_ENFD;
			mp->m_qflags &= ~XFS_ALL_QUOTA_ACTIVE;
350 351 352
		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
			   !strcmp(this_char, MNTOPT_UQUOTA) ||
			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
353 354
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
					 XFS_UQUOTA_ENFD);
355 356
		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
357 358
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_UQUOTA_ENFD;
359 360
		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
361 362
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
363
		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
364 365
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
366 367
		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
368 369
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
370
		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
371 372
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
373
		} else if (!strcmp(this_char, MNTOPT_DELAYLOG)) {
374 375
			xfs_warn(mp,
	"delaylog is the default now, option is deprecated.");
376
		} else if (!strcmp(this_char, MNTOPT_NODELAYLOG)) {
377
			xfs_warn(mp,
378
	"nodelaylog support has been removed, option is deprecated.");
379 380 381 382
		} else if (!strcmp(this_char, MNTOPT_DISCARD)) {
			mp->m_flags |= XFS_MOUNT_DISCARD;
		} else if (!strcmp(this_char, MNTOPT_NODISCARD)) {
			mp->m_flags &= ~XFS_MOUNT_DISCARD;
383
		} else if (!strcmp(this_char, "ihashsize")) {
384 385
			xfs_warn(mp,
	"ihashsize no longer used, option is deprecated.");
386
		} else if (!strcmp(this_char, "osyncisdsync")) {
387 388
			xfs_warn(mp,
	"osyncisdsync has no effect, option is deprecated.");
389
		} else if (!strcmp(this_char, "osyncisosync")) {
390 391
			xfs_warn(mp,
	"osyncisosync has no effect, option is deprecated.");
392
		} else if (!strcmp(this_char, "irixsgid")) {
393 394
			xfs_warn(mp,
	"irixsgid is now a sysctl(2) variable, option is deprecated.");
395
		} else {
396
			xfs_warn(mp, "unknown mount option [%s].", this_char);
397 398 399 400
			return EINVAL;
		}
	}

401 402 403 404 405
	/*
	 * no recovery flag requires a read-only mount
	 */
	if ((mp->m_flags & XFS_MOUNT_NORECOVERY) &&
	    !(mp->m_flags & XFS_MOUNT_RDONLY)) {
406
		xfs_warn(mp, "no-recovery mounts must be read-only.");
407
		return EINVAL;
408 409
	}

410
	if ((mp->m_flags & XFS_MOUNT_NOALIGN) && (dsunit || dswidth)) {
411 412
		xfs_warn(mp,
	"sunit and swidth options incompatible with the noalign option");
413 414 415
		return EINVAL;
	}

C
Christoph Hellwig 已提交
416 417
#ifndef CONFIG_XFS_QUOTA
	if (XFS_IS_QUOTA_RUNNING(mp)) {
418
		xfs_warn(mp, "quota support not available in this kernel.");
C
Christoph Hellwig 已提交
419 420 421 422
		return EINVAL;
	}
#endif

423 424
	if ((mp->m_qflags & (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE)) &&
	    (mp->m_qflags & (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE))) {
425
		xfs_warn(mp, "cannot mount with both project and group quota");
426 427 428 429
		return EINVAL;
	}

	if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
430
		xfs_warn(mp, "sunit and swidth must be specified together");
431 432 433 434
		return EINVAL;
	}

	if (dsunit && (dswidth % dsunit != 0)) {
435 436
		xfs_warn(mp,
	"stripe width (%d) must be a multiple of the stripe unit (%d)",
437 438 439 440
			dswidth, dsunit);
		return EINVAL;
	}

441 442 443 444 445 446 447 448
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.
		 */
449
		if (dsunit) {
450 451
			mp->m_dalign = dsunit;
			mp->m_flags |= XFS_MOUNT_RETERR;
452
		}
453 454 455 456 457 458 459 460 461

		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)) {
462
		xfs_warn(mp, "invalid logbufs value: %d [not %d-%d]",
463 464 465 466 467 468 469 470
			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))) {
471 472
		xfs_warn(mp,
			"invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
473 474 475 476 477 478 479
			mp->m_logbsize);
		return XFS_ERROR(EINVAL);
	}

	if (iosizelog) {
		if (iosizelog > XFS_MAX_IO_LOG ||
		    iosizelog < XFS_MIN_IO_LOG) {
480
			xfs_warn(mp, "invalid log iosize: %d [not %d-%d]",
481 482 483 484 485 486 487 488
				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;
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
	}

	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 */
506
		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
507 508 509 510 511 512 513 514
		{ 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_ATTR2,		"," MNTOPT_ATTR2 },
		{ XFS_MOUNT_FILESTREAMS,	"," MNTOPT_FILESTREAM },
		{ XFS_MOUNT_GRPID,		"," MNTOPT_GRPID },
515
		{ XFS_MOUNT_DISCARD,		"," MNTOPT_DISCARD },
516
		{ XFS_MOUNT_SMALL_INUMS,	"," MNTOPT_32BITINODE },
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
		{ 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);

563 564 565 566 567 568 569 570 571 572 573 574 575
	/* 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);
	}
576 577 578 579 580 581

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

	return 0;
}
L
Linus Torvalds 已提交
582 583 584 585 586 587 588 589 590
__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).
591
	 * __block_write_begin does this in an [unsigned] long...
L
Linus Torvalds 已提交
592 593 594 595 596 597 598 599 600 601 602 603
	 *      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
604
# if defined(CONFIG_LBDAF)
L
Linus Torvalds 已提交
605 606 607 608 609 610 611 612 613 614 615
	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;
}

616 617 618 619
xfs_agnumber_t
xfs_set_inode32(struct xfs_mount *mp)
{
	xfs_agnumber_t	index = 0;
620
	xfs_agnumber_t	maxagi = 0;
621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
	xfs_sb_t	*sbp = &mp->m_sb;
	xfs_agnumber_t	max_metadata;
	xfs_agino_t	agino =	XFS_OFFBNO_TO_AGINO(mp, sbp->sb_agblocks -1, 0);
	xfs_ino_t	ino = XFS_AGINO_TO_INO(mp, sbp->sb_agcount -1, agino);
	xfs_perag_t	*pag;

	/* Calculate how much should be reserved for inodes to meet
	 * the max inode percentage.
	 */
	if (mp->m_maxicount) {
		__uint64_t	icount;

		icount = sbp->sb_dblocks * sbp->sb_imax_pct;
		do_div(icount, 100);
		icount += sbp->sb_agblocks - 1;
		do_div(icount, sbp->sb_agblocks);
		max_metadata = icount;
	} else {
		max_metadata = sbp->sb_agcount;
	}

	for (index = 0; index < sbp->sb_agcount; index++) {
		ino = XFS_AGINO_TO_INO(mp, index, agino);
644

645
		if (ino > XFS_MAXINUMBER_32) {
646 647 648 649 650
			pag = xfs_perag_get(mp, index);
			pag->pagi_inodeok = 0;
			pag->pagf_metadata = 0;
			xfs_perag_put(pag);
			continue;
651 652 653 654
		}

		pag = xfs_perag_get(mp, index);
		pag->pagi_inodeok = 1;
655
		maxagi++;
656 657 658 659
		if (index < max_metadata)
			pag->pagf_metadata = 1;
		xfs_perag_put(pag);
	}
660 661 662 663
	mp->m_flags |= (XFS_MOUNT_32BITINODES |
			XFS_MOUNT_SMALL_INUMS);

	return maxagi;
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
}

xfs_agnumber_t
xfs_set_inode64(struct xfs_mount *mp)
{
	xfs_agnumber_t index = 0;

	for (index = 0; index < mp->m_sb.sb_agcount; index++) {
		struct xfs_perag	*pag;

		pag = xfs_perag_get(mp, index);
		pag->pagi_inodeok = 1;
		pag->pagf_metadata = 0;
		xfs_perag_put(pag);
	}

	/* There is no need for lock protection on m_flags,
	 * the rw_semaphore of the VFS superblock is locked
	 * during mount/umount/remount operations, so this is
	 * enough to avoid concurency on the m_flags field
	 */
	mp->m_flags &= ~(XFS_MOUNT_32BITINODES |
			 XFS_MOUNT_SMALL_INUMS);
	return index;
}

H
Hannes Eder 已提交
690
STATIC int
L
Linus Torvalds 已提交
691 692 693 694 695 696 697
xfs_blkdev_get(
	xfs_mount_t		*mp,
	const char		*name,
	struct block_device	**bdevp)
{
	int			error = 0;

698 699
	*bdevp = blkdev_get_by_path(name, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
				    mp);
L
Linus Torvalds 已提交
700 701
	if (IS_ERR(*bdevp)) {
		error = PTR_ERR(*bdevp);
702
		xfs_warn(mp, "Invalid device [%s], error=%d\n", name, error);
L
Linus Torvalds 已提交
703 704 705 706 707
	}

	return -error;
}

H
Hannes Eder 已提交
708
STATIC void
L
Linus Torvalds 已提交
709 710 711 712
xfs_blkdev_put(
	struct block_device	*bdev)
{
	if (bdev)
713
		blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
L
Linus Torvalds 已提交
714 715
}

716 717 718 719
void
xfs_blkdev_issue_flush(
	xfs_buftarg_t		*buftarg)
{
720
	blkdev_issue_flush(buftarg->bt_bdev, GFP_NOFS, NULL);
721
}
L
Linus Torvalds 已提交
722

723 724 725 726 727
STATIC void
xfs_close_devices(
	struct xfs_mount	*mp)
{
	if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp) {
728
		struct block_device *logdev = mp->m_logdev_targp->bt_bdev;
729
		xfs_free_buftarg(mp, mp->m_logdev_targp);
730
		xfs_blkdev_put(logdev);
731 732
	}
	if (mp->m_rtdev_targp) {
733
		struct block_device *rtdev = mp->m_rtdev_targp->bt_bdev;
734
		xfs_free_buftarg(mp, mp->m_rtdev_targp);
735
		xfs_blkdev_put(rtdev);
736
	}
737
	xfs_free_buftarg(mp, mp->m_ddev_targp);
738 739 740 741 742 743 744 745 746 747 748 749 750 751
}

/*
 * 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(
752
	struct xfs_mount	*mp)
753 754 755 756 757 758 759 760
{
	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.
	 */
761 762
	if (mp->m_logname) {
		error = xfs_blkdev_get(mp, mp->m_logname, &logdev);
763 764 765 766
		if (error)
			goto out;
	}

767 768
	if (mp->m_rtname) {
		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev);
769 770 771 772
		if (error)
			goto out_close_logdev;

		if (rtdev == ddev || rtdev == logdev) {
773 774
			xfs_warn(mp,
	"Cannot mount filesystem with identical rtdev and ddev/logdev.");
775 776 777 778 779 780 781 782 783
			error = EINVAL;
			goto out_close_rtdev;
		}
	}

	/*
	 * Setup xfs_mount buffer target pointers
	 */
	error = ENOMEM;
784
	mp->m_ddev_targp = xfs_alloc_buftarg(mp, ddev, 0, mp->m_fsname);
785 786 787 788
	if (!mp->m_ddev_targp)
		goto out_close_rtdev;

	if (rtdev) {
789 790
		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev, 1,
							mp->m_fsname);
791 792 793 794 795
		if (!mp->m_rtdev_targp)
			goto out_free_ddev_targ;
	}

	if (logdev && logdev != ddev) {
796 797
		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev, 1,
							mp->m_fsname);
798 799 800 801 802 803 804 805 806 807
		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)
808
		xfs_free_buftarg(mp, mp->m_rtdev_targp);
809
 out_free_ddev_targ:
810
	xfs_free_buftarg(mp, mp->m_ddev_targp);
811 812 813 814 815 816 817 818 819 820
 out_close_rtdev:
	if (rtdev)
		xfs_blkdev_put(rtdev);
 out_close_logdev:
	if (logdev && logdev != ddev)
		xfs_blkdev_put(logdev);
 out:
	return error;
}

821 822 823 824 825 826 827 828
/*
 * Setup xfs_mount buffer target pointers based on superblock
 */
STATIC int
xfs_setup_devices(
	struct xfs_mount	*mp)
{
	int			error;
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
	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;
}
856

857 858 859 860 861 862 863 864 865 866 867 868 869 870
STATIC int
xfs_init_mount_workqueues(
	struct xfs_mount	*mp)
{
	mp->m_data_workqueue = alloc_workqueue("xfs-data/%s",
			WQ_MEM_RECLAIM, 0, mp->m_fsname);
	if (!mp->m_data_workqueue)
		goto out;

	mp->m_unwritten_workqueue = alloc_workqueue("xfs-conv/%s",
			WQ_MEM_RECLAIM, 0, mp->m_fsname);
	if (!mp->m_unwritten_workqueue)
		goto out_destroy_data_iodone_queue;

871 872 873 874
	mp->m_cil_workqueue = alloc_workqueue("xfs-cil/%s",
			WQ_MEM_RECLAIM, 0, mp->m_fsname);
	if (!mp->m_cil_workqueue)
		goto out_destroy_unwritten;
875 876 877 878 879 880 881 882 883 884 885

	mp->m_reclaim_workqueue = alloc_workqueue("xfs-reclaim/%s",
			WQ_NON_REENTRANT, 0, mp->m_fsname);
	if (!mp->m_reclaim_workqueue)
		goto out_destroy_cil;

	mp->m_log_workqueue = alloc_workqueue("xfs-log/%s",
			WQ_NON_REENTRANT, 0, mp->m_fsname);
	if (!mp->m_log_workqueue)
		goto out_destroy_reclaim;

886 887 888 889 890
	mp->m_eofblocks_workqueue = alloc_workqueue("xfs-eofblocks/%s",
			WQ_NON_REENTRANT, 0, mp->m_fsname);
	if (!mp->m_eofblocks_workqueue)
		goto out_destroy_log;

891 892
	return 0;

893 894
out_destroy_log:
	destroy_workqueue(mp->m_log_workqueue);
895 896 897 898
out_destroy_reclaim:
	destroy_workqueue(mp->m_reclaim_workqueue);
out_destroy_cil:
	destroy_workqueue(mp->m_cil_workqueue);
899 900
out_destroy_unwritten:
	destroy_workqueue(mp->m_unwritten_workqueue);
901 902 903 904 905 906 907 908 909 910
out_destroy_data_iodone_queue:
	destroy_workqueue(mp->m_data_workqueue);
out:
	return -ENOMEM;
}

STATIC void
xfs_destroy_mount_workqueues(
	struct xfs_mount	*mp)
{
911
	destroy_workqueue(mp->m_eofblocks_workqueue);
912 913
	destroy_workqueue(mp->m_log_workqueue);
	destroy_workqueue(mp->m_reclaim_workqueue);
914
	destroy_workqueue(mp->m_cil_workqueue);
915 916 917 918
	destroy_workqueue(mp->m_data_workqueue);
	destroy_workqueue(mp->m_unwritten_workqueue);
}

919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936
/*
 * Flush all dirty data to disk. Must not be called while holding an XFS_ILOCK
 * or a page lock. We use sync_inodes_sb() here to ensure we block while waiting
 * for IO to complete so that we effectively throttle multiple callers to the
 * rate at which IO is completing.
 */
void
xfs_flush_inodes(
	struct xfs_mount	*mp)
{
	struct super_block	*sb = mp->m_super;

	if (down_read_trylock(&sb->s_umount)) {
		sync_inodes_sb(sb);
		up_read(&sb->s_umount);
	}
}

937
/* Catch misguided souls that try to use this interface on XFS */
L
Linus Torvalds 已提交
938
STATIC struct inode *
939
xfs_fs_alloc_inode(
L
Linus Torvalds 已提交
940 941
	struct super_block	*sb)
{
942
	BUG();
943
	return NULL;
L
Linus Torvalds 已提交
944 945
}

946
/*
947 948
 * Now that the generic code is guaranteed not to be accessing
 * the linux inode, we can reclaim the inode.
949
 */
L
Linus Torvalds 已提交
950
STATIC void
951
xfs_fs_destroy_inode(
952
	struct inode		*inode)
L
Linus Torvalds 已提交
953
{
954 955
	struct xfs_inode	*ip = XFS_I(inode);

956
	trace_xfs_destroy_inode(ip);
957 958

	XFS_STATS_INC(vn_reclaim);
959 960 961 962 963 964 965 966 967 968 969 970 971 972

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

	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));

	/*
973 974 975 976 977
	 * 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.
978 979
	 */
out_reclaim:
980
	xfs_inode_set_reclaim_tag(ip);
L
Linus Torvalds 已提交
981 982
}

983 984 985 986
/*
 * 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
987
 * from the slab. This avoids the need to repeatedly initialise
988 989 990
 * fields in the xfs inode that left in the initialise state
 * when freeing the inode.
 */
991 992
STATIC void
xfs_fs_inode_init_once(
993 994 995 996 997
	void			*inode)
{
	struct xfs_inode	*ip = inode;

	memset(ip, 0, sizeof(struct xfs_inode));
998 999 1000 1001 1002

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

	/* xfs inode */
1003 1004 1005 1006 1007 1008 1009
	atomic_set(&ip->i_pincount, 0);
	spin_lock_init(&ip->i_flags_lock);

	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
		     "xfsino", ip->i_ino);
}

L
Linus Torvalds 已提交
1010
STATIC void
1011
xfs_fs_evict_inode(
L
Linus Torvalds 已提交
1012 1013
	struct inode		*inode)
{
1014
	xfs_inode_t		*ip = XFS_I(inode);
1015

1016 1017
	ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock));

1018
	trace_xfs_evict_inode(ip);
1019

1020
	truncate_inode_pages(&inode->i_data, 0);
1021
	clear_inode(inode);
1022 1023 1024 1025 1026
	XFS_STATS_INC(vn_rele);
	XFS_STATS_INC(vn_remove);
	XFS_STATS_DEC(vn_active);

	xfs_inactive(ip);
1027
}
L
Linus Torvalds 已提交
1028

1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
/*
 * We do an unlocked check for XFS_IDONTCACHE here because we are already
 * serialised against cache hits here via the inode->i_lock and igrab() in
 * xfs_iget_cache_hit(). Hence a lookup that might clear this flag will not be
 * racing with us, and it avoids needing to grab a spinlock here for every inode
 * we drop the final reference on.
 */
STATIC int
xfs_fs_drop_inode(
	struct inode		*inode)
{
	struct xfs_inode	*ip = XFS_I(inode);

	return generic_drop_inode(inode) || (ip->i_flags & XFS_IDONTCACHE);
}

1045 1046 1047 1048 1049 1050 1051 1052 1053
STATIC void
xfs_free_fsname(
	struct xfs_mount	*mp)
{
	kfree(mp->m_fsname);
	kfree(mp->m_rtname);
	kfree(mp->m_logname);
}

L
Linus Torvalds 已提交
1054
STATIC void
1055
xfs_fs_put_super(
L
Linus Torvalds 已提交
1056 1057
	struct super_block	*sb)
{
1058
	struct xfs_mount	*mp = XFS_M(sb);
L
Linus Torvalds 已提交
1059

1060 1061 1062
	xfs_filestream_unmount(mp);
	xfs_unmountfs(mp);

1063
	xfs_freesb(mp);
1064
	xfs_icsb_destroy_counters(mp);
1065
	xfs_destroy_mount_workqueues(mp);
1066
	xfs_close_devices(mp);
1067
	xfs_free_fsname(mp);
1068
	kfree(mp);
L
Linus Torvalds 已提交
1069 1070 1071
}

STATIC int
1072
xfs_fs_sync_fs(
L
Linus Torvalds 已提交
1073 1074 1075
	struct super_block	*sb,
	int			wait)
{
1076
	struct xfs_mount	*mp = XFS_M(sb);
L
Linus Torvalds 已提交
1077

1078
	/*
1079
	 * Doing anything during the async pass would be counterproductive.
1080
	 */
1081
	if (!wait)
1082 1083
		return 0;

1084
	xfs_log_force(mp, XFS_LOG_SYNC);
1085
	if (laptop_mode) {
L
Linus Torvalds 已提交
1086 1087
		/*
		 * The disk must be active because we're syncing.
1088
		 * We schedule log work now (now that the disk is
L
Linus Torvalds 已提交
1089 1090
		 * active) instead of later (when it might not be).
		 */
1091
		flush_delayed_work(&mp->m_log->l_work);
L
Linus Torvalds 已提交
1092 1093
	}

1094
	return 0;
L
Linus Torvalds 已提交
1095 1096 1097
}

STATIC int
1098
xfs_fs_statfs(
1099
	struct dentry		*dentry,
L
Linus Torvalds 已提交
1100 1101
	struct kstatfs		*statp)
{
1102 1103
	struct xfs_mount	*mp = XFS_M(dentry->d_sb);
	xfs_sb_t		*sbp = &mp->m_sb;
C
Christoph Hellwig 已提交
1104
	struct xfs_inode	*ip = XFS_I(dentry->d_inode);
1105 1106
	__uint64_t		fakeinos, id;
	xfs_extlen_t		lsize;
1107
	__int64_t		ffree;
1108 1109 1110 1111 1112 1113 1114 1115

	statp->f_type = XFS_SB_MAGIC;
	statp->f_namelen = MAXNAMELEN - 1;

	id = huge_encode_dev(mp->m_ddev_targp->bt_dev);
	statp->f_fsid.val[0] = (u32)id;
	statp->f_fsid.val[1] = (u32)(id >> 32);

1116
	xfs_icsb_sync_counters(mp, XFS_ICSB_LAZY_COUNT);
1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127

	spin_lock(&mp->m_sb_lock);
	statp->f_bsize = sbp->sb_blocksize;
	lsize = sbp->sb_logstart ? sbp->sb_logblocks : 0;
	statp->f_blocks = sbp->sb_dblocks - lsize;
	statp->f_bfree = statp->f_bavail =
				sbp->sb_fdblocks - XFS_ALLOC_SET_ASIDE(mp);
	fakeinos = statp->f_bfree << sbp->sb_inopblog;
	statp->f_files =
	    MIN(sbp->sb_icount + fakeinos, (__uint64_t)XFS_MAXINUMBER);
	if (mp->m_maxicount)
1128 1129 1130
		statp->f_files = min_t(typeof(statp->f_files),
					statp->f_files,
					mp->m_maxicount);
1131 1132 1133 1134 1135

	/* make sure statp->f_ffree does not underflow */
	ffree = statp->f_files - (sbp->sb_icount - sbp->sb_ifree);
	statp->f_ffree = max_t(__int64_t, ffree, 0);

1136 1137
	spin_unlock(&mp->m_sb_lock);

1138
	if ((ip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
C
Christoph Hellwig 已提交
1139 1140 1141
	    ((mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))) ==
			      (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
		xfs_qm_statvfs(ip, statp);
1142
	return 0;
L
Linus Torvalds 已提交
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
STATIC void
xfs_save_resvblks(struct xfs_mount *mp)
{
	__uint64_t resblks = 0;

	mp->m_resblks_save = mp->m_resblks;
	xfs_reserve_blocks(mp, &resblks, NULL);
}

STATIC void
xfs_restore_resvblks(struct xfs_mount *mp)
{
	__uint64_t resblks;

	if (mp->m_resblks_save) {
		resblks = mp->m_resblks_save;
		mp->m_resblks_save = 0;
	} else
		resblks = xfs_default_resblks(mp);

	xfs_reserve_blocks(mp, &resblks, NULL);
}

1168 1169 1170 1171 1172
/*
 * Trigger writeback of all the dirty metadata in the file system.
 *
 * This ensures that the metadata is written to their location on disk rather
 * than just existing in transactions in the log. This means after a quiesce
1173 1174
 * there is no log replay required to write the inodes to disk - this is the
 * primary difference between a sync and a quiesce.
1175
 *
1176 1177
 * Note: xfs_log_quiesce() stops background log work - the callers must ensure
 * it is started again when appropriate.
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
 */
void
xfs_quiesce_attr(
	struct xfs_mount	*mp)
{
	int	error = 0;

	/* wait for all modifications to complete */
	while (atomic_read(&mp->m_active_trans) > 0)
		delay(100);

	/* force the log to unpin objects from the now complete transactions */
	xfs_log_force(mp, XFS_LOG_SYNC);

	/* reclaim inodes to do any IO before the freeze completes */
	xfs_reclaim_inodes(mp, 0);
	xfs_reclaim_inodes(mp, SYNC_WAIT);

	/* Push the superblock and write an unmount record */
	error = xfs_log_sbcount(mp);
	if (error)
		xfs_warn(mp, "xfs_attr_quiesce: failed to log sb changes. "
				"Frozen image may not be consistent.");
	/*
1202 1203
	 * Just warn here till VFS can correctly support
	 * read-only remount without racing.
1204
	 */
1205
	WARN_ON(atomic_read(&mp->m_active_trans) != 0);
1206

1207
	xfs_log_quiesce(mp);
1208 1209
}

L
Linus Torvalds 已提交
1210
STATIC int
1211
xfs_fs_remount(
L
Linus Torvalds 已提交
1212 1213 1214 1215
	struct super_block	*sb,
	int			*flags,
	char			*options)
{
1216
	struct xfs_mount	*mp = XFS_M(sb);
1217 1218
	substring_t		args[MAX_OPT_ARGS];
	char			*p;
1219
	int			error;
L
Linus Torvalds 已提交
1220

1221 1222
	while ((p = strsep(&options, ",")) != NULL) {
		int token;
1223

1224 1225
		if (!*p)
			continue;
1226

1227 1228 1229
		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_barrier:
1230
			mp->m_flags |= XFS_MOUNT_BARRIER;
1231 1232
			break;
		case Opt_nobarrier:
1233
			mp->m_flags &= ~XFS_MOUNT_BARRIER;
1234
			break;
1235
		case Opt_inode64:
1236
			mp->m_maxagi = xfs_set_inode64(mp);
1237
			break;
1238 1239 1240
		case Opt_inode32:
			mp->m_maxagi = xfs_set_inode32(mp);
			break;
1241
		default:
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
			/*
			 * Logically we would return an error here to prevent
			 * users from believing they might have changed
			 * mount options using remount which can't be changed.
			 *
			 * But unfortunately mount(8) adds all options from
			 * mtab and fstab to the mount arguments in some cases
			 * so we can't blindly reject options, but have to
			 * check for each specified option if it actually
			 * differs from the currently set option and only
			 * reject it if that's the case.
			 *
			 * Until that is implemented we return success for
			 * every remount request, and silently ignore all
			 * options that we can't actually change.
			 */
#if 0
1259 1260
			xfs_info(mp,
		"mount option \"%s\" not supported for remount\n", p);
1261
			return -EINVAL;
1262
#else
1263
			break;
1264
#endif
1265
		}
1266 1267
	}

1268
	/* ro -> rw */
1269 1270
	if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) {
		mp->m_flags &= ~XFS_MOUNT_RDONLY;
1271 1272 1273 1274 1275 1276 1277 1278

		/*
		 * If this is the first remount to writeable state we
		 * might have some superblock changes to update.
		 */
		if (mp->m_update_flags) {
			error = xfs_mount_log_sb(mp, mp->m_update_flags);
			if (error) {
1279
				xfs_warn(mp, "failed to write sb changes");
1280 1281 1282 1283
				return error;
			}
			mp->m_update_flags = 0;
		}
1284 1285 1286 1287 1288

		/*
		 * Fill out the reserve pool if it is empty. Use the stashed
		 * value if it is non-zero, otherwise go with the default.
		 */
1289
		xfs_restore_resvblks(mp);
1290
		xfs_log_work_queue(mp);
1291 1292 1293 1294
	}

	/* rw -> ro */
	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) {
1295
		/*
1296 1297 1298 1299 1300
		 * Before we sync the metadata, we need to free up the reserve
		 * block pool so that the used block count in the superblock on
		 * disk is correct at the end of the remount. Stash the current
		 * reserve pool size so that if we get remounted rw, we can
		 * return it to the same size.
1301
		 */
1302
		xfs_save_resvblks(mp);
1303
		xfs_quiesce_attr(mp);
1304 1305 1306
		mp->m_flags |= XFS_MOUNT_RDONLY;
	}

1307
	return 0;
L
Linus Torvalds 已提交
1308 1309
}

1310 1311
/*
 * Second stage of a freeze. The data is already frozen so we only
1312
 * need to take care of the metadata. Once that's done write a dummy
1313 1314
 * record to dirty the log in case of a crash while frozen.
 */
1315 1316
STATIC int
xfs_fs_freeze(
L
Linus Torvalds 已提交
1317 1318
	struct super_block	*sb)
{
1319 1320
	struct xfs_mount	*mp = XFS_M(sb);

1321
	xfs_save_resvblks(mp);
1322
	xfs_quiesce_attr(mp);
1323
	return -xfs_fs_log_dummy(mp);
L
Linus Torvalds 已提交
1324 1325
}

1326 1327 1328 1329 1330 1331 1332
STATIC int
xfs_fs_unfreeze(
	struct super_block	*sb)
{
	struct xfs_mount	*mp = XFS_M(sb);

	xfs_restore_resvblks(mp);
1333
	xfs_log_work_queue(mp);
1334 1335 1336
	return 0;
}

L
Linus Torvalds 已提交
1337
STATIC int
1338
xfs_fs_show_options(
L
Linus Torvalds 已提交
1339
	struct seq_file		*m,
1340
	struct dentry		*root)
L
Linus Torvalds 已提交
1341
{
1342
	return -xfs_showargs(XFS_M(root->d_sb), m);
L
Linus Torvalds 已提交
1343 1344
}

1345 1346 1347 1348 1349 1350 1351 1352 1353 1354
/*
 * This function fills in xfs_mount_t fields based on mount args.
 * Note: the superblock _has_ now been read in.
 */
STATIC int
xfs_finish_flags(
	struct xfs_mount	*mp)
{
	int			ronly = (mp->m_flags & XFS_MOUNT_RDONLY);

1355
	/* Fail a mount where the logbuf is smaller than the log stripe */
1356
	if (xfs_sb_version_haslogv2(&mp->m_sb)) {
1357 1358
		if (mp->m_logbsize <= 0 &&
		    mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE) {
1359
			mp->m_logbsize = mp->m_sb.sb_logsunit;
1360 1361
		} else if (mp->m_logbsize > 0 &&
			   mp->m_logbsize < mp->m_sb.sb_logsunit) {
1362 1363
			xfs_warn(mp,
		"logbuf size must be greater than or equal to log stripe size");
1364 1365 1366 1367
			return XFS_ERROR(EINVAL);
		}
	} else {
		/* Fail a mount if the logbuf is larger than 32K */
1368
		if (mp->m_logbsize > XLOG_BIG_RECORD_BSIZE) {
1369 1370
			xfs_warn(mp,
		"logbuf size for version 1 logs must be 16K or 32K");
1371 1372 1373 1374 1375 1376 1377 1378 1379
			return XFS_ERROR(EINVAL);
		}
	}

	/*
	 * mkfs'ed attr2 will turn on attr2 mount unless explicitly
	 * told by noattr2 to turn it off
	 */
	if (xfs_sb_version_hasattr2(&mp->m_sb) &&
1380
	    !(mp->m_flags & XFS_MOUNT_NOATTR2))
1381 1382 1383 1384 1385 1386
		mp->m_flags |= XFS_MOUNT_ATTR2;

	/*
	 * prohibit r/w mounts of read-only filesystems
	 */
	if ((mp->m_sb.sb_flags & XFS_SBF_READONLY) && !ronly) {
1387 1388
		xfs_warn(mp,
			"cannot mount a read-only filesystem as read-write");
1389 1390 1391 1392 1393 1394
		return XFS_ERROR(EROFS);
	}

	return 0;
}

L
Linus Torvalds 已提交
1395
STATIC int
1396
xfs_fs_fill_super(
L
Linus Torvalds 已提交
1397 1398 1399 1400
	struct super_block	*sb,
	void			*data,
	int			silent)
{
1401
	struct inode		*root;
1402
	struct xfs_mount	*mp = NULL;
1403
	int			flags = 0, error = ENOMEM;
1404

1405 1406
	mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL);
	if (!mp)
1407
		goto out;
L
Linus Torvalds 已提交
1408

1409 1410 1411
	spin_lock_init(&mp->m_sb_lock);
	mutex_init(&mp->m_growlock);
	atomic_set(&mp->m_active_trans, 0);
1412
	INIT_DELAYED_WORK(&mp->m_reclaim_work, xfs_reclaim_worker);
1413
	INIT_DELAYED_WORK(&mp->m_eofblocks_work, xfs_eofblocks_worker);
1414

1415 1416
	mp->m_super = sb;
	sb->s_fs_info = mp;
L
Linus Torvalds 已提交
1417

C
Christoph Hellwig 已提交
1418
	error = xfs_parseargs(mp, (char *)data);
1419
	if (error)
1420
		goto out_free_fsname;
L
Linus Torvalds 已提交
1421 1422

	sb_min_blocksize(sb, BBSIZE);
1423
	sb->s_xattr = xfs_xattr_handlers;
1424
	sb->s_export_op = &xfs_export_operations;
1425
#ifdef CONFIG_XFS_QUOTA
1426
	sb->s_qcop = &xfs_quotactl_operations;
1427
#endif
1428
	sb->s_op = &xfs_super_operations;
L
Linus Torvalds 已提交
1429

1430
	if (silent)
1431 1432
		flags |= XFS_MFSI_QUIET;

1433
	error = xfs_open_devices(mp);
1434
	if (error)
C
Christoph Hellwig 已提交
1435
		goto out_free_fsname;
1436

1437
	error = xfs_init_mount_workqueues(mp);
1438 1439
	if (error)
		goto out_close_devices;
1440

1441 1442 1443 1444
	error = xfs_icsb_init_counters(mp);
	if (error)
		goto out_destroy_workqueues;

1445 1446
	error = xfs_readsb(mp, flags);
	if (error)
1447 1448 1449
		goto out_destroy_counters;

	error = xfs_finish_flags(mp);
1450
	if (error)
1451
		goto out_free_sb;
1452

1453
	error = xfs_setup_devices(mp);
1454
	if (error)
1455
		goto out_free_sb;
1456 1457 1458

	error = xfs_filestream_mount(mp);
	if (error)
1459
		goto out_free_sb;
1460

1461 1462 1463 1464
	/*
	 * we must configure the block size in the superblock before we run the
	 * full mount process as the mount process can lookup and cache inodes.
	 */
1465 1466 1467
	sb->s_magic = XFS_SB_MAGIC;
	sb->s_blocksize = mp->m_sb.sb_blocksize;
	sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1;
L
Linus Torvalds 已提交
1468
	sb->s_maxbytes = xfs_max_file_offset(sb->s_blocksize_bits);
1469
	sb->s_max_links = XFS_MAXLINK;
L
Linus Torvalds 已提交
1470 1471 1472
	sb->s_time_gran = 1;
	set_posix_acl_flag(sb);

1473
	error = xfs_mountfs(mp);
1474
	if (error)
1475
		goto out_filestream_unmount;
1476

1477
	root = igrab(VFS_I(mp->m_rootip));
1478
	if (!root) {
C
Christoph Hellwig 已提交
1479
		error = ENOENT;
1480
		goto out_unmount;
C
Christoph Hellwig 已提交
1481
	}
1482 1483
	if (is_bad_inode(root)) {
		error = EINVAL;
1484
		goto out_unmount;
L
Linus Torvalds 已提交
1485
	}
1486
	sb->s_root = d_make_root(root);
1487 1488
	if (!sb->s_root) {
		error = ENOMEM;
1489
		goto out_unmount;
L
Linus Torvalds 已提交
1490
	}
1491

L
Linus Torvalds 已提交
1492
	return 0;
1493

1494
 out_filestream_unmount:
1495
	xfs_filestream_unmount(mp);
1496 1497
 out_free_sb:
	xfs_freesb(mp);
1498
 out_destroy_counters:
1499
	xfs_icsb_destroy_counters(mp);
1500 1501
out_destroy_workqueues:
	xfs_destroy_mount_workqueues(mp);
1502
 out_close_devices:
1503
	xfs_close_devices(mp);
1504 1505
 out_free_fsname:
	xfs_free_fsname(mp);
1506
	kfree(mp);
1507
 out:
1508
	return -error;
1509

1510
 out_unmount:
1511
	xfs_filestream_unmount(mp);
1512
	xfs_unmountfs(mp);
1513
	goto out_free_sb;
L
Linus Torvalds 已提交
1514 1515
}

A
Al Viro 已提交
1516 1517
STATIC struct dentry *
xfs_fs_mount(
L
Linus Torvalds 已提交
1518 1519 1520
	struct file_system_type	*fs_type,
	int			flags,
	const char		*dev_name,
A
Al Viro 已提交
1521
	void			*data)
L
Linus Torvalds 已提交
1522
{
A
Al Viro 已提交
1523
	return mount_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super);
1524 1525
}

1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
static int
xfs_fs_nr_cached_objects(
	struct super_block	*sb)
{
	return xfs_reclaim_inodes_count(XFS_M(sb));
}

static void
xfs_fs_free_cached_objects(
	struct super_block	*sb,
	int			nr_to_scan)
{
	xfs_reclaim_inodes_nr(XFS_M(sb), nr_to_scan);
}

1541
static const struct super_operations xfs_super_operations = {
1542 1543
	.alloc_inode		= xfs_fs_alloc_inode,
	.destroy_inode		= xfs_fs_destroy_inode,
1544
	.evict_inode		= xfs_fs_evict_inode,
1545
	.drop_inode		= xfs_fs_drop_inode,
1546
	.put_super		= xfs_fs_put_super,
1547
	.sync_fs		= xfs_fs_sync_fs,
1548
	.freeze_fs		= xfs_fs_freeze,
1549
	.unfreeze_fs		= xfs_fs_unfreeze,
1550 1551 1552
	.statfs			= xfs_fs_statfs,
	.remount_fs		= xfs_fs_remount,
	.show_options		= xfs_fs_show_options,
1553 1554
	.nr_cached_objects	= xfs_fs_nr_cached_objects,
	.free_cached_objects	= xfs_fs_free_cached_objects,
L
Linus Torvalds 已提交
1555 1556
};

A
Andrew Morton 已提交
1557
static struct file_system_type xfs_fs_type = {
L
Linus Torvalds 已提交
1558 1559
	.owner			= THIS_MODULE,
	.name			= "xfs",
A
Al Viro 已提交
1560
	.mount			= xfs_fs_mount,
L
Linus Torvalds 已提交
1561 1562 1563 1564
	.kill_sb		= kill_block_super,
	.fs_flags		= FS_REQUIRES_DEV,
};

1565 1566 1567 1568 1569 1570
STATIC int __init
xfs_init_zones(void)
{

	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
	if (!xfs_ioend_zone)
1571
		goto out;
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586

	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
						  xfs_ioend_zone);
	if (!xfs_ioend_pool)
		goto out_destroy_ioend_zone;

	xfs_log_ticket_zone = kmem_zone_init(sizeof(xlog_ticket_t),
						"xfs_log_ticket");
	if (!xfs_log_ticket_zone)
		goto out_destroy_ioend_pool;

	xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t),
						"xfs_bmap_free_item");
	if (!xfs_bmap_free_item_zone)
		goto out_destroy_log_ticket_zone;
1587

1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599
	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
						"xfs_btree_cur");
	if (!xfs_btree_cur_zone)
		goto out_destroy_bmap_free_item_zone;

	xfs_da_state_zone = kmem_zone_init(sizeof(xfs_da_state_t),
						"xfs_da_state");
	if (!xfs_da_state_zone)
		goto out_destroy_btree_cur_zone;

	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
	if (!xfs_ifork_zone)
1600
		goto out_destroy_da_state_zone;
1601 1602 1603 1604 1605

	xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
	if (!xfs_trans_zone)
		goto out_destroy_ifork_zone;

1606 1607 1608 1609 1610 1611
	xfs_log_item_desc_zone =
		kmem_zone_init(sizeof(struct xfs_log_item_desc),
			       "xfs_log_item_desc");
	if (!xfs_log_item_desc_zone)
		goto out_destroy_trans_zone;

1612 1613 1614 1615 1616
	/*
	 * The size of the zone allocated buf log item is the maximum
	 * size possible under XFS.  This wastes a little bit of memory,
	 * but it is much faster.
	 */
1617 1618
	xfs_buf_item_zone = kmem_zone_init(sizeof(struct xfs_buf_log_item),
					   "xfs_buf_item");
1619
	if (!xfs_buf_item_zone)
1620
		goto out_destroy_log_item_desc_zone;
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635

	xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
			((XFS_EFD_MAX_FAST_EXTENTS - 1) *
				 sizeof(xfs_extent_t))), "xfs_efd_item");
	if (!xfs_efd_zone)
		goto out_destroy_buf_item_zone;

	xfs_efi_zone = kmem_zone_init((sizeof(xfs_efi_log_item_t) +
			((XFS_EFI_MAX_FAST_EXTENTS - 1) *
				sizeof(xfs_extent_t))), "xfs_efi_item");
	if (!xfs_efi_zone)
		goto out_destroy_efd_zone;

	xfs_inode_zone =
		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
1636 1637
			KM_ZONE_HWALIGN | KM_ZONE_RECLAIM | KM_ZONE_SPREAD,
			xfs_fs_inode_init_once);
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656
	if (!xfs_inode_zone)
		goto out_destroy_efi_zone;

	xfs_ili_zone =
		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
					KM_ZONE_SPREAD, NULL);
	if (!xfs_ili_zone)
		goto out_destroy_inode_zone;

	return 0;

 out_destroy_inode_zone:
	kmem_zone_destroy(xfs_inode_zone);
 out_destroy_efi_zone:
	kmem_zone_destroy(xfs_efi_zone);
 out_destroy_efd_zone:
	kmem_zone_destroy(xfs_efd_zone);
 out_destroy_buf_item_zone:
	kmem_zone_destroy(xfs_buf_item_zone);
1657 1658
 out_destroy_log_item_desc_zone:
	kmem_zone_destroy(xfs_log_item_desc_zone);
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
 out_destroy_trans_zone:
	kmem_zone_destroy(xfs_trans_zone);
 out_destroy_ifork_zone:
	kmem_zone_destroy(xfs_ifork_zone);
 out_destroy_da_state_zone:
	kmem_zone_destroy(xfs_da_state_zone);
 out_destroy_btree_cur_zone:
	kmem_zone_destroy(xfs_btree_cur_zone);
 out_destroy_bmap_free_item_zone:
	kmem_zone_destroy(xfs_bmap_free_item_zone);
 out_destroy_log_ticket_zone:
	kmem_zone_destroy(xfs_log_ticket_zone);
 out_destroy_ioend_pool:
	mempool_destroy(xfs_ioend_pool);
 out_destroy_ioend_zone:
	kmem_zone_destroy(xfs_ioend_zone);
 out:
	return -ENOMEM;
}

STATIC void
xfs_destroy_zones(void)
{
1682 1683 1684 1685 1686
	/*
	 * Make sure all delayed rcu free are flushed before we
	 * destroy caches.
	 */
	rcu_barrier();
1687 1688 1689 1690 1691
	kmem_zone_destroy(xfs_ili_zone);
	kmem_zone_destroy(xfs_inode_zone);
	kmem_zone_destroy(xfs_efi_zone);
	kmem_zone_destroy(xfs_efd_zone);
	kmem_zone_destroy(xfs_buf_item_zone);
1692
	kmem_zone_destroy(xfs_log_item_desc_zone);
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702
	kmem_zone_destroy(xfs_trans_zone);
	kmem_zone_destroy(xfs_ifork_zone);
	kmem_zone_destroy(xfs_da_state_zone);
	kmem_zone_destroy(xfs_btree_cur_zone);
	kmem_zone_destroy(xfs_bmap_free_item_zone);
	kmem_zone_destroy(xfs_log_ticket_zone);
	mempool_destroy(xfs_ioend_pool);
	kmem_zone_destroy(xfs_ioend_zone);

}
L
Linus Torvalds 已提交
1703

1704 1705 1706
STATIC int __init
xfs_init_workqueues(void)
{
1707 1708 1709 1710 1711 1712 1713 1714
	/*
	 * The allocation workqueue can be used in memory reclaim situations
	 * (writepage path), and parallelism is only limited by the number of
	 * AGs in all the filesystems mounted. Hence use the default large
	 * max_active value for this workqueue.
	 */
	xfs_alloc_wq = alloc_workqueue("xfsalloc", WQ_MEM_RECLAIM, 0);
	if (!xfs_alloc_wq)
1715
		return -ENOMEM;
1716

1717 1718 1719
	return 0;
}

1720
STATIC void
1721 1722
xfs_destroy_workqueues(void)
{
1723
	destroy_workqueue(xfs_alloc_wq);
1724 1725
}

L
Linus Torvalds 已提交
1726
STATIC int __init
1727
init_xfs_fs(void)
L
Linus Torvalds 已提交
1728 1729 1730
{
	int			error;

1731 1732
	printk(KERN_INFO XFS_VERSION_STRING " with "
			 XFS_BUILD_OPTIONS " enabled\n");
L
Linus Torvalds 已提交
1733

1734
	xfs_dir_startup();
L
Linus Torvalds 已提交
1735

1736
	error = xfs_init_zones();
1737 1738 1739
	if (error)
		goto out;

1740
	error = xfs_init_workqueues();
1741
	if (error)
1742
		goto out_destroy_zones;
1743

1744 1745 1746 1747
	error = xfs_mru_cache_init();
	if (error)
		goto out_destroy_wq;

1748 1749 1750
	error = xfs_filestream_init();
	if (error)
		goto out_mru_cache_uninit;
L
Linus Torvalds 已提交
1751

1752
	error = xfs_buf_init();
1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
	if (error)
		goto out_filestream_uninit;

	error = xfs_init_procfs();
	if (error)
		goto out_buf_terminate;

	error = xfs_sysctl_register();
	if (error)
		goto out_cleanup_procfs;
L
Linus Torvalds 已提交
1763

1764 1765 1766
	error = xfs_qm_init();
	if (error)
		goto out_sysctl_unregister;
L
Linus Torvalds 已提交
1767 1768 1769

	error = register_filesystem(&xfs_fs_type);
	if (error)
1770
		goto out_qm_exit;
L
Linus Torvalds 已提交
1771 1772
	return 0;

1773 1774
 out_qm_exit:
	xfs_qm_exit();
1775 1776 1777 1778 1779
 out_sysctl_unregister:
	xfs_sysctl_unregister();
 out_cleanup_procfs:
	xfs_cleanup_procfs();
 out_buf_terminate:
1780
	xfs_buf_terminate();
1781 1782 1783 1784
 out_filestream_uninit:
	xfs_filestream_uninit();
 out_mru_cache_uninit:
	xfs_mru_cache_uninit();
1785 1786
 out_destroy_wq:
	xfs_destroy_workqueues();
1787
 out_destroy_zones:
1788
	xfs_destroy_zones();
1789
 out:
L
Linus Torvalds 已提交
1790 1791 1792 1793
	return error;
}

STATIC void __exit
1794
exit_xfs_fs(void)
L
Linus Torvalds 已提交
1795
{
1796
	xfs_qm_exit();
L
Linus Torvalds 已提交
1797
	unregister_filesystem(&xfs_fs_type);
1798 1799
	xfs_sysctl_unregister();
	xfs_cleanup_procfs();
1800
	xfs_buf_terminate();
1801 1802
	xfs_filestream_uninit();
	xfs_mru_cache_uninit();
1803
	xfs_destroy_workqueues();
1804
	xfs_destroy_zones();
L
Linus Torvalds 已提交
1805 1806 1807 1808 1809 1810 1811 1812
}

module_init(init_xfs_fs);
module_exit(exit_xfs_fs);

MODULE_AUTHOR("Silicon Graphics, Inc.");
MODULE_DESCRIPTION(XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled");
MODULE_LICENSE("GPL");
反馈
建议
客服 返回
顶部