xfs_super.c 48.3 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
 */
#include "xfs.h"
19
#include "xfs_bit.h"
L
Linus Torvalds 已提交
20
#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 30
#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"
31
#include "xfs_alloc_btree.h"
L
Linus Torvalds 已提交
32 33
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
34
#include "xfs_attr_sf.h"
L
Linus Torvalds 已提交
35 36
#include "xfs_dinode.h"
#include "xfs_inode.h"
37
#include "xfs_btree.h"
38
#include "xfs_btree_trace.h"
39
#include "xfs_ialloc.h"
L
Linus Torvalds 已提交
40 41 42 43
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
C
Christoph Hellwig 已提交
44
#include "xfs_fsops.h"
L
Linus Torvalds 已提交
45 46 47 48 49
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
50
#include "xfs_vnodeops.h"
L
Linus Torvalds 已提交
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 59
#include "xfs_da_btree.h"
#include "xfs_dir2_trace.h"
#include "xfs_extfree_item.h"
#include "xfs_mru_cache.h"
#include "xfs_inode_item.h"
60
#include "xfs_sync.h"
L
Linus Torvalds 已提交
61 62 63 64

#include <linux/namei.h>
#include <linux/init.h>
#include <linux/mount.h>
65
#include <linux/mempool.h>
L
Linus Torvalds 已提交
66
#include <linux/writeback.h>
67
#include <linux/kthread.h>
68
#include <linux/freezer.h>
69
#include <linux/parser.h>
L
Linus Torvalds 已提交
70

71 72
static struct super_operations xfs_super_operations;
static kmem_zone_t *xfs_ioend_zone;
73
mempool_t *xfs_ioend_pool;
L
Linus Torvalds 已提交
74

75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
#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_INO64	"ino64"		/* force inodes into 64-bit range */
#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) */

123 124 125 126 127 128 129 130 131 132
/*
 * 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
};

133
static const match_table_t tokens = {
134 135 136 137 138 139
	{Opt_barrier, "barrier"},
	{Opt_nobarrier, "nobarrier"},
	{Opt_err, NULL}
};


140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162
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;
}

163 164 165 166 167 168 169
/*
 * 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.
 */
170 171 172 173
STATIC int
xfs_parseargs(
	struct xfs_mount	*mp,
	char			*options,
174
	char			**mtpt)
175
{
176
	struct super_block	*sb = mp->m_super;
177
	char			*this_char, *value, *eov;
178 179 180
	int			dsunit = 0;
	int			dswidth = 0;
	int			iosize = 0;
181
	int			dmapi_implies_ikeep = 1;
182
	__uint8_t		iosizelog = 0;
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200

	/*
	 * 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;
201

202 203 204 205 206
	/*
	 * These can be overridden by the mount option parsing.
	 */
	mp->m_logbufs = -1;
	mp->m_logbsize = -1;
207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223

	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;
			}
224
			mp->m_logbufs = simple_strtoul(value, &eov, 10);
225 226 227 228 229 230 231
		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
232
			mp->m_logbsize = suffix_strtoul(value, &eov, 10);
233 234 235 236 237 238 239
		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
240 241 242
			mp->m_logname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_logname)
				return ENOMEM;
243 244 245 246 247 248 249
		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
250 251 252
			*mtpt = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!*mtpt)
				return ENOMEM;
253 254 255 256 257 258 259
		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
260 261 262
			mp->m_rtname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
			if (!mp->m_rtname)
				return ENOMEM;
263 264 265 266 267 268 269 270
		} 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);
C
Christoph Hellwig 已提交
271
			iosizelog = ffs(iosize) - 1;
272 273 274 275 276 277 278 279
		} 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);
280
			iosizelog = ffs(iosize) - 1;
281 282 283 284 285 286 287
		} 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)) {
288
			mp->m_flags |= XFS_MOUNT_WSYNC;
289
		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
290
			mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
291
		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
292
			mp->m_flags |= XFS_MOUNT_NORECOVERY;
293
		} else if (!strcmp(this_char, MNTOPT_INO64)) {
294 295 296 297
#if XFS_BIG_INUMS
			mp->m_flags |= XFS_MOUNT_INO64;
			mp->m_inoadd = XFS_INO64_OFFSET;
#else
298 299 300 301 302 303
			cmn_err(CE_WARN,
				"XFS: %s option not allowed on this system",
				this_char);
			return EINVAL;
#endif
		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
304
			mp->m_flags |= XFS_MOUNT_NOALIGN;
305
		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
306
			mp->m_flags |= XFS_MOUNT_SWALLOC;
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323
		} 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)) {
324
			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
325 326 327 328 329 330 331
#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)) {
332
			mp->m_flags |= XFS_MOUNT_NOUUID;
333
		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
334
			mp->m_flags |= XFS_MOUNT_BARRIER;
335
		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
336
			mp->m_flags &= ~XFS_MOUNT_BARRIER;
337
		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
338
			mp->m_flags |= XFS_MOUNT_IKEEP;
339
		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
340
			dmapi_implies_ikeep = 0;
341
			mp->m_flags &= ~XFS_MOUNT_IKEEP;
342
		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
343
			mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
344
		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
345
			mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
346
		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
347
			mp->m_flags |= XFS_MOUNT_ATTR2;
348
		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
349 350
			mp->m_flags &= ~XFS_MOUNT_ATTR2;
			mp->m_flags |= XFS_MOUNT_NOATTR2;
351
		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
352
			mp->m_flags |= XFS_MOUNT_FILESTREAMS;
353
		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
354 355
			mp->m_qflags &= ~(XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
					  XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
356
					  XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
357
					  XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD);
358 359 360
		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
			   !strcmp(this_char, MNTOPT_UQUOTA) ||
			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
361 362
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
					 XFS_UQUOTA_ENFD);
363 364
		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
365 366
			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_UQUOTA_ENFD;
367 368
		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
369 370
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
371
		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
372 373
			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
374 375
		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
376 377
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
					 XFS_OQUOTA_ENFD);
378
		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
379 380
			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
381
		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
382
			mp->m_flags |= XFS_MOUNT_DMAPI;
383
		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
384
			mp->m_flags |= XFS_MOUNT_DMAPI;
385
		} else if (!strcmp(this_char, MNTOPT_DMI)) {
386
			mp->m_flags |= XFS_MOUNT_DMAPI;
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
		} 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;
		}
	}

404 405 406 407 408 409 410
	/*
	 * 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;
411 412
	}

413
	if ((mp->m_flags & XFS_MOUNT_NOALIGN) && (dsunit || dswidth)) {
414 415 416 417 418
		cmn_err(CE_WARN,
	"XFS: sunit and swidth options incompatible with the noalign option");
		return EINVAL;
	}

419 420
	if ((mp->m_qflags & (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE)) &&
	    (mp->m_qflags & (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE))) {
421 422 423 424 425
		cmn_err(CE_WARN,
			"XFS: cannot mount with both project and group quota");
		return EINVAL;
	}

426
	if ((mp->m_flags & XFS_MOUNT_DMAPI) && (!*mtpt || *mtpt[0] == '\0')) {
427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
		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.
	 */
454 455
	if ((mp->m_flags & XFS_MOUNT_DMAPI) && dmapi_implies_ikeep)
		mp->m_flags |= XFS_MOUNT_IKEEP;
456

457 458 459 460 461 462 463 464
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.
		 */
465
		if (dsunit) {
466 467
			mp->m_dalign = dsunit;
			mp->m_flags |= XFS_MOUNT_RETERR;
468
		}
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511

		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;
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528
	}

	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 */
529
		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601
		{ XFS_MOUNT_WSYNC,		"," MNTOPT_WSYNC },
		{ XFS_MOUNT_INO64,		"," MNTOPT_INO64 },
		{ 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);

	if (mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
		seq_puts(m, "," MNTOPT_PRJQUOTA);
	else if (mp->m_qflags & XFS_PQUOTA_ACCT)
		seq_puts(m, "," MNTOPT_PQUOTANOENF);

	if (mp->m_qflags & (XFS_GQUOTA_ACCT|XFS_OQUOTA_ENFD))
		seq_puts(m, "," MNTOPT_GRPQUOTA);
	else if (mp->m_qflags & XFS_GQUOTA_ACCT)
		seq_puts(m, "," MNTOPT_GQUOTANOENF);

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

	return 0;
}
L
Linus Torvalds 已提交
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635
__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
# if defined(CONFIG_LBD)
	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;
}

H
Hannes Eder 已提交
636
STATIC int
L
Linus Torvalds 已提交
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);
L
Linus Torvalds 已提交
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;
}

H
Hannes Eder 已提交
653
STATIC void
L
Linus Torvalds 已提交
654 655 656 657
xfs_blkdev_put(
	struct block_device	*bdev)
{
	if (bdev)
658
		close_bdev_exclusive(bdev, FMODE_READ|FMODE_WRITE);
L
Linus Torvalds 已提交
659 660
}

661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
/*
 * 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;
}

void
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
}
L
Linus Torvalds 已提交
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(
D
David Chinner 已提交
867
	struct xfs_ail		*ailp,
868 869
	xfs_lsn_t		threshold_lsn)
{
D
David Chinner 已提交
870 871
	ailp->xa_target = threshold_lsn;
	wake_up_process(ailp->xa_task);
872 873
}

H
Hannes Eder 已提交
874
STATIC int
875 876 877
xfsaild(
	void	*data)
{
D
David Chinner 已提交
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();

D
David Chinner 已提交
890 891
		ASSERT(ailp->xa_mount->m_log);
		if (XFS_FORCED_SHUTDOWN(ailp->xa_mount))
892 893
			continue;

D
David Chinner 已提交
894
		tout = xfsaild_push(ailp, &last_pushed_lsn);
895 896 897 898 899 900 901
	}

	return 0;
}	/* xfsaild */

int
xfsaild_start(
D
David Chinner 已提交
902
	struct xfs_ail	*ailp)
903
{
D
David Chinner 已提交
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(
D
David Chinner 已提交
913
	struct xfs_ail	*ailp)
914
{
D
David Chinner 已提交
915
	kthread_stop(ailp->xa_task);
916 917 918
}


919
/* Catch misguided souls that try to use this interface on XFS */
L
Linus Torvalds 已提交
920
STATIC struct inode *
921
xfs_fs_alloc_inode(
L
Linus Torvalds 已提交
922 923
	struct super_block	*sb)
{
924
	BUG();
925
	return NULL;
L
Linus Torvalds 已提交
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
 */
L
Linus Torvalds 已提交
932
STATIC void
933
xfs_fs_destroy_inode(
934
	struct inode	*inode)
L
Linus Torvalds 已提交
935
{
936 937 938 939 940
	xfs_inode_t		*ip = XFS_I(inode);

	XFS_STATS_INC(vn_reclaim);
	if (xfs_reclaim(ip))
		panic("%s: cannot reclaim 0x%p\n", __func__, inode);
L
Linus Torvalds 已提交
941 942
}

943 944 945 946 947 948 949 950
/*
 * 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.
 */
951 952
STATIC void
xfs_fs_inode_init_once(
953 954 955 956 957
	void			*inode)
{
	struct xfs_inode	*ip = inode;

	memset(ip, 0, sizeof(struct xfs_inode));
958 959 960 961 962

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

	/* xfs inode */
963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
	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);
	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
}

L
Linus Torvalds 已提交
980 981 982 983
/*
 * Attempt to flush the inode, this will actually fail
 * if the inode is pinned, but we dirty the inode again
 * at the point when it is unpinned after a log write,
984
 * since this is when the inode itself becomes flushable.
L
Linus Torvalds 已提交
985 986
 */
STATIC int
987
xfs_fs_write_inode(
L
Linus Torvalds 已提交
988 989 990
	struct inode		*inode,
	int			sync)
{
991
	struct xfs_inode	*ip = XFS_I(inode);
992
	struct xfs_mount	*mp = ip->i_mount;
993
	int			error = 0;
L
Linus Torvalds 已提交
994

995
	xfs_itrace_entry(ip);
996 997 998 999

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

1000
	if (sync) {
1001 1002
		error = xfs_wait_on_pages(ip, 0, -1);
		if (error)
1003
			goto out;
L
Linus Torvalds 已提交
1004
	}
1005

1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
	/*
	 * Bypass inodes which have already been cleaned by
	 * the inode flush clustering code inside xfs_iflush
	 */
	if (xfs_inode_clean(ip))
		goto out;

	/*
	 * We make this non-blocking if the inode is contended, return
	 * EAGAIN to indicate to the caller that they did not succeed.
	 * This prevents the flush path from blocking on inodes inside
	 * another operation right now, they get caught later by xfs_sync.
	 */
	if (sync) {
		xfs_ilock(ip, XFS_ILOCK_SHARED);
		xfs_iflock(ip);

		error = xfs_iflush(ip, XFS_IFLUSH_SYNC);
	} else {
		error = EAGAIN;
		if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED))
			goto out;
		if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip))
			goto out_unlock;

		error = xfs_iflush(ip, XFS_IFLUSH_ASYNC_NOBLOCK);
	}

 out_unlock:
	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 out:
1037 1038 1039 1040 1041
	/*
	 * if we failed to write out the inode then mark
	 * it dirty again so we'll try again later.
	 */
	if (error)
1042
		xfs_mark_inode_dirty_sync(ip);
L
Linus Torvalds 已提交
1043 1044 1045 1046
	return -error;
}

STATIC void
1047
xfs_fs_clear_inode(
L
Linus Torvalds 已提交
1048 1049
	struct inode		*inode)
{
1050
	xfs_inode_t		*ip = XFS_I(inode);
1051

1052 1053 1054 1055 1056 1057
	xfs_itrace_entry(ip);
	XFS_STATS_INC(vn_rele);
	XFS_STATS_INC(vn_remove);
	XFS_STATS_DEC(vn_active);

	xfs_inactive(ip);
1058
}
L
Linus Torvalds 已提交
1059

1060 1061 1062 1063 1064 1065 1066 1067 1068
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
1069
STATIC void
1070
xfs_fs_put_super(
L
Linus Torvalds 已提交
1071 1072
	struct super_block	*sb)
{
1073
	struct xfs_mount	*mp = XFS_M(sb);
1074 1075
	struct xfs_inode	*rip = mp->m_rootip;
	int			unmount_event_flags = 0;
L
Linus Torvalds 已提交
1076

1077
	xfs_syncd_stop(mp);
1078
	xfs_sync_inodes(mp, SYNC_ATTR|SYNC_DELWRI);
1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108

#ifdef HAVE_DMAPI
	if (mp->m_flags & XFS_MOUNT_DMAPI) {
		unmount_event_flags =
			(mp->m_dmevmask & (1 << DM_EVENT_UNMOUNT)) ?
				0 : DM_FLAGS_UNWANTED;
		/*
		 * Ignore error from dmapi here, first unmount is not allowed
		 * to fail anyway, and second we wouldn't want to fail a
		 * unmount because of dmapi.
		 */
		XFS_SEND_PREUNMOUNT(mp, rip, DM_RIGHT_NULL, rip, DM_RIGHT_NULL,
				NULL, NULL, 0, 0, unmount_event_flags);
	}
#endif

	/*
	 * Blow away any referenced inode in the filestreams cache.
	 * This can and will cause log traffic as inodes go inactive
	 * here.
	 */
	xfs_filestream_unmount(mp);

	XFS_bflush(mp->m_ddev_targp);

	if (mp->m_flags & XFS_MOUNT_DMAPI) {
		XFS_SEND_UNMOUNT(mp, rip, DM_RIGHT_NULL, 0, 0,
				unmount_event_flags);
	}

1109
	xfs_unmountfs(mp);
1110
	xfs_freesb(mp);
C
Christoph Hellwig 已提交
1111
	xfs_icsb_destroy_counters(mp);
1112
	xfs_close_devices(mp);
1113 1114
	xfs_qmops_put(mp);
	xfs_dmops_put(mp);
1115
	xfs_free_fsname(mp);
C
Christoph Hellwig 已提交
1116
	kfree(mp);
L
Linus Torvalds 已提交
1117 1118 1119
}

STATIC void
1120
xfs_fs_write_super(
L
Linus Torvalds 已提交
1121 1122
	struct super_block	*sb)
{
1123
	if (!(sb->s_flags & MS_RDONLY))
1124
		xfs_sync_fsdata(XFS_M(sb), 0);
L
Linus Torvalds 已提交
1125 1126 1127 1128
	sb->s_dirt = 0;
}

STATIC int
1129
xfs_fs_sync_super(
L
Linus Torvalds 已提交
1130 1131 1132
	struct super_block	*sb,
	int			wait)
{
1133
	struct xfs_mount	*mp = XFS_M(sb);
1134
	int			error;
L
Linus Torvalds 已提交
1135

1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
	/*
	 * Treat a sync operation like a freeze.  This is to work
	 * around a race in sync_inodes() which works in two phases
	 * - an asynchronous flush, which can write out an inode
	 * without waiting for file size updates to complete, and a
	 * synchronous flush, which wont do anything because the
	 * async flush removed the inode's dirty flag.  Also
	 * sync_inodes() will not see any files that just have
	 * outstanding transactions to be flushed because we don't
	 * dirty the Linux inode until after the transaction I/O
	 * completes.
	 */
1148 1149 1150 1151
	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE))
		error = xfs_quiesce_data(mp);
	else
		error = xfs_sync_fsdata(mp, 0);
L
Linus Torvalds 已提交
1152 1153 1154
	sb->s_dirt = 0;

	if (unlikely(laptop_mode)) {
1155
		int	prev_sync_seq = mp->m_sync_seq;
L
Linus Torvalds 已提交
1156 1157 1158 1159 1160 1161

		/*
		 * The disk must be active because we're syncing.
		 * We schedule xfssyncd now (now that the disk is
		 * active) instead of later (when it might not be).
		 */
1162
		wake_up_process(mp->m_sync_task);
L
Linus Torvalds 已提交
1163 1164 1165 1166 1167 1168
		/*
		 * We have to wait for the sync iteration to complete.
		 * If we don't, the disk activity caused by the sync
		 * will come after the sync is completed, and that
		 * triggers another sync from laptop mode.
		 */
1169 1170
		wait_event(mp->m_wait_single_sync_task,
				mp->m_sync_seq != prev_sync_seq);
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175 1176
	}

	return -error;
}

STATIC int
1177
xfs_fs_statfs(
1178
	struct dentry		*dentry,
L
Linus Torvalds 已提交
1179 1180
	struct kstatfs		*statp)
{
C
Christoph Hellwig 已提交
1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
	struct xfs_mount	*mp = XFS_M(dentry->d_sb);
	xfs_sb_t		*sbp = &mp->m_sb;
	__uint64_t		fakeinos, id;
	xfs_extlen_t		lsize;

	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);

1193
	xfs_icsb_sync_counters(mp, XFS_ICSB_LAZY_COUNT);
C
Christoph Hellwig 已提交
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218

	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;
#if XFS_BIG_INUMS
	fakeinos += mp->m_inoadd;
#endif
	statp->f_files =
	    MIN(sbp->sb_icount + fakeinos, (__uint64_t)XFS_MAXINUMBER);
	if (mp->m_maxicount)
#if XFS_BIG_INUMS
		if (!mp->m_inoadd)
#endif
			statp->f_files = min_t(typeof(statp->f_files),
						statp->f_files,
						mp->m_maxicount);
	statp->f_ffree = statp->f_files - (sbp->sb_icount - sbp->sb_ifree);
	spin_unlock(&mp->m_sb_lock);

	XFS_QM_DQSTATVFS(XFS_I(dentry->d_inode), statp);
	return 0;
L
Linus Torvalds 已提交
1219 1220 1221
}

STATIC int
1222
xfs_fs_remount(
L
Linus Torvalds 已提交
1223 1224 1225 1226
	struct super_block	*sb,
	int			*flags,
	char			*options)
{
1227
	struct xfs_mount	*mp = XFS_M(sb);
1228 1229
	substring_t		args[MAX_OPT_ARGS];
	char			*p;
1230
	int			error;
L
Linus Torvalds 已提交
1231

1232 1233
	while ((p = strsep(&options, ",")) != NULL) {
		int token;
1234

1235 1236
		if (!*p)
			continue;
1237

1238 1239 1240
		token = match_token(p, tokens, args);
		switch (token) {
		case Opt_barrier:
1241
			mp->m_flags |= XFS_MOUNT_BARRIER;
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251

			/*
			 * Test if barriers are actually working if we can,
			 * else delay this check until the filesystem is
			 * marked writeable.
			 */
			if (!(mp->m_flags & XFS_MOUNT_RDONLY))
				xfs_mountfs_check_barriers(mp);
			break;
		case Opt_nobarrier:
1252
			mp->m_flags &= ~XFS_MOUNT_BARRIER;
1253 1254
			break;
		default:
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271
			/*
			 * 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
1272 1273 1274
			printk(KERN_INFO
	"XFS: mount option \"%s\" not supported for remount\n", p);
			return -EINVAL;
1275
#else
1276
			break;
1277
#endif
1278
		}
1279 1280
	}

1281
	/* ro -> rw */
1282 1283 1284 1285
	if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) {
		mp->m_flags &= ~XFS_MOUNT_RDONLY;
		if (mp->m_flags & XFS_MOUNT_BARRIER)
			xfs_mountfs_check_barriers(mp);
1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299

		/*
		 * 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) {
				cmn_err(CE_WARN,
					"XFS: failed to write sb changes");
				return error;
			}
			mp->m_update_flags = 0;
		}
1300 1301 1302 1303
	}

	/* rw -> ro */
	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) {
1304
		xfs_quiesce_data(mp);
D
David Chinner 已提交
1305
		xfs_quiesce_attr(mp);
1306 1307 1308
		mp->m_flags |= XFS_MOUNT_RDONLY;
	}

1309
	return 0;
L
Linus Torvalds 已提交
1310 1311
}

C
Christoph Hellwig 已提交
1312 1313
/*
 * Second stage of a freeze. The data is already frozen so we only
D
David Chinner 已提交
1314
 * need to take care of the metadata. Once that's done write a dummy
C
Christoph Hellwig 已提交
1315 1316
 * record to dirty the log in case of a crash while frozen.
 */
1317 1318
STATIC int
xfs_fs_freeze(
L
Linus Torvalds 已提交
1319 1320
	struct super_block	*sb)
{
C
Christoph Hellwig 已提交
1321 1322
	struct xfs_mount	*mp = XFS_M(sb);

D
David Chinner 已提交
1323
	xfs_quiesce_attr(mp);
1324
	return -xfs_fs_log_dummy(mp);
L
Linus Torvalds 已提交
1325 1326 1327
}

STATIC int
1328
xfs_fs_show_options(
L
Linus Torvalds 已提交
1329 1330 1331
	struct seq_file		*m,
	struct vfsmount		*mnt)
{
1332
	return -xfs_showargs(XFS_M(mnt->mnt_sb), m);
L
Linus Torvalds 已提交
1333 1334
}

1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
/*
 * 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);

1345
	/* Fail a mount where the logbuf is smaller than the log stripe */
1346
	if (xfs_sb_version_haslogv2(&mp->m_sb)) {
1347 1348
		if (mp->m_logbsize <= 0 &&
		    mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE) {
1349
			mp->m_logbsize = mp->m_sb.sb_logsunit;
1350 1351
		} else if (mp->m_logbsize > 0 &&
			   mp->m_logbsize < mp->m_sb.sb_logsunit) {
1352 1353 1354 1355 1356 1357
			cmn_err(CE_WARN,
	"XFS: logbuf size must be greater than or equal to log stripe size");
			return XFS_ERROR(EINVAL);
		}
	} else {
		/* Fail a mount if the logbuf is larger than 32K */
1358
		if (mp->m_logbsize > XLOG_BIG_RECORD_BSIZE) {
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369
			cmn_err(CE_WARN,
	"XFS: logbuf size for version 1 logs must be 16K or 32K");
			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) &&
1370
	    !(mp->m_flags & XFS_MOUNT_NOATTR2))
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
		mp->m_flags |= XFS_MOUNT_ATTR2;

	/*
	 * prohibit r/w mounts of read-only filesystems
	 */
	if ((mp->m_sb.sb_flags & XFS_SBF_READONLY) && !ronly) {
		cmn_err(CE_WARN,
	"XFS: cannot mount a read-only filesystem as read-write");
		return XFS_ERROR(EROFS);
	}

	return 0;
}

L
Linus Torvalds 已提交
1385
STATIC int
1386
xfs_fs_fill_super(
L
Linus Torvalds 已提交
1387 1388 1389 1390
	struct super_block	*sb,
	void			*data,
	int			silent)
{
1391
	struct inode		*root;
1392
	struct xfs_mount	*mp = NULL;
C
Christoph Hellwig 已提交
1393
	int			flags = 0, error = ENOMEM;
1394
	char			*mtpt = NULL;
1395

C
Christoph Hellwig 已提交
1396 1397
	mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL);
	if (!mp)
1398
		goto out;
L
Linus Torvalds 已提交
1399

C
Christoph Hellwig 已提交
1400 1401 1402
	spin_lock_init(&mp->m_sb_lock);
	mutex_init(&mp->m_growlock);
	atomic_set(&mp->m_active_trans, 0);
1403 1404 1405 1406
	INIT_LIST_HEAD(&mp->m_sync_list);
	spin_lock_init(&mp->m_sync_lock);
	init_waitqueue_head(&mp->m_wait_single_sync_task);

C
Christoph Hellwig 已提交
1407 1408
	mp->m_super = sb;
	sb->s_fs_info = mp;
L
Linus Torvalds 已提交
1409

1410
	error = xfs_parseargs(mp, (char *)data, &mtpt);
1411
	if (error)
1412
		goto out_free_fsname;
L
Linus Torvalds 已提交
1413 1414

	sb_min_blocksize(sb, BBSIZE);
1415
	sb->s_xattr = xfs_xattr_handlers;
1416
	sb->s_export_op = &xfs_export_operations;
1417
#ifdef CONFIG_XFS_QUOTA
1418
	sb->s_qcop = &xfs_quotactl_operations;
1419
#endif
1420
	sb->s_op = &xfs_super_operations;
L
Linus Torvalds 已提交
1421

1422
	error = xfs_dmops_get(mp);
1423
	if (error)
1424 1425
		goto out_free_fsname;
	error = xfs_qmops_get(mp);
1426
	if (error)
1427
		goto out_put_dmops;
L
Linus Torvalds 已提交
1428

1429
	if (silent)
1430 1431
		flags |= XFS_MFSI_QUIET;

1432
	error = xfs_open_devices(mp);
1433 1434
	if (error)
		goto out_put_qmops;
1435

C
Christoph Hellwig 已提交
1436 1437 1438
	if (xfs_icsb_init_counters(mp))
		mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;

1439 1440
	error = xfs_readsb(mp, flags);
	if (error)
1441 1442 1443
		goto out_destroy_counters;

	error = xfs_finish_flags(mp);
1444
	if (error)
1445
		goto out_free_sb;
1446

1447
	error = xfs_setup_devices(mp);
1448
	if (error)
1449
		goto out_free_sb;
1450 1451 1452 1453 1454 1455

	if (mp->m_flags & XFS_MOUNT_BARRIER)
		xfs_mountfs_check_barriers(mp);

	error = xfs_filestream_mount(mp);
	if (error)
1456
		goto out_free_sb;
1457

C
Christoph Hellwig 已提交
1458
	error = xfs_mountfs(mp);
1459
	if (error)
1460
		goto out_filestream_unmount;
1461

1462
	XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, mtpt, mp->m_fsname);
1463

L
Linus Torvalds 已提交
1464
	sb->s_dirt = 1;
C
Christoph Hellwig 已提交
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 1469 1470 1471
	sb->s_maxbytes = xfs_max_file_offset(sb->s_blocksize_bits);
	sb->s_time_gran = 1;
	set_posix_acl_flag(sb);

1472
	root = igrab(VFS_I(mp->m_rootip));
1473
	if (!root) {
C
Christoph Hellwig 已提交
1474
		error = ENOENT;
L
Linus Torvalds 已提交
1475
		goto fail_unmount;
C
Christoph Hellwig 已提交
1476
	}
1477 1478
	if (is_bad_inode(root)) {
		error = EINVAL;
L
Linus Torvalds 已提交
1479 1480
		goto fail_vnrele;
	}
1481 1482 1483
	sb->s_root = d_alloc_root(root);
	if (!sb->s_root) {
		error = ENOMEM;
L
Linus Torvalds 已提交
1484 1485
		goto fail_vnrele;
	}
1486

1487 1488
	error = xfs_syncd_init(mp);
	if (error)
L
Linus Torvalds 已提交
1489
		goto fail_vnrele;
1490

1491
	kfree(mtpt);
L
Linus Torvalds 已提交
1492

1493
	xfs_itrace_exit(XFS_I(sb->s_root->d_inode));
L
Linus Torvalds 已提交
1494 1495
	return 0;

1496 1497
 out_filestream_unmount:
	xfs_filestream_unmount(mp);
1498 1499
 out_free_sb:
	xfs_freesb(mp);
1500
 out_destroy_counters:
C
Christoph Hellwig 已提交
1501
	xfs_icsb_destroy_counters(mp);
1502 1503
	xfs_close_devices(mp);
 out_put_qmops:
1504
	xfs_qmops_put(mp);
1505
 out_put_dmops:
1506
	xfs_dmops_put(mp);
1507 1508 1509
 out_free_fsname:
	xfs_free_fsname(mp);
	kfree(mtpt);
C
Christoph Hellwig 已提交
1510
	kfree(mp);
1511
 out:
C
Christoph Hellwig 已提交
1512
	return -error;
1513 1514

 fail_vnrele:
L
Linus Torvalds 已提交
1515 1516 1517 1518
	if (sb->s_root) {
		dput(sb->s_root);
		sb->s_root = NULL;
	} else {
1519
		iput(root);
L
Linus Torvalds 已提交
1520 1521
	}

1522
 fail_unmount:
1523 1524 1525 1526 1527 1528 1529 1530 1531
	/*
	 * Blow away any referenced inode in the filestreams cache.
	 * This can and will cause log traffic as inodes go inactive
	 * here.
	 */
	xfs_filestream_unmount(mp);

	XFS_bflush(mp->m_ddev_targp);

1532
	xfs_unmountfs(mp);
1533
	goto out_free_sb;
L
Linus Torvalds 已提交
1534 1535
}

1536
STATIC int
1537
xfs_fs_get_sb(
L
Linus Torvalds 已提交
1538 1539 1540
	struct file_system_type	*fs_type,
	int			flags,
	const char		*dev_name,
1541 1542
	void			*data,
	struct vfsmount		*mnt)
L
Linus Torvalds 已提交
1543
{
1544 1545
	return get_sb_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super,
			   mnt);
1546 1547
}

1548
static struct super_operations xfs_super_operations = {
1549 1550 1551 1552 1553 1554 1555
	.alloc_inode		= xfs_fs_alloc_inode,
	.destroy_inode		= xfs_fs_destroy_inode,
	.write_inode		= xfs_fs_write_inode,
	.clear_inode		= xfs_fs_clear_inode,
	.put_super		= xfs_fs_put_super,
	.write_super		= xfs_fs_write_super,
	.sync_fs		= xfs_fs_sync_super,
1556
	.freeze_fs		= xfs_fs_freeze,
1557 1558 1559
	.statfs			= xfs_fs_statfs,
	.remount_fs		= xfs_fs_remount,
	.show_options		= xfs_fs_show_options,
L
Linus Torvalds 已提交
1560 1561
};

A
Andrew Morton 已提交
1562
static struct file_system_type xfs_fs_type = {
L
Linus Torvalds 已提交
1563 1564
	.owner			= THIS_MODULE,
	.name			= "xfs",
1565
	.get_sb			= xfs_fs_get_sb,
L
Linus Torvalds 已提交
1566 1567 1568 1569
	.kill_sb		= kill_block_super,
	.fs_flags		= FS_REQUIRES_DEV,
};

1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582
STATIC int __init
xfs_alloc_trace_bufs(void)
{
#ifdef XFS_ALLOC_TRACE
	xfs_alloc_trace_buf = ktrace_alloc(XFS_ALLOC_TRACE_SIZE, KM_MAYFAIL);
	if (!xfs_alloc_trace_buf)
		goto out;
#endif
#ifdef XFS_BMAP_TRACE
	xfs_bmap_trace_buf = ktrace_alloc(XFS_BMAP_TRACE_SIZE, KM_MAYFAIL);
	if (!xfs_bmap_trace_buf)
		goto out_free_alloc_trace;
#endif
1583 1584 1585 1586 1587 1588 1589 1590 1591 1592
#ifdef XFS_BTREE_TRACE
	xfs_allocbt_trace_buf = ktrace_alloc(XFS_ALLOCBT_TRACE_SIZE,
					     KM_MAYFAIL);
	if (!xfs_allocbt_trace_buf)
		goto out_free_bmap_trace;

	xfs_inobt_trace_buf = ktrace_alloc(XFS_INOBT_TRACE_SIZE, KM_MAYFAIL);
	if (!xfs_inobt_trace_buf)
		goto out_free_allocbt_trace;

1593 1594
	xfs_bmbt_trace_buf = ktrace_alloc(XFS_BMBT_TRACE_SIZE, KM_MAYFAIL);
	if (!xfs_bmbt_trace_buf)
1595
		goto out_free_inobt_trace;
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616
#endif
#ifdef XFS_ATTR_TRACE
	xfs_attr_trace_buf = ktrace_alloc(XFS_ATTR_TRACE_SIZE, KM_MAYFAIL);
	if (!xfs_attr_trace_buf)
		goto out_free_bmbt_trace;
#endif
#ifdef XFS_DIR2_TRACE
	xfs_dir2_trace_buf = ktrace_alloc(XFS_DIR2_GTRACE_SIZE, KM_MAYFAIL);
	if (!xfs_dir2_trace_buf)
		goto out_free_attr_trace;
#endif

	return 0;

#ifdef XFS_DIR2_TRACE
 out_free_attr_trace:
#endif
#ifdef XFS_ATTR_TRACE
	ktrace_free(xfs_attr_trace_buf);
 out_free_bmbt_trace:
#endif
1617
#ifdef XFS_BTREE_TRACE
1618
	ktrace_free(xfs_bmbt_trace_buf);
1619 1620 1621 1622
 out_free_inobt_trace:
	ktrace_free(xfs_inobt_trace_buf);
 out_free_allocbt_trace:
	ktrace_free(xfs_allocbt_trace_buf);
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
 out_free_bmap_trace:
#endif
#ifdef XFS_BMAP_TRACE
	ktrace_free(xfs_bmap_trace_buf);
 out_free_alloc_trace:
#endif
#ifdef XFS_ALLOC_TRACE
	ktrace_free(xfs_alloc_trace_buf);
 out:
#endif
	return -ENOMEM;
}

STATIC void
xfs_free_trace_bufs(void)
{
#ifdef XFS_DIR2_TRACE
	ktrace_free(xfs_dir2_trace_buf);
#endif
#ifdef XFS_ATTR_TRACE
	ktrace_free(xfs_attr_trace_buf);
#endif
1645
#ifdef XFS_BTREE_TRACE
1646
	ktrace_free(xfs_bmbt_trace_buf);
1647 1648
	ktrace_free(xfs_inobt_trace_buf);
	ktrace_free(xfs_allocbt_trace_buf);
1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
#endif
#ifdef XFS_BMAP_TRACE
	ktrace_free(xfs_bmap_trace_buf);
#endif
#ifdef XFS_ALLOC_TRACE
	ktrace_free(xfs_alloc_trace_buf);
#endif
}

STATIC int __init
xfs_init_zones(void)
{

	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
	if (!xfs_ioend_zone)
1664
		goto out;
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679

	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;
1680

1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
	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_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
	if (!xfs_dabuf_zone)
		goto out_destroy_da_state_zone;

	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
	if (!xfs_ifork_zone)
		goto out_destroy_dabuf_zone;

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

	/*
	 * 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.
	 */
	xfs_buf_item_zone = kmem_zone_init((sizeof(xfs_buf_log_item_t) +
				(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
				  NBWORD) * sizeof(int))), "xfs_buf_item");
	if (!xfs_buf_item_zone)
		goto out_destroy_trans_zone;

	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",
1728 1729
			KM_ZONE_HWALIGN | KM_ZONE_RECLAIM | KM_ZONE_SPREAD,
			xfs_fs_inode_init_once);
1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802
	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;

#ifdef CONFIG_XFS_POSIX_ACL
	xfs_acl_zone = kmem_zone_init(sizeof(xfs_acl_t), "xfs_acl");
	if (!xfs_acl_zone)
		goto out_destroy_ili_zone;
#endif

	return 0;

#ifdef CONFIG_XFS_POSIX_ACL
 out_destroy_ili_zone:
#endif
	kmem_zone_destroy(xfs_ili_zone);
 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);
 out_destroy_trans_zone:
	kmem_zone_destroy(xfs_trans_zone);
 out_destroy_ifork_zone:
	kmem_zone_destroy(xfs_ifork_zone);
 out_destroy_dabuf_zone:
	kmem_zone_destroy(xfs_dabuf_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)
{
#ifdef CONFIG_XFS_POSIX_ACL
	kmem_zone_destroy(xfs_acl_zone);
#endif
	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);
	kmem_zone_destroy(xfs_trans_zone);
	kmem_zone_destroy(xfs_ifork_zone);
	kmem_zone_destroy(xfs_dabuf_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
1803 1804

STATIC int __init
1805
init_xfs_fs(void)
L
Linus Torvalds 已提交
1806 1807 1808
{
	int			error;

1809 1810
	printk(KERN_INFO XFS_VERSION_STRING " with "
			 XFS_BUILD_OPTIONS " enabled\n");
L
Linus Torvalds 已提交
1811 1812

	ktrace_init(64);
1813
	xfs_ioend_init();
1814
	xfs_dir_startup();
L
Linus Torvalds 已提交
1815

1816
	error = xfs_init_zones();
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
	if (error)
		goto out;

	error = xfs_alloc_trace_bufs();
	if (error)
		goto out_destroy_zones;

	error = xfs_mru_cache_init();
	if (error)
		goto out_free_trace_buffers;

	error = xfs_filestream_init();
	if (error)
		goto out_mru_cache_uninit;
L
Linus Torvalds 已提交
1831

1832
	error = xfs_buf_init();
1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
	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
1843 1844 1845 1846 1847

	vfs_initquota();

	error = register_filesystem(&xfs_fs_type);
	if (error)
1848
		goto out_sysctl_unregister;
L
Linus Torvalds 已提交
1849 1850
	return 0;

1851 1852 1853 1854 1855
 out_sysctl_unregister:
	xfs_sysctl_unregister();
 out_cleanup_procfs:
	xfs_cleanup_procfs();
 out_buf_terminate:
1856
	xfs_buf_terminate();
1857 1858 1859 1860 1861 1862 1863
 out_filestream_uninit:
	xfs_filestream_uninit();
 out_mru_cache_uninit:
	xfs_mru_cache_uninit();
 out_free_trace_buffers:
	xfs_free_trace_bufs();
 out_destroy_zones:
1864
	xfs_destroy_zones();
1865
 out:
L
Linus Torvalds 已提交
1866 1867 1868 1869
	return error;
}

STATIC void __exit
1870
exit_xfs_fs(void)
L
Linus Torvalds 已提交
1871 1872 1873
{
	vfs_exitquota();
	unregister_filesystem(&xfs_fs_type);
1874 1875
	xfs_sysctl_unregister();
	xfs_cleanup_procfs();
1876
	xfs_buf_terminate();
1877 1878 1879
	xfs_filestream_uninit();
	xfs_mru_cache_uninit();
	xfs_free_trace_bufs();
1880
	xfs_destroy_zones();
L
Linus Torvalds 已提交
1881 1882 1883 1884 1885 1886 1887 1888 1889
	ktrace_uninit();
}

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");