xfs_super.c 38.1 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 21
#include "xfs_log.h"
#include "xfs_clnt.h"
22
#include "xfs_inum.h"
L
Linus Torvalds 已提交
23 24
#include "xfs_trans.h"
#include "xfs_sb.h"
25
#include "xfs_ag.h"
L
Linus Torvalds 已提交
26 27 28 29 30 31
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
32
#include "xfs_alloc_btree.h"
L
Linus Torvalds 已提交
33 34
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
35
#include "xfs_attr_sf.h"
L
Linus Torvalds 已提交
36 37
#include "xfs_dinode.h"
#include "xfs_inode.h"
38 39
#include "xfs_btree.h"
#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"
51
#include "xfs_vfsops.h"
L
Linus Torvalds 已提交
52
#include "xfs_version.h"
53
#include "xfs_log_priv.h"
54
#include "xfs_trans_priv.h"
L
Linus Torvalds 已提交
55 56 57 58

#include <linux/namei.h>
#include <linux/init.h>
#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>
L
Linus Torvalds 已提交
63

64 65 66 67
static struct quotactl_ops xfs_quotactl_operations;
static struct super_operations xfs_super_operations;
static kmem_zone_t *xfs_vnode_zone;
static kmem_zone_t *xfs_ioend_zone;
68
mempool_t *xfs_ioend_pool;
L
Linus Torvalds 已提交
69 70 71

STATIC struct xfs_mount_args *
xfs_args_allocate(
72 73
	struct super_block	*sb,
	int			silent)
L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82 83 84 85
{
	struct xfs_mount_args	*args;

	args = kmem_zalloc(sizeof(struct xfs_mount_args), KM_SLEEP);
	args->logbufs = args->logbufsize = -1;
	strncpy(args->fsname, sb->s_id, MAXNAMELEN);

	/* Copy the already-parsed mount(2) flags we're interested in */
	if (sb->s_flags & MS_DIRSYNC)
		args->flags |= XFSMNT_DIRSYNC;
	if (sb->s_flags & MS_SYNCHRONOUS)
		args->flags |= XFSMNT_WSYNC;
86 87
	if (silent)
		args->flags |= XFSMNT_QUIET;
L
Linus Torvalds 已提交
88 89 90 91 92
	args->flags |= XFSMNT_32BITINODES;

	return args;
}

93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
#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) */

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;
}

STATIC int
xfs_parseargs(
	struct xfs_mount	*mp,
	char			*options,
	struct xfs_mount_args	*args,
	int			update)
{
	char			*this_char, *value, *eov;
	int			dsunit, dswidth, vol_dsunit, vol_dswidth;
	int			iosize;
174
	int			dmapi_implies_ikeep = 1;
175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304

	args->flags |= XFSMNT_BARRIER;
	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;

	if (!options)
		goto done;

	iosize = dsunit = dswidth = vol_dsunit = vol_dswidth = 0;

	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;
			}
			args->logbufs = simple_strtoul(value, &eov, 10);
		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			args->logbufsize = suffix_strtoul(value, &eov, 10);
		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			strncpy(args->logname, value, MAXNAMELEN);
		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			strncpy(args->mtpt, value, MAXNAMELEN);
		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
			if (!value || !*value) {
				cmn_err(CE_WARN,
					"XFS: %s option requires an argument",
					this_char);
				return EINVAL;
			}
			strncpy(args->rtname, value, MAXNAMELEN);
		} 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);
			args->flags |= XFSMNT_IOSIZE;
			args->iosizelog = (uint8_t) iosize;
		} 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);
			args->flags |= XFSMNT_IOSIZE;
			args->iosizelog = ffs(iosize) - 1;
		} 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)) {
			args->flags |= XFSMNT_WSYNC;
		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
			args->flags |= XFSMNT_OSYNCISOSYNC;
		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
			args->flags |= XFSMNT_NORECOVERY;
		} else if (!strcmp(this_char, MNTOPT_INO64)) {
			args->flags |= XFSMNT_INO64;
#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_NOALIGN)) {
			args->flags |= XFSMNT_NOALIGN;
		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
			args->flags |= XFSMNT_SWALLOC;
		} 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)) {
			args->flags &= ~XFSMNT_32BITINODES;
#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)) {
			args->flags |= XFSMNT_NOUUID;
		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
			args->flags |= XFSMNT_BARRIER;
		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
			args->flags &= ~XFSMNT_BARRIER;
		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
305
			args->flags |= XFSMNT_IKEEP;
306
		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
307 308
			dmapi_implies_ikeep = 0;
			args->flags &= ~XFSMNT_IKEEP;
309 310 311 312 313 314 315 316
		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
			args->flags2 &= ~XFSMNT2_COMPAT_IOSIZE;
		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
			args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
			args->flags |= XFSMNT_ATTR2;
		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
			args->flags &= ~XFSMNT_ATTR2;
317
			args->flags |= XFSMNT_NOATTR2;
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
			args->flags2 |= XFSMNT2_FILESTREAMS;
		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
			args->flags &= ~(XFSMNT_UQUOTAENF|XFSMNT_UQUOTA);
			args->flags &= ~(XFSMNT_GQUOTAENF|XFSMNT_GQUOTA);
		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
			   !strcmp(this_char, MNTOPT_UQUOTA) ||
			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
			args->flags |= XFSMNT_UQUOTA | XFSMNT_UQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
			args->flags |= XFSMNT_UQUOTA;
			args->flags &= ~XFSMNT_UQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
			args->flags |= XFSMNT_PQUOTA | XFSMNT_PQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
			args->flags |= XFSMNT_PQUOTA;
			args->flags &= ~XFSMNT_PQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
			args->flags |= XFSMNT_GQUOTA | XFSMNT_GQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
			args->flags |= XFSMNT_GQUOTA;
			args->flags &= ~XFSMNT_GQUOTAENF;
		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
			args->flags |= XFSMNT_DMAPI;
		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
			args->flags |= XFSMNT_DMAPI;
		} else if (!strcmp(this_char, MNTOPT_DMI)) {
			args->flags |= XFSMNT_DMAPI;
		} 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;
		}
	}

	if (args->flags & XFSMNT_NORECOVERY) {
		if ((mp->m_flags & XFS_MOUNT_RDONLY) == 0) {
			cmn_err(CE_WARN,
				"XFS: no-recovery mounts must be read-only.");
			return EINVAL;
		}
	}

	if ((args->flags & XFSMNT_NOALIGN) && (dsunit || dswidth)) {
		cmn_err(CE_WARN,
	"XFS: sunit and swidth options incompatible with the noalign option");
		return EINVAL;
	}

	if ((args->flags & XFSMNT_GQUOTA) && (args->flags & XFSMNT_PQUOTA)) {
		cmn_err(CE_WARN,
			"XFS: cannot mount with both project and group quota");
		return EINVAL;
	}

	if ((args->flags & XFSMNT_DMAPI) && *args->mtpt == '\0') {
		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.
	 */
414 415
	if ((args->flags & XFSMNT_DMAPI) && dmapi_implies_ikeep)
		args->flags |= XFSMNT_IKEEP;
416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449

	if ((args->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
		if (dsunit) {
			args->sunit = dsunit;
			args->flags |= XFSMNT_RETERR;
		} else {
			args->sunit = vol_dsunit;
		}
		dswidth ? (args->swidth = dswidth) :
			  (args->swidth = vol_dswidth);
	} else {
		args->sunit = args->swidth = 0;
	}

done:
	if (args->flags & XFSMNT_32BITINODES)
		mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
	if (args->flags2)
		args->flags |= XFSMNT_FLAGS2;
	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 */
450
		{ XFS_MOUNT_IKEEP,		"," MNTOPT_IKEEP },
451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
		{ 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
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
__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;
}

557
STATIC_INLINE void
L
Linus Torvalds 已提交
558 559 560
xfs_set_inodeops(
	struct inode		*inode)
{
561 562
	switch (inode->i_mode & S_IFMT) {
	case S_IFREG:
563
		inode->i_op = &xfs_inode_operations;
564
		inode->i_fop = &xfs_file_operations;
565
		inode->i_mapping->a_ops = &xfs_address_space_operations;
566 567
		break;
	case S_IFDIR:
568
		inode->i_op = &xfs_dir_inode_operations;
569
		inode->i_fop = &xfs_dir_file_operations;
570 571
		break;
	case S_IFLNK:
572
		inode->i_op = &xfs_symlink_inode_operations;
573
		if (!(XFS_I(inode)->i_df.if_flags & XFS_IFINLINE))
574
			inode->i_mapping->a_ops = &xfs_address_space_operations;
575 576
		break;
	default:
577
		inode->i_op = &xfs_inode_operations;
L
Linus Torvalds 已提交
578
		init_special_inode(inode, inode->i_mode, inode->i_rdev);
579
		break;
L
Linus Torvalds 已提交
580 581 582
	}
}

583
STATIC_INLINE void
L
Linus Torvalds 已提交
584 585
xfs_revalidate_inode(
	xfs_mount_t		*mp,
586
	bhv_vnode_t		*vp,
L
Linus Torvalds 已提交
587 588
	xfs_inode_t		*ip)
{
589
	struct inode		*inode = vn_to_inode(vp);
L
Linus Torvalds 已提交
590

591
	inode->i_mode	= ip->i_d.di_mode;
L
Linus Torvalds 已提交
592 593 594
	inode->i_nlink	= ip->i_d.di_nlink;
	inode->i_uid	= ip->i_d.di_uid;
	inode->i_gid	= ip->i_d.di_gid;
595 596 597 598 599 600 601 602 603

	switch (inode->i_mode & S_IFMT) {
	case S_IFBLK:
	case S_IFCHR:
		inode->i_rdev =
			MKDEV(sysv_major(ip->i_df.if_u2.if_rdev) & 0x1ff,
			      sysv_minor(ip->i_df.if_u2.if_rdev));
		break;
	default:
L
Linus Torvalds 已提交
604
		inode->i_rdev = 0;
605
		break;
L
Linus Torvalds 已提交
606
	}
607

L
Linus Torvalds 已提交
608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631
	inode->i_generation = ip->i_d.di_gen;
	i_size_write(inode, ip->i_d.di_size);
	inode->i_atime.tv_sec	= ip->i_d.di_atime.t_sec;
	inode->i_atime.tv_nsec	= ip->i_d.di_atime.t_nsec;
	inode->i_mtime.tv_sec	= ip->i_d.di_mtime.t_sec;
	inode->i_mtime.tv_nsec	= ip->i_d.di_mtime.t_nsec;
	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
	if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE)
		inode->i_flags |= S_IMMUTABLE;
	else
		inode->i_flags &= ~S_IMMUTABLE;
	if (ip->i_d.di_flags & XFS_DIFLAG_APPEND)
		inode->i_flags |= S_APPEND;
	else
		inode->i_flags &= ~S_APPEND;
	if (ip->i_d.di_flags & XFS_DIFLAG_SYNC)
		inode->i_flags |= S_SYNC;
	else
		inode->i_flags &= ~S_SYNC;
	if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME)
		inode->i_flags |= S_NOATIME;
	else
		inode->i_flags &= ~S_NOATIME;
632
	xfs_iflags_clear(ip, XFS_IMODIFIED);
L
Linus Torvalds 已提交
633 634 635 636
}

void
xfs_initialize_vnode(
637
	struct xfs_mount	*mp,
638
	bhv_vnode_t		*vp,
639
	struct xfs_inode	*ip)
L
Linus Torvalds 已提交
640
{
641
	struct inode		*inode = vn_to_inode(vp);
L
Linus Torvalds 已提交
642

643 644 645
	if (!ip->i_vnode) {
		ip->i_vnode = vp;
		inode->i_private = ip;
L
Linus Torvalds 已提交
646 647 648 649 650 651 652 653 654
	}

	/*
	 * We need to set the ops vectors, and unlock the inode, but if
	 * we have been called during the new inode create process, it is
	 * too early to fill in the Linux inode.  We will get called a
	 * second time once the inode is properly set up, and then we can
	 * finish our work.
	 */
655
	if (ip->i_d.di_mode != 0 && (inode->i_state & I_NEW)) {
656
		xfs_revalidate_inode(mp, vp, ip);
L
Linus Torvalds 已提交
657
		xfs_set_inodeops(inode);
658

659
		xfs_iflags_clear(ip, XFS_INEW);
L
Linus Torvalds 已提交
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
		barrier();

		unlock_new_inode(inode);
	}
}

int
xfs_blkdev_get(
	xfs_mount_t		*mp,
	const char		*name,
	struct block_device	**bdevp)
{
	int			error = 0;

	*bdevp = open_bdev_excl(name, 0, mp);
	if (IS_ERR(*bdevp)) {
		error = PTR_ERR(*bdevp);
		printk("XFS: Invalid device [%s], error=%d\n", name, error);
	}

	return -error;
}

void
xfs_blkdev_put(
	struct block_device	*bdev)
{
	if (bdev)
		close_bdev_excl(bdev);
}

691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
/*
 * 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;
734
		return;
735 736
	}

737
	if (mp->m_ddev_targp->bt_bdev->bd_disk->queue->ordered ==
738 739 740 741
					QUEUE_ORDERED_NONE) {
		xfs_fs_cmn_err(CE_NOTE, mp,
		  "Disabling barriers, not supported by the underlying device");
		mp->m_flags &= ~XFS_MOUNT_BARRIER;
742
		return;
743 744
	}

745 746 747 748 749 750 751
	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;
	}

752 753 754 755 756
	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;
757
		return;
758 759 760 761 762 763 764
	}
}

void
xfs_blkdev_issue_flush(
	xfs_buftarg_t		*buftarg)
{
765
	blkdev_issue_flush(buftarg->bt_bdev, NULL);
766
}
L
Linus Torvalds 已提交
767

768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
/*
 * XFS AIL push thread support
 */
void
xfsaild_wakeup(
	xfs_mount_t		*mp,
	xfs_lsn_t		threshold_lsn)
{
	mp->m_ail.xa_target = threshold_lsn;
	wake_up_process(mp->m_ail.xa_task);
}

int
xfsaild(
	void	*data)
{
	xfs_mount_t	*mp = (xfs_mount_t *)data;
	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();

		ASSERT(mp->m_log);
		if (XFS_FORCED_SHUTDOWN(mp))
			continue;

		tout = xfsaild_push(mp, &last_pushed_lsn);
	}

	return 0;
}	/* xfsaild */

int
xfsaild_start(
	xfs_mount_t	*mp)
{
	mp->m_ail.xa_target = 0;
	mp->m_ail.xa_task = kthread_run(xfsaild, mp, "xfsaild");
	if (IS_ERR(mp->m_ail.xa_task))
		return -PTR_ERR(mp->m_ail.xa_task);
	return 0;
}

void
xfsaild_stop(
	xfs_mount_t	*mp)
{
	kthread_stop(mp->m_ail.xa_task);
}



L
Linus Torvalds 已提交
826
STATIC struct inode *
827
xfs_fs_alloc_inode(
L
Linus Torvalds 已提交
828 829
	struct super_block	*sb)
{
830
	bhv_vnode_t		*vp;
L
Linus Torvalds 已提交
831

832 833
	vp = kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
	if (unlikely(!vp))
L
Linus Torvalds 已提交
834
		return NULL;
835
	return vn_to_inode(vp);
L
Linus Torvalds 已提交
836 837 838
}

STATIC void
839
xfs_fs_destroy_inode(
L
Linus Torvalds 已提交
840 841
	struct inode		*inode)
{
842
	kmem_zone_free(xfs_vnode_zone, vn_from_inode(inode));
L
Linus Torvalds 已提交
843 844 845
}

STATIC void
846
xfs_fs_inode_init_once(
847
	void			*vnode)
L
Linus Torvalds 已提交
848
{
C
Christoph Lameter 已提交
849
	inode_init_once(vn_to_inode((bhv_vnode_t *)vnode));
L
Linus Torvalds 已提交
850 851
}

852
STATIC int __init
853
xfs_init_zones(void)
L
Linus Torvalds 已提交
854
{
855
	xfs_vnode_zone = kmem_zone_init_flags(sizeof(bhv_vnode_t), "xfs_vnode",
856 857
					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
					KM_ZONE_SPREAD,
858
					xfs_fs_inode_init_once);
859 860 861 862 863 864 865
	if (!xfs_vnode_zone)
		goto out;

	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
	if (!xfs_ioend_zone)
		goto out_destroy_vnode_zone;

866 867
	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
						  xfs_ioend_zone);
868 869
	if (!xfs_ioend_pool)
		goto out_free_ioend_zone;
L
Linus Torvalds 已提交
870
	return 0;
871 872 873 874 875 876 877

 out_free_ioend_zone:
	kmem_zone_destroy(xfs_ioend_zone);
 out_destroy_vnode_zone:
	kmem_zone_destroy(xfs_vnode_zone);
 out:
	return -ENOMEM;
L
Linus Torvalds 已提交
878 879 880
}

STATIC void
881
xfs_destroy_zones(void)
L
Linus Torvalds 已提交
882
{
883 884 885
	mempool_destroy(xfs_ioend_pool);
	kmem_zone_destroy(xfs_vnode_zone);
	kmem_zone_destroy(xfs_ioend_zone);
L
Linus Torvalds 已提交
886 887 888 889 890 891
}

/*
 * 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,
892
 * since this is when the inode itself becomes flushable.
L
Linus Torvalds 已提交
893 894
 */
STATIC int
895
xfs_fs_write_inode(
L
Linus Torvalds 已提交
896 897 898
	struct inode		*inode,
	int			sync)
{
899 900
	int			error = 0;
	int			flags = 0;
L
Linus Torvalds 已提交
901

902
	xfs_itrace_entry(XFS_I(inode));
903 904 905
	if (sync) {
		filemap_fdatawait(inode->i_mapping);
		flags |= FLUSH_SYNC;
L
Linus Torvalds 已提交
906
	}
907
	error = xfs_inode_flush(XFS_I(inode), flags);
908 909 910 911 912 913
	/*
	 * if we failed to write out the inode then mark
	 * it dirty again so we'll try again later.
	 */
	if (error)
		mark_inode_dirty_sync(inode);
914

L
Linus Torvalds 已提交
915 916 917 918
	return -error;
}

STATIC void
919
xfs_fs_clear_inode(
L
Linus Torvalds 已提交
920 921
	struct inode		*inode)
{
922
	xfs_inode_t		*ip = XFS_I(inode);
923

924
	/*
925
	 * ip can be null when xfs_iget_core calls xfs_idestroy if we
926 927
	 * find an inode with di_mode == 0 but without IGET_CREATE set.
	 */
928
	if (ip) {
929
		xfs_itrace_entry(ip);
930 931 932 933 934 935 936 937
		XFS_STATS_INC(vn_rele);
		XFS_STATS_INC(vn_remove);
		XFS_STATS_INC(vn_reclaim);
		XFS_STATS_DEC(vn_active);

		xfs_inactive(ip);
		xfs_iflags_clear(ip, XFS_IMODIFIED);
		if (xfs_reclaim(ip))
938
			panic("%s: cannot reclaim 0x%p\n", __func__, inode);
939
	}
940

941
	ASSERT(XFS_I(inode) == NULL);
942
}
L
Linus Torvalds 已提交
943 944 945 946 947 948 949 950 951 952

/*
 * Enqueue a work item to be picked up by the vfs xfssyncd thread.
 * Doing this has two advantages:
 * - It saves on stack space, which is tight in certain situations
 * - It can be used (with care) as a mechanism to avoid deadlocks.
 * Flushing while allocating in a full filesystem requires both.
 */
STATIC void
xfs_syncd_queue_work(
953
	struct xfs_mount *mp,
L
Linus Torvalds 已提交
954
	void		*data,
955
	void		(*syncer)(struct xfs_mount *, void *))
L
Linus Torvalds 已提交
956
{
957
	struct bhv_vfs_sync_work *work;
L
Linus Torvalds 已提交
958

959
	work = kmem_alloc(sizeof(struct bhv_vfs_sync_work), KM_SLEEP);
L
Linus Torvalds 已提交
960 961 962
	INIT_LIST_HEAD(&work->w_list);
	work->w_syncer = syncer;
	work->w_data = data;
963 964 965 966 967
	work->w_mount = mp;
	spin_lock(&mp->m_sync_lock);
	list_add_tail(&work->w_list, &mp->m_sync_list);
	spin_unlock(&mp->m_sync_lock);
	wake_up_process(mp->m_sync_task);
L
Linus Torvalds 已提交
968 969 970 971 972 973 974 975 976 977
}

/*
 * Flush delayed allocate data, attempting to free up reserved space
 * from existing allocations.  At this point a new allocation attempt
 * has failed with ENOSPC and we are in the process of scratching our
 * heads, looking about for more room...
 */
STATIC void
xfs_flush_inode_work(
978 979
	struct xfs_mount *mp,
	void		*arg)
L
Linus Torvalds 已提交
980
{
981 982 983
	struct inode	*inode = arg;
	filemap_flush(inode->i_mapping);
	iput(inode);
L
Linus Torvalds 已提交
984 985 986 987 988 989
}

void
xfs_flush_inode(
	xfs_inode_t	*ip)
{
990
	struct inode	*inode = ip->i_vnode;
L
Linus Torvalds 已提交
991 992

	igrab(inode);
993
	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
994
	delay(msecs_to_jiffies(500));
L
Linus Torvalds 已提交
995 996 997 998 999 1000 1001 1002
}

/*
 * This is the "bigger hammer" version of xfs_flush_inode_work...
 * (IOW, "If at first you don't succeed, use a Bigger Hammer").
 */
STATIC void
xfs_flush_device_work(
1003 1004
	struct xfs_mount *mp,
	void		*arg)
L
Linus Torvalds 已提交
1005
{
1006
	struct inode	*inode = arg;
C
Christoph Hellwig 已提交
1007
	sync_blockdev(mp->m_super->s_bdev);
1008
	iput(inode);
L
Linus Torvalds 已提交
1009 1010 1011 1012 1013 1014
}

void
xfs_flush_device(
	xfs_inode_t	*ip)
{
1015
	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
L
Linus Torvalds 已提交
1016 1017

	igrab(inode);
1018
	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
1019
	delay(msecs_to_jiffies(500));
L
Linus Torvalds 已提交
1020 1021 1022 1023
	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
}

STATIC void
1024 1025
xfs_sync_worker(
	struct xfs_mount *mp,
L
Linus Torvalds 已提交
1026 1027 1028 1029
	void		*unused)
{
	int		error;

1030
	if (!(mp->m_flags & XFS_MOUNT_RDONLY))
1031
		error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR);
1032 1033
	mp->m_sync_seq++;
	wake_up(&mp->m_wait_single_sync_task);
L
Linus Torvalds 已提交
1034 1035 1036 1037 1038 1039
}

STATIC int
xfssyncd(
	void			*arg)
{
1040
	struct xfs_mount	*mp = arg;
L
Linus Torvalds 已提交
1041
	long			timeleft;
1042
	bhv_vfs_sync_work_t	*work, *n;
1043
	LIST_HEAD		(tmp);
L
Linus Torvalds 已提交
1044

1045
	set_freezable();
1046
	timeleft = xfs_syncd_centisecs * msecs_to_jiffies(10);
L
Linus Torvalds 已提交
1047
	for (;;) {
1048
		timeleft = schedule_timeout_interruptible(timeleft);
L
Linus Torvalds 已提交
1049
		/* swsusp */
1050
		try_to_freeze();
1051
		if (kthread_should_stop() && list_empty(&mp->m_sync_list))
L
Linus Torvalds 已提交
1052 1053
			break;

1054
		spin_lock(&mp->m_sync_lock);
L
Linus Torvalds 已提交
1055 1056 1057 1058 1059
		/*
		 * We can get woken by laptop mode, to do a sync -
		 * that's the (only!) case where the list would be
		 * empty with time remaining.
		 */
1060
		if (!timeleft || list_empty(&mp->m_sync_list)) {
L
Linus Torvalds 已提交
1061
			if (!timeleft)
1062 1063
				timeleft = xfs_syncd_centisecs *
							msecs_to_jiffies(10);
1064 1065 1066
			INIT_LIST_HEAD(&mp->m_sync_work.w_list);
			list_add_tail(&mp->m_sync_work.w_list,
					&mp->m_sync_list);
L
Linus Torvalds 已提交
1067
		}
1068
		list_for_each_entry_safe(work, n, &mp->m_sync_list, w_list)
L
Linus Torvalds 已提交
1069
			list_move(&work->w_list, &tmp);
1070
		spin_unlock(&mp->m_sync_lock);
L
Linus Torvalds 已提交
1071 1072

		list_for_each_entry_safe(work, n, &tmp, w_list) {
1073
			(*work->w_syncer)(mp, work->w_data);
L
Linus Torvalds 已提交
1074
			list_del(&work->w_list);
1075
			if (work == &mp->m_sync_work)
L
Linus Torvalds 已提交
1076
				continue;
1077
			kmem_free(work);
L
Linus Torvalds 已提交
1078 1079 1080 1081 1082 1083 1084
		}
	}

	return 0;
}

STATIC void
1085
xfs_fs_put_super(
L
Linus Torvalds 已提交
1086 1087
	struct super_block	*sb)
{
1088
	struct xfs_mount	*mp = XFS_M(sb);
L
Linus Torvalds 已提交
1089 1090
	int			error;

1091 1092
	kthread_stop(mp->m_sync_task);

1093 1094
	xfs_sync(mp, SYNC_ATTR | SYNC_DELWRI);
	error = xfs_unmount(mp, 0, NULL);
C
Christoph Hellwig 已提交
1095
	if (error)
1096
		printk("XFS: unmount got error=%d\n", error);
L
Linus Torvalds 已提交
1097 1098 1099
}

STATIC void
1100
xfs_fs_write_super(
L
Linus Torvalds 已提交
1101 1102
	struct super_block	*sb)
{
1103
	if (!(sb->s_flags & MS_RDONLY))
1104
		xfs_sync(XFS_M(sb), SYNC_FSDATA);
L
Linus Torvalds 已提交
1105 1106 1107 1108
	sb->s_dirt = 0;
}

STATIC int
1109
xfs_fs_sync_super(
L
Linus Torvalds 已提交
1110 1111 1112
	struct super_block	*sb,
	int			wait)
{
1113
	struct xfs_mount	*mp = XFS_M(sb);
1114 1115
	int			error;
	int			flags;
L
Linus Torvalds 已提交
1116

1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129
	/*
	 * 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.
	 */
	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE)) {
1130 1131 1132 1133 1134 1135 1136 1137
		/*
		 * First stage of freeze - no more writers will make progress
		 * now we are here, so we flush delwri and delalloc buffers
		 * here, then wait for all I/O to complete.  Data is frozen at
		 * that point. Metadata is not frozen, transactions can still
		 * occur here so don't bother flushing the buftarg (i.e
		 * SYNC_QUIESCE) because it'll just get dirty again.
		 */
1138
		flags = SYNC_DATA_QUIESCE;
1139
	} else
1140
		flags = SYNC_FSDATA;
L
Linus Torvalds 已提交
1141

1142
	error = xfs_sync(mp, flags);
L
Linus Torvalds 已提交
1143 1144 1145
	sb->s_dirt = 0;

	if (unlikely(laptop_mode)) {
1146
		int	prev_sync_seq = mp->m_sync_seq;
L
Linus Torvalds 已提交
1147 1148 1149 1150 1151 1152

		/*
		 * 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).
		 */
1153
		wake_up_process(mp->m_sync_task);
L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159
		/*
		 * 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.
		 */
1160 1161
		wait_event(mp->m_wait_single_sync_task,
				mp->m_sync_seq != prev_sync_seq);
L
Linus Torvalds 已提交
1162 1163 1164 1165 1166 1167
	}

	return -error;
}

STATIC int
1168
xfs_fs_statfs(
1169
	struct dentry		*dentry,
L
Linus Torvalds 已提交
1170 1171
	struct kstatfs		*statp)
{
C
Christoph Hellwig 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
	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);

1184
	xfs_icsb_sync_counters(mp, XFS_ICSB_LAZY_COUNT);
C
Christoph Hellwig 已提交
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209

	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
1210 1211 1212
}

STATIC int
1213
xfs_fs_remount(
L
Linus Torvalds 已提交
1214 1215 1216 1217
	struct super_block	*sb,
	int			*flags,
	char			*options)
{
1218
	struct xfs_mount	*mp = XFS_M(sb);
1219
	struct xfs_mount_args	*args = xfs_args_allocate(sb, 0);
L
Linus Torvalds 已提交
1220 1221
	int			error;

1222
	error = xfs_parseargs(mp, options, args, 1);
L
Linus Torvalds 已提交
1223
	if (!error)
1224
		error = xfs_mntupdate(mp, flags, args);
1225
	kmem_free(args);
L
Linus Torvalds 已提交
1226 1227 1228
	return -error;
}

C
Christoph Hellwig 已提交
1229 1230 1231 1232 1233
/*
 * Second stage of a freeze. The data is already frozen so we only
 * need to take care of themetadata. Once that's done write a dummy
 * record to dirty the log in case of a crash while frozen.
 */
L
Linus Torvalds 已提交
1234
STATIC void
1235
xfs_fs_lockfs(
L
Linus Torvalds 已提交
1236 1237
	struct super_block	*sb)
{
C
Christoph Hellwig 已提交
1238 1239 1240 1241
	struct xfs_mount	*mp = XFS_M(sb);

	xfs_attr_quiesce(mp);
	xfs_fs_log_dummy(mp);
L
Linus Torvalds 已提交
1242 1243 1244
}

STATIC int
1245
xfs_fs_show_options(
L
Linus Torvalds 已提交
1246 1247 1248
	struct seq_file		*m,
	struct vfsmount		*mnt)
{
1249
	return -xfs_showargs(XFS_M(mnt->mnt_sb), m);
L
Linus Torvalds 已提交
1250 1251
}

1252
STATIC int
1253
xfs_fs_quotasync(
1254 1255 1256
	struct super_block	*sb,
	int			type)
{
1257
	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XQUOTASYNC, 0, NULL);
1258 1259
}

L
Linus Torvalds 已提交
1260
STATIC int
1261
xfs_fs_getxstate(
L
Linus Torvalds 已提交
1262 1263 1264
	struct super_block	*sb,
	struct fs_quota_stat	*fqs)
{
1265
	return -XFS_QM_QUOTACTL(XFS_M(sb), Q_XGETQSTAT, 0, (caddr_t)fqs);
L
Linus Torvalds 已提交
1266 1267 1268
}

STATIC int
1269
xfs_fs_setxstate(
L
Linus Torvalds 已提交
1270 1271 1272 1273
	struct super_block	*sb,
	unsigned int		flags,
	int			op)
{
1274
	return -XFS_QM_QUOTACTL(XFS_M(sb), op, 0, (caddr_t)&flags);
L
Linus Torvalds 已提交
1275 1276 1277
}

STATIC int
1278
xfs_fs_getxquota(
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283
	struct super_block	*sb,
	int			type,
	qid_t			id,
	struct fs_disk_quota	*fdq)
{
1284
	return -XFS_QM_QUOTACTL(XFS_M(sb),
1285 1286 1287
				 (type == USRQUOTA) ? Q_XGETQUOTA :
				  ((type == GRPQUOTA) ? Q_XGETGQUOTA :
				   Q_XGETPQUOTA), id, (caddr_t)fdq);
L
Linus Torvalds 已提交
1288 1289 1290
}

STATIC int
1291
xfs_fs_setxquota(
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296
	struct super_block	*sb,
	int			type,
	qid_t			id,
	struct fs_disk_quota	*fdq)
{
1297
	return -XFS_QM_QUOTACTL(XFS_M(sb),
1298 1299 1300
				 (type == USRQUOTA) ? Q_XSETQLIM :
				  ((type == GRPQUOTA) ? Q_XSETGQLIM :
				   Q_XSETPQLIM), id, (caddr_t)fdq);
L
Linus Torvalds 已提交
1301 1302 1303
}

STATIC int
1304
xfs_fs_fill_super(
L
Linus Torvalds 已提交
1305 1306 1307 1308
	struct super_block	*sb,
	void			*data,
	int			silent)
{
1309
	struct inode		*root;
1310
	struct xfs_mount	*mp = NULL;
1311
	struct xfs_mount_args	*args = xfs_args_allocate(sb, silent);
1312
	int			error;
L
Linus Torvalds 已提交
1313

1314
	mp = xfs_mount_init();
L
Linus Torvalds 已提交
1315

1316 1317 1318 1319
	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
1320 1321
	mp->m_super = sb;
	sb->s_fs_info = mp;
L
Linus Torvalds 已提交
1322

1323 1324 1325
	if (sb->s_flags & MS_RDONLY)
		mp->m_flags |= XFS_MOUNT_RDONLY;

1326 1327
	error = xfs_parseargs(mp, (char *)data, args, 0);
	if (error)
L
Linus Torvalds 已提交
1328 1329 1330
		goto fail_vfsop;

	sb_min_blocksize(sb, BBSIZE);
1331 1332 1333
	sb->s_export_op = &xfs_export_operations;
	sb->s_qcop = &xfs_quotactl_operations;
	sb->s_op = &xfs_super_operations;
L
Linus Torvalds 已提交
1334

1335 1336
	error = xfs_mount(mp, args, NULL);
	if (error)
L
Linus Torvalds 已提交
1337 1338 1339
		goto fail_vfsop;

	sb->s_dirt = 1;
C
Christoph Hellwig 已提交
1340 1341 1342
	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
1343 1344 1345 1346
	sb->s_maxbytes = xfs_max_file_offset(sb->s_blocksize_bits);
	sb->s_time_gran = 1;
	set_posix_acl_flag(sb);

1347 1348
	root = igrab(mp->m_rootip->i_vnode);
	if (!root) {
C
Christoph Hellwig 已提交
1349
		error = ENOENT;
L
Linus Torvalds 已提交
1350
		goto fail_unmount;
C
Christoph Hellwig 已提交
1351
	}
1352 1353
	if (is_bad_inode(root)) {
		error = EINVAL;
L
Linus Torvalds 已提交
1354 1355
		goto fail_vnrele;
	}
1356 1357 1358
	sb->s_root = d_alloc_root(root);
	if (!sb->s_root) {
		error = ENOMEM;
L
Linus Torvalds 已提交
1359 1360
		goto fail_vnrele;
	}
1361 1362 1363 1364 1365 1366

	mp->m_sync_work.w_syncer = xfs_sync_worker;
	mp->m_sync_work.w_mount = mp;
	mp->m_sync_task = kthread_run(xfssyncd, mp, "xfssyncd");
	if (IS_ERR(mp->m_sync_task)) {
		error = -PTR_ERR(mp->m_sync_task);
L
Linus Torvalds 已提交
1367
		goto fail_vnrele;
1368 1369
	}

1370
	xfs_itrace_exit(XFS_I(sb->s_root->d_inode));
L
Linus Torvalds 已提交
1371

1372
	kmem_free(args);
L
Linus Torvalds 已提交
1373 1374 1375 1376 1377 1378 1379
	return 0;

fail_vnrele:
	if (sb->s_root) {
		dput(sb->s_root);
		sb->s_root = NULL;
	} else {
1380
		iput(root);
L
Linus Torvalds 已提交
1381 1382 1383
	}

fail_unmount:
1384
	xfs_unmount(mp, 0, NULL);
L
Linus Torvalds 已提交
1385 1386

fail_vfsop:
1387
	kmem_free(args);
L
Linus Torvalds 已提交
1388 1389 1390
	return -error;
}

1391
STATIC int
1392
xfs_fs_get_sb(
L
Linus Torvalds 已提交
1393 1394 1395
	struct file_system_type	*fs_type,
	int			flags,
	const char		*dev_name,
1396 1397
	void			*data,
	struct vfsmount		*mnt)
L
Linus Torvalds 已提交
1398
{
1399 1400
	return get_sb_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super,
			   mnt);
1401 1402
}

1403
static struct super_operations xfs_super_operations = {
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
	.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,
	.write_super_lockfs	= xfs_fs_lockfs,
	.statfs			= xfs_fs_statfs,
	.remount_fs		= xfs_fs_remount,
	.show_options		= xfs_fs_show_options,
L
Linus Torvalds 已提交
1415 1416
};

1417
static struct quotactl_ops xfs_quotactl_operations = {
1418 1419 1420 1421 1422
	.quota_sync		= xfs_fs_quotasync,
	.get_xstate		= xfs_fs_getxstate,
	.set_xstate		= xfs_fs_setxstate,
	.get_xquota		= xfs_fs_getxquota,
	.set_xquota		= xfs_fs_setxquota,
L
Linus Torvalds 已提交
1423 1424
};

A
Andrew Morton 已提交
1425
static struct file_system_type xfs_fs_type = {
L
Linus Torvalds 已提交
1426 1427
	.owner			= THIS_MODULE,
	.name			= "xfs",
1428
	.get_sb			= xfs_fs_get_sb,
L
Linus Torvalds 已提交
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
	.kill_sb		= kill_block_super,
	.fs_flags		= FS_REQUIRES_DEV,
};


STATIC int __init
init_xfs_fs( void )
{
	int			error;
	static char		message[] __initdata = KERN_INFO \
		XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n";

	printk(message);

	ktrace_init(64);

1445
	error = xfs_init_zones();
L
Linus Torvalds 已提交
1446
	if (error < 0)
1447
		goto undo_zones;
L
Linus Torvalds 已提交
1448

1449
	error = xfs_buf_init();
L
Linus Torvalds 已提交
1450
	if (error < 0)
1451
		goto undo_buffers;
L
Linus Torvalds 已提交
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463

	vn_init();
	xfs_init();
	uuid_init();
	vfs_initquota();

	error = register_filesystem(&xfs_fs_type);
	if (error)
		goto undo_register;
	return 0;

undo_register:
1464
	xfs_buf_terminate();
L
Linus Torvalds 已提交
1465

1466
undo_buffers:
1467
	xfs_destroy_zones();
L
Linus Torvalds 已提交
1468

1469
undo_zones:
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478
	return error;
}

STATIC void __exit
exit_xfs_fs( void )
{
	vfs_exitquota();
	unregister_filesystem(&xfs_fs_type);
	xfs_cleanup();
1479
	xfs_buf_terminate();
1480
	xfs_destroy_zones();
L
Linus Torvalds 已提交
1481 1482 1483 1484 1485 1486 1487 1488 1489
	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");