xfs_itable.c 25.6 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2 3
 * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
 * 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_fs.h"
L
Linus Torvalds 已提交
20
#include "xfs_types.h"
21
#include "xfs_bit.h"
L
Linus Torvalds 已提交
22
#include "xfs_log.h"
23
#include "xfs_inum.h"
L
Linus Torvalds 已提交
24 25
#include "xfs_trans.h"
#include "xfs_sb.h"
26
#include "xfs_ag.h"
L
Linus Torvalds 已提交
27 28 29 30
#include "xfs_dir2.h"
#include "xfs_dmapi.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 37 38 39
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_error.h"
40
#include "xfs_btree.h"
L
Linus Torvalds 已提交
41

42 43 44 45 46 47
int
xfs_internal_inum(
	xfs_mount_t	*mp,
	xfs_ino_t	ino)
{
	return (ino == mp->m_sb.sb_rbmino || ino == mp->m_sb.sb_rsumino ||
48
		(xfs_sb_version_hasquota(&mp->m_sb) &&
49 50 51
		 (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino)));
}

L
Linus Torvalds 已提交
52 53 54 55 56 57 58 59
STATIC int
xfs_bulkstat_one_iget(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	ino,		/* inode number to get data for */
	xfs_daddr_t	bno,		/* starting bno of inode cluster */
	xfs_bstat_t	*buf,		/* return buffer */
	int		*stat)		/* BULKSTAT_RV_... */
{
60
	xfs_icdinode_t	*dic;	/* dinode core info pointer */
L
Linus Torvalds 已提交
61 62 63
	xfs_inode_t	*ip;		/* incore inode pointer */
	int		error;

64 65
	error = xfs_iget(mp, NULL, ino,
			 XFS_IGET_BULKSTAT, XFS_ILOCK_SHARED, &ip, bno);
L
Linus Torvalds 已提交
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
	if (error) {
		*stat = BULKSTAT_RV_NOTHING;
		return error;
	}

	ASSERT(ip != NULL);
	ASSERT(ip->i_blkno != (xfs_daddr_t)0);

	dic = &ip->i_d;

	/* xfs_iget returns the following without needing
	 * further change.
	 */
	buf->bs_nlink = dic->di_nlink;
	buf->bs_projid = dic->di_projid;
	buf->bs_ino = ino;
	buf->bs_mode = dic->di_mode;
	buf->bs_uid = dic->di_uid;
	buf->bs_gid = dic->di_gid;
	buf->bs_size = dic->di_size;
86
	vn_atime_to_bstime(VFS_I(ip), &buf->bs_atime);
L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
	buf->bs_mtime.tv_sec = dic->di_mtime.t_sec;
	buf->bs_mtime.tv_nsec = dic->di_mtime.t_nsec;
	buf->bs_ctime.tv_sec = dic->di_ctime.t_sec;
	buf->bs_ctime.tv_nsec = dic->di_ctime.t_nsec;
	buf->bs_xflags = xfs_ip2xflags(ip);
	buf->bs_extsize = dic->di_extsize << mp->m_sb.sb_blocklog;
	buf->bs_extents = dic->di_nextents;
	buf->bs_gen = dic->di_gen;
	memset(buf->bs_pad, 0, sizeof(buf->bs_pad));
	buf->bs_dmevmask = dic->di_dmevmask;
	buf->bs_dmstate = dic->di_dmstate;
	buf->bs_aextents = dic->di_anextents;

	switch (dic->di_format) {
	case XFS_DINODE_FMT_DEV:
		buf->bs_rdev = ip->i_df.if_u2.if_rdev;
		buf->bs_blksize = BLKDEV_IOSIZE;
		buf->bs_blocks = 0;
		break;
	case XFS_DINODE_FMT_LOCAL:
	case XFS_DINODE_FMT_UUID:
		buf->bs_rdev = 0;
		buf->bs_blksize = mp->m_sb.sb_blocksize;
		buf->bs_blocks = 0;
		break;
	case XFS_DINODE_FMT_EXTENTS:
	case XFS_DINODE_FMT_BTREE:
		buf->bs_rdev = 0;
		buf->bs_blksize = mp->m_sb.sb_blocksize;
		buf->bs_blocks = dic->di_nblocks + ip->i_delayed_blks;
		break;
	}

	xfs_iput(ip, XFS_ILOCK_SHARED);
	return error;
}

124
STATIC void
L
Linus Torvalds 已提交
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145
xfs_bulkstat_one_dinode(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	ino,		/* inode number to get data for */
	xfs_dinode_t	*dip,		/* dinode inode pointer */
	xfs_bstat_t	*buf)		/* return buffer */
{
	xfs_dinode_core_t *dic;		/* dinode core info pointer */

	dic = &dip->di_core;

	/*
	 * The inode format changed when we moved the link count and
	 * made it 32 bits long.  If this is an old format inode,
	 * convert it in memory to look like a new one.  If it gets
	 * flushed to disk we will convert back before flushing or
	 * logging it.  We zero out the new projid field and the old link
	 * count field.  We'll handle clearing the pad field (the remains
	 * of the old uuid field) when we actually convert the inode to
	 * the new format. We don't change the version number so that we
	 * can distinguish this from a real new format inode.
	 */
146 147
	if (dic->di_version == XFS_DINODE_VERSION_1) {
		buf->bs_nlink = be16_to_cpu(dic->di_onlink);
L
Linus Torvalds 已提交
148 149
		buf->bs_projid = 0;
	} else {
150 151
		buf->bs_nlink = be32_to_cpu(dic->di_nlink);
		buf->bs_projid = be16_to_cpu(dic->di_projid);
L
Linus Torvalds 已提交
152 153 154
	}

	buf->bs_ino = ino;
155 156 157 158 159 160 161 162 163 164
	buf->bs_mode = be16_to_cpu(dic->di_mode);
	buf->bs_uid = be32_to_cpu(dic->di_uid);
	buf->bs_gid = be32_to_cpu(dic->di_gid);
	buf->bs_size = be64_to_cpu(dic->di_size);
	buf->bs_atime.tv_sec = be32_to_cpu(dic->di_atime.t_sec);
	buf->bs_atime.tv_nsec = be32_to_cpu(dic->di_atime.t_nsec);
	buf->bs_mtime.tv_sec = be32_to_cpu(dic->di_mtime.t_sec);
	buf->bs_mtime.tv_nsec = be32_to_cpu(dic->di_mtime.t_nsec);
	buf->bs_ctime.tv_sec = be32_to_cpu(dic->di_ctime.t_sec);
	buf->bs_ctime.tv_nsec = be32_to_cpu(dic->di_ctime.t_nsec);
165
	buf->bs_xflags = xfs_dic2xflags(dip);
166 167 168
	buf->bs_extsize = be32_to_cpu(dic->di_extsize) << mp->m_sb.sb_blocklog;
	buf->bs_extents = be32_to_cpu(dic->di_nextents);
	buf->bs_gen = be32_to_cpu(dic->di_gen);
L
Linus Torvalds 已提交
169
	memset(buf->bs_pad, 0, sizeof(buf->bs_pad));
170 171 172
	buf->bs_dmevmask = be32_to_cpu(dic->di_dmevmask);
	buf->bs_dmstate = be16_to_cpu(dic->di_dmstate);
	buf->bs_aextents = be16_to_cpu(dic->di_anextents);
L
Linus Torvalds 已提交
173

174
	switch (dic->di_format) {
L
Linus Torvalds 已提交
175
	case XFS_DINODE_FMT_DEV:
176
		buf->bs_rdev = be32_to_cpu(dip->di_u.di_dev);
L
Linus Torvalds 已提交
177 178 179 180 181 182 183 184 185 186 187 188 189
		buf->bs_blksize = BLKDEV_IOSIZE;
		buf->bs_blocks = 0;
		break;
	case XFS_DINODE_FMT_LOCAL:
	case XFS_DINODE_FMT_UUID:
		buf->bs_rdev = 0;
		buf->bs_blksize = mp->m_sb.sb_blocksize;
		buf->bs_blocks = 0;
		break;
	case XFS_DINODE_FMT_EXTENTS:
	case XFS_DINODE_FMT_BTREE:
		buf->bs_rdev = 0;
		buf->bs_blksize = mp->m_sb.sb_blocksize;
190
		buf->bs_blocks = be64_to_cpu(dic->di_nblocks);
L
Linus Torvalds 已提交
191 192 193 194
		break;
	}
}

195 196 197 198 199 200 201 202 203 204
STATIC int
xfs_bulkstat_one_fmt(
	void			__user *ubuffer,
	const xfs_bstat_t	*buffer)
{
	if (copy_to_user(ubuffer, buffer, sizeof(*buffer)))
		return -EFAULT;
	return sizeof(*buffer);
}

L
Linus Torvalds 已提交
205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
/*
 * Return stat information for one inode.
 * Return 0 if ok, else errno.
 */
int		       		/* error status */
xfs_bulkstat_one(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	ino,		/* inode number to get data for */
	void		__user *buffer,	/* buffer to place output in */
	int		ubsize,		/* size of buffer */
	void		*private_data,	/* my private data */
	xfs_daddr_t	bno,		/* starting bno of inode cluster */
	int		*ubused,	/* bytes used by me */
	void		*dibuff,	/* on-disk inode buffer */
	int		*stat)		/* BULKSTAT_RV_... */
{
	xfs_bstat_t	*buf;		/* return buffer */
	int		error = 0;	/* error value */
	xfs_dinode_t	*dip;		/* dinode inode pointer */
224
	bulkstat_one_fmt_pf formatter = private_data ? : xfs_bulkstat_one_fmt;
L
Linus Torvalds 已提交
225 226

	dip = (xfs_dinode_t *)dibuff;
227
	*stat = BULKSTAT_RV_NOTHING;
L
Linus Torvalds 已提交
228

229
	if (!buffer || xfs_internal_inum(mp, ino))
L
Linus Torvalds 已提交
230
		return XFS_ERROR(EINVAL);
231
	if (ubsize < sizeof(*buf))
L
Linus Torvalds 已提交
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
		return XFS_ERROR(ENOMEM);

	buf = kmem_alloc(sizeof(*buf), KM_SLEEP);

	if (dip == NULL) {
		/* We're not being passed a pointer to a dinode.  This happens
		 * if BULKSTAT_FG_IGET is selected.  Do the iget.
		 */
		error = xfs_bulkstat_one_iget(mp, ino, bno, buf, stat);
		if (error)
			goto out_free;
	} else {
		xfs_bulkstat_one_dinode(mp, ino, dip, buf);
	}

247 248
	error = formatter(buffer, buf);
	if (error < 0)  {
249
		error = EFAULT;
L
Linus Torvalds 已提交
250 251 252 253 254
		goto out_free;
	}

	*stat = BULKSTAT_RV_DIDONE;
	if (ubused)
255
		*ubused = error;
L
Linus Torvalds 已提交
256 257

 out_free:
258
	kmem_free(buf);
L
Linus Torvalds 已提交
259 260 261
	return error;
}

262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
/*
 * Test to see whether we can use the ondisk inode directly, based
 * on the given bulkstat flags, filling in dipp accordingly.
 * Returns zero if the inode is dodgey.
 */
STATIC int
xfs_bulkstat_use_dinode(
	xfs_mount_t	*mp,
	int		flags,
	xfs_buf_t	*bp,
	int		clustidx,
	xfs_dinode_t	**dipp)
{
	xfs_dinode_t	*dip;
	unsigned int	aformat;

	*dipp = NULL;
	if (!bp || (flags & BULKSTAT_FG_IGET))
		return 1;
	dip = (xfs_dinode_t *)
			xfs_buf_offset(bp, clustidx << mp->m_sb.sb_inodelog);
283
	/*
284
	 * Check the buffer containing the on-disk inode for di_mode == 0.
285 286 287 288 289
	 * This is to prevent xfs_bulkstat from picking up just reclaimed
	 * inodes that have their in-core state initialized but not flushed
	 * to disk yet. This is a temporary hack that would require a proper
	 * fix in the future.
	 */
290
	if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC ||
291
	    !XFS_DINODE_GOOD_VERSION(dip->di_core.di_version) ||
292
	    !dip->di_core.di_mode)
293 294 295 296 297 298
		return 0;
	if (flags & BULKSTAT_FG_QUICK) {
		*dipp = dip;
		return 1;
	}
	/* BULKSTAT_FG_INLINE: if attr fork is local, or not there, use it */
299
	aformat = dip->di_core.di_aformat;
300
	if ((XFS_DFORK_Q(dip) == 0) ||
301 302 303 304 305 306 307 308
	    (aformat == XFS_DINODE_FMT_LOCAL) ||
	    (aformat == XFS_DINODE_FMT_EXTENTS && !dip->di_core.di_anextents)) {
		*dipp = dip;
		return 1;
	}
	return 1;
}

309 310
#define XFS_BULKSTAT_UBLEFT(ubleft)	((ubleft) >= statstruct_size)

L
Linus Torvalds 已提交
311 312 313 314 315 316 317 318 319 320 321 322 323
/*
 * Return stat information in bulk (by-inode) for the filesystem.
 */
int					/* error status */
xfs_bulkstat(
	xfs_mount_t		*mp,	/* mount point for filesystem */
	xfs_ino_t		*lastinop, /* last inode returned */
	int			*ubcountp, /* size of buffer/count returned */
	bulkstat_one_pf		formatter, /* func that'd fill a single buf */
	void			*private_data,/* private data for formatter */
	size_t			statstruct_size, /* sizeof struct filling */
	char			__user *ubuffer, /* buffer with inode stats */
	int			flags,	/* defined in xfs_itable.h */
324
	int			*done)	/* 1 if there are more stats to get */
L
Linus Torvalds 已提交
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
{
	xfs_agblock_t		agbno=0;/* allocation group block number */
	xfs_buf_t		*agbp;	/* agi header buffer */
	xfs_agi_t		*agi;	/* agi header data */
	xfs_agino_t		agino;	/* inode # in allocation group */
	xfs_agnumber_t		agno;	/* allocation group number */
	xfs_daddr_t		bno;	/* inode cluster start daddr */
	int			chunkidx; /* current index into inode chunk */
	int			clustidx; /* current index into inode cluster */
	xfs_btree_cur_t		*cur;	/* btree cursor for ialloc btree */
	int			end_of_ag; /* set if we've seen the ag end */
	int			error;	/* error code */
	int                     fmterror;/* bulkstat formatter result */
	__int32_t		gcnt;	/* current btree rec's count */
	xfs_inofree_t		gfree;	/* current btree rec's free mask */
	xfs_agino_t		gino;	/* current btree rec's start inode */
	int			i;	/* loop index */
	int			icount;	/* count of inodes good in irbuf */
343
	size_t			irbsize; /* size of irec buffer in bytes */
L
Linus Torvalds 已提交
344
	xfs_ino_t		ino;	/* inode number (filesystem) */
345 346 347
	xfs_inobt_rec_incore_t	*irbp;	/* current irec buffer pointer */
	xfs_inobt_rec_incore_t	*irbuf;	/* start of irec buffer */
	xfs_inobt_rec_incore_t	*irbufend; /* end of good irec buffer entries */
348
	xfs_ino_t		lastino; /* last inode number returned */
L
Linus Torvalds 已提交
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
	int			nbcluster; /* # of blocks in a cluster */
	int			nicluster; /* # of inodes in a cluster */
	int			nimask;	/* mask for inode clusters */
	int			nirbuf;	/* size of irbuf */
	int			rval;	/* return value error code */
	int			tmp;	/* result value from btree calls */
	int			ubcount; /* size of user's buffer */
	int			ubleft;	/* bytes left in user's buffer */
	char			__user *ubufp;	/* pointer into user's buffer */
	int			ubelem;	/* spaces used in user's buffer */
	int			ubused;	/* bytes used by formatter */
	xfs_buf_t		*bp;	/* ptr to on-disk inode cluster buf */
	xfs_dinode_t		*dip;	/* ptr into bp for specific inode */
	xfs_inode_t		*ip;	/* ptr to in-core inode struct */

	/*
	 * Get the last inode value, see if there's nothing to do.
	 */
	ino = (xfs_ino_t)*lastinop;
368
	lastino = ino;
L
Linus Torvalds 已提交
369 370 371 372 373 374 375 376 377
	dip = NULL;
	agno = XFS_INO_TO_AGNO(mp, ino);
	agino = XFS_INO_TO_AGINO(mp, ino);
	if (agno >= mp->m_sb.sb_agcount ||
	    ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
		*done = 1;
		*ubcountp = 0;
		return 0;
	}
378 379 380
	if (!ubcountp || *ubcountp <= 0) {
		return EINVAL;
	}
L
Linus Torvalds 已提交
381 382 383 384 385 386 387 388 389 390 391
	ubcount = *ubcountp; /* statstruct's */
	ubleft = ubcount * statstruct_size; /* bytes */
	*ubcountp = ubelem = 0;
	*done = 0;
	fmterror = 0;
	ubufp = ubuffer;
	nicluster = mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp) ?
		mp->m_sb.sb_inopblock :
		(XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog);
	nimask = ~(nicluster - 1);
	nbcluster = nicluster >> mp->m_sb.sb_inopblog;
392
	irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4,
393
				   KM_SLEEP | KM_MAYFAIL | KM_LARGE);
394 395
	nirbuf = irbsize / sizeof(*irbuf);

L
Linus Torvalds 已提交
396 397 398 399 400
	/*
	 * Loop over the allocation groups, starting from the last
	 * inode returned; 0 means start of the allocation group.
	 */
	rval = 0;
401 402
	while (XFS_BULKSTAT_UBLEFT(ubleft) && agno < mp->m_sb.sb_agcount) {
		cond_resched();
L
Linus Torvalds 已提交
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
		bp = NULL;
		down_read(&mp->m_peraglock);
		error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp);
		up_read(&mp->m_peraglock);
		if (error) {
			/*
			 * Skip this allocation group and go to the next one.
			 */
			agno++;
			agino = 0;
			continue;
		}
		agi = XFS_BUF_TO_AGI(agbp);
		/*
		 * Allocate and initialize a btree cursor for ialloc btree.
		 */
		cur = xfs_btree_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_INO,
420
						(xfs_inode_t *)0, 0);
L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456
		irbp = irbuf;
		irbufend = irbuf + nirbuf;
		end_of_ag = 0;
		/*
		 * If we're returning in the middle of an allocation group,
		 * we need to get the remainder of the chunk we're in.
		 */
		if (agino > 0) {
			/*
			 * Lookup the inode chunk that this inode lives in.
			 */
			error = xfs_inobt_lookup_le(cur, agino, 0, 0, &tmp);
			if (!error &&	/* no I/O error */
			    tmp &&	/* lookup succeeded */
					/* got the record, should always work */
			    !(error = xfs_inobt_get_rec(cur, &gino, &gcnt,
				    &gfree, &i)) &&
			    i == 1 &&
					/* this is the right chunk */
			    agino < gino + XFS_INODES_PER_CHUNK &&
					/* lastino was not last in chunk */
			    (chunkidx = agino - gino + 1) <
				    XFS_INODES_PER_CHUNK &&
					/* there are some left allocated */
			    XFS_INOBT_MASKN(chunkidx,
				    XFS_INODES_PER_CHUNK - chunkidx) & ~gfree) {
				/*
				 * Grab the chunk record.  Mark all the
				 * uninteresting inodes (because they're
				 * before our start point) free.
				 */
				for (i = 0; i < chunkidx; i++) {
					if (XFS_INOBT_MASK(i) & ~gfree)
						gcnt++;
				}
				gfree |= XFS_INOBT_MASKN(0, chunkidx);
457 458 459
				irbp->ir_startino = gino;
				irbp->ir_freecount = gcnt;
				irbp->ir_free = gfree;
L
Linus Torvalds 已提交
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
				irbp++;
				agino = gino + XFS_INODES_PER_CHUNK;
				icount = XFS_INODES_PER_CHUNK - gcnt;
			} else {
				/*
				 * If any of those tests failed, bump the
				 * inode number (just in case).
				 */
				agino++;
				icount = 0;
			}
			/*
			 * In any case, increment to the next record.
			 */
			if (!error)
				error = xfs_inobt_increment(cur, 0, &tmp);
		} else {
			/*
			 * Start of ag.  Lookup the first inode chunk.
			 */
			error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &tmp);
			icount = 0;
		}
		/*
		 * Loop through inode btree records in this ag,
		 * until we run out of inodes or space in the buffer.
		 */
		while (irbp < irbufend && icount < ubcount) {
			/*
			 * Loop as long as we're unable to read the
			 * inode btree.
			 */
			while (error) {
				agino += XFS_INODES_PER_CHUNK;
				if (XFS_AGINO_TO_AGBNO(mp, agino) >=
495
						be32_to_cpu(agi->agi_length))
L
Linus Torvalds 已提交
496 497 498
					break;
				error = xfs_inobt_lookup_ge(cur, agino, 0, 0,
							    &tmp);
499
				cond_resched();
L
Linus Torvalds 已提交
500 501 502 503 504 505 506 507 508 509 510 511 512 513
			}
			/*
			 * If ran off the end of the ag either with an error,
			 * or the normal way, set end and stop collecting.
			 */
			if (error ||
			    (error = xfs_inobt_get_rec(cur, &gino, &gcnt,
				    &gfree, &i)) ||
			    i == 0) {
				end_of_ag = 1;
				break;
			}
			/*
			 * If this chunk has any allocated inodes, save it.
514
			 * Also start read-ahead now for this chunk.
L
Linus Torvalds 已提交
515 516
			 */
			if (gcnt < XFS_INODES_PER_CHUNK) {
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
				/*
				 * Loop over all clusters in the next chunk.
				 * Do a readahead if there are any allocated
				 * inodes in that cluster.
				 */
				for (agbno = XFS_AGINO_TO_AGBNO(mp, gino),
				     chunkidx = 0;
				     chunkidx < XFS_INODES_PER_CHUNK;
				     chunkidx += nicluster,
				     agbno += nbcluster) {
					if (XFS_INOBT_MASKN(chunkidx,
							    nicluster) & ~gfree)
						xfs_btree_reada_bufs(mp, agno,
							agbno, nbcluster);
				}
				irbp->ir_startino = gino;
				irbp->ir_freecount = gcnt;
				irbp->ir_free = gfree;
L
Linus Torvalds 已提交
535 536 537 538 539 540 541 542
				irbp++;
				icount += XFS_INODES_PER_CHUNK - gcnt;
			}
			/*
			 * Set agino to after this chunk and bump the cursor.
			 */
			agino = gino + XFS_INODES_PER_CHUNK;
			error = xfs_inobt_increment(cur, 0, &tmp);
543
			cond_resched();
L
Linus Torvalds 已提交
544 545 546 547 548 549 550 551 552 553 554 555 556
		}
		/*
		 * Drop the btree buffers and the agi buffer.
		 * We can't hold any of the locks these represent
		 * when calling iget.
		 */
		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
		xfs_buf_relse(agbp);
		/*
		 * Now format all the good inodes into the user's buffer.
		 */
		irbufend = irbp;
		for (irbp = irbuf;
557
		     irbp < irbufend && XFS_BULKSTAT_UBLEFT(ubleft); irbp++) {
L
Linus Torvalds 已提交
558 559 560
			/*
			 * Now process this chunk of inodes.
			 */
561
			for (agino = irbp->ir_startino, chunkidx = clustidx = 0;
562
			     XFS_BULKSTAT_UBLEFT(ubleft) &&
563
				irbp->ir_freecount < XFS_INODES_PER_CHUNK;
L
Linus Torvalds 已提交
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
			     chunkidx++, clustidx++, agino++) {
				ASSERT(chunkidx < XFS_INODES_PER_CHUNK);
				/*
				 * Recompute agbno if this is the
				 * first inode of the cluster.
				 *
				 * Careful with clustidx.   There can be
				 * multple clusters per chunk, a single
				 * cluster per chunk or a cluster that has
				 * inodes represented from several different
				 * chunks (if blocksize is large).
				 *
				 * Because of this, the starting clustidx is
				 * initialized to zero in this loop but must
				 * later be reset after reading in the cluster
				 * buffer.
				 */
				if ((chunkidx & (nicluster - 1)) == 0) {
					agbno = XFS_AGINO_TO_AGBNO(mp,
583
							irbp->ir_startino) +
L
Linus Torvalds 已提交
584 585 586
						((chunkidx & nimask) >>
						 mp->m_sb.sb_inopblog);

587 588
					if (flags & (BULKSTAT_FG_QUICK |
						     BULKSTAT_FG_INLINE)) {
L
Linus Torvalds 已提交
589 590 591 592 593 594 595 596 597 598
						ino = XFS_AGINO_TO_INO(mp, agno,
								       agino);
						bno = XFS_AGB_TO_DADDR(mp, agno,
								       agbno);

						/*
						 * Get the inode cluster buffer
						 */
						if (bp)
							xfs_buf_relse(bp);
599 600 601 602 603 604
						ip = xfs_inode_alloc(mp, ino);
						if (!ip) {
							bp = NULL;
							rval = ENOMEM;
							break;
						}
L
Linus Torvalds 已提交
605
						error = xfs_itobp(mp, NULL, ip,
606
								&dip, &bp, bno,
607 608
								XFS_IMAP_BULKSTAT,
								XFS_BUF_LOCK);
L
Linus Torvalds 已提交
609 610
						if (!error)
							clustidx = ip->i_boffset / mp->m_sb.sb_inodesize;
611
						xfs_idestroy(ip);
L
Linus Torvalds 已提交
612 613 614 615
						if (XFS_TEST_ERROR(error != 0,
								   mp, XFS_ERRTAG_BULKSTAT_READ_CHUNK,
								   XFS_RANDOM_BULKSTAT_READ_CHUNK)) {
							bp = NULL;
616 617
							ubleft = 0;
							rval = error;
L
Linus Torvalds 已提交
618 619 620 621
							break;
						}
					}
				}
622 623
				ino = XFS_AGINO_TO_INO(mp, agno, agino);
				bno = XFS_AGB_TO_DADDR(mp, agno, agbno);
L
Linus Torvalds 已提交
624 625 626
				/*
				 * Skip if this inode is free.
				 */
627 628
				if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free) {
					lastino = ino;
L
Linus Torvalds 已提交
629
					continue;
630
				}
L
Linus Torvalds 已提交
631 632 633 634
				/*
				 * Count used inodes as free so we can tell
				 * when the chunk is used up.
				 */
635
				irbp->ir_freecount++;
636
				if (!xfs_bulkstat_use_dinode(mp, flags, bp,
637 638
							     clustidx, &dip)) {
					lastino = ino;
639
					continue;
640
				}
641 642 643 644 645 646 647 648
				/*
				 * If we need to do an iget, cannot hold bp.
				 * Drop it, until starting the next cluster.
				 */
				if ((flags & BULKSTAT_FG_INLINE) && !dip) {
					if (bp)
						xfs_buf_relse(bp);
					bp = NULL;
L
Linus Torvalds 已提交
649 650 651 652 653 654
				}

				/*
				 * Get the inode and fill in a single buffer.
				 * BULKSTAT_FG_QUICK uses dip to fill it in.
				 * BULKSTAT_FG_IGET uses igets.
655 656
				 * BULKSTAT_FG_INLINE uses dip if we have an
				 * inline attr fork, else igets.
L
Linus Torvalds 已提交
657 658 659 660 661 662 663 664 665
				 * See: xfs_bulkstat_one & xfs_dm_bulkstat_one.
				 * This is also used to count inodes/blks, etc
				 * in xfs_qm_quotacheck.
				 */
				ubused = statstruct_size;
				error = formatter(mp, ino, ubufp,
						ubleft, private_data,
						bno, &ubused, dip, &fmterror);
				if (fmterror == BULKSTAT_RV_NOTHING) {
666 667
					if (error && error != ENOENT &&
						error != EINVAL) {
668
						ubleft = 0;
669 670 671 672
						rval = error;
						break;
					}
					lastino = ino;
L
Linus Torvalds 已提交
673 674 675 676 677 678 679 680 681 682 683 684 685 686
					continue;
				}
				if (fmterror == BULKSTAT_RV_GIVEUP) {
					ubleft = 0;
					ASSERT(error);
					rval = error;
					break;
				}
				if (ubufp)
					ubufp += ubused;
				ubleft -= ubused;
				ubelem++;
				lastino = ino;
			}
687 688

			cond_resched();
L
Linus Torvalds 已提交
689 690 691 692 693 694 695 696
		}

		if (bp)
			xfs_buf_relse(bp);

		/*
		 * Set up for the next loop iteration.
		 */
697
		if (XFS_BULKSTAT_UBLEFT(ubleft)) {
L
Linus Torvalds 已提交
698 699 700
			if (end_of_ag) {
				agno++;
				agino = 0;
701 702
			} else
				agino = XFS_INO_TO_AGINO(mp, lastino);
L
Linus Torvalds 已提交
703 704 705 706 707 708
		} else
			break;
	}
	/*
	 * Done, we're either out of filesystem or space to put the data.
	 */
709
	kmem_free(irbuf);
L
Linus Torvalds 已提交
710
	*ubcountp = ubelem;
711 712 713 714 715
	/*
	 * Found some inodes, return them now and return the error next time.
	 */
	if (ubelem)
		rval = 0;
L
Linus Torvalds 已提交
716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
	if (agno >= mp->m_sb.sb_agcount) {
		/*
		 * If we ran out of filesystem, mark lastino as off
		 * the end of the filesystem, so the next call
		 * will return immediately.
		 */
		*lastinop = (xfs_ino_t)XFS_AGINO_TO_INO(mp, agno, 0);
		*done = 1;
	} else
		*lastinop = (xfs_ino_t)lastino;

	return rval;
}

/*
 * Return stat information in bulk (by-inode) for the filesystem.
 * Special case for non-sequential one inode bulkstat.
 */
int					/* error status */
xfs_bulkstat_single(
	xfs_mount_t		*mp,	/* mount point for filesystem */
	xfs_ino_t		*lastinop, /* inode to return */
	char			__user *buffer, /* buffer with inode stats */
739
	int			*done)	/* 1 if there are more stats to get */
L
Linus Torvalds 已提交
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
{
	int			count;	/* count value for bulkstat call */
	int			error;	/* return value */
	xfs_ino_t		ino;	/* filesystem inode number */
	int			res;	/* result from bs1 */

	/*
	 * note that requesting valid inode numbers which are not allocated
	 * to inodes will most likely cause xfs_itobp to generate warning
	 * messages about bad magic numbers. This is ok. The fact that
	 * the inode isn't actually an inode is handled by the
	 * error check below. Done this way to make the usual case faster
	 * at the expense of the error case.
	 */

	ino = (xfs_ino_t)*lastinop;
	error = xfs_bulkstat_one(mp, ino, buffer, sizeof(xfs_bstat_t),
				 NULL, 0, NULL, NULL, &res);
	if (error) {
		/*
		 * Special case way failed, do it the "long" way
		 * to see if that works.
		 */
		(*lastinop)--;
		count = 1;
		if (xfs_bulkstat(mp, lastinop, &count, xfs_bulkstat_one,
				NULL, sizeof(xfs_bstat_t), buffer,
				BULKSTAT_FG_IGET, done))
			return error;
		if (count == 0 || (xfs_ino_t)*lastinop != ino)
			return error == EFSCORRUPTED ?
				XFS_ERROR(EINVAL) : error;
		else
			return 0;
	}
	*done = 0;
	return 0;
}

779 780 781 782 783 784 785 786 787 788 789 790 791
int
xfs_inumbers_fmt(
	void			__user *ubuffer, /* buffer to write to */
	const xfs_inogrp_t	*buffer,	/* buffer to read from */
	long			count,		/* # of elements to read */
	long			*written)	/* # of bytes written */
{
	if (copy_to_user(ubuffer, buffer, count * sizeof(*buffer)))
		return -EFAULT;
	*written = count * sizeof(*buffer);
	return 0;
}

L
Linus Torvalds 已提交
792 793 794 795 796 797 798 799
/*
 * Return inode number table for the filesystem.
 */
int					/* error status */
xfs_inumbers(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	*lastino,	/* last inode returned */
	int		*count,		/* size of buffer/count returned */
800 801
	void		__user *ubuffer,/* buffer with inode descriptions */
	inumbers_fmt_pf	formatter)
L
Linus Torvalds 已提交
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
{
	xfs_buf_t	*agbp;
	xfs_agino_t	agino;
	xfs_agnumber_t	agno;
	int		bcount;
	xfs_inogrp_t	*buffer;
	int		bufidx;
	xfs_btree_cur_t	*cur;
	int		error;
	__int32_t	gcnt;
	xfs_inofree_t	gfree;
	xfs_agino_t	gino;
	int		i;
	xfs_ino_t	ino;
	int		left;
	int		tmp;

	ino = (xfs_ino_t)*lastino;
	agno = XFS_INO_TO_AGNO(mp, ino);
	agino = XFS_INO_TO_AGINO(mp, ino);
	left = *count;
	*count = 0;
824
	bcount = MIN(left, (int)(PAGE_SIZE / sizeof(*buffer)));
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853
	buffer = kmem_alloc(bcount * sizeof(*buffer), KM_SLEEP);
	error = bufidx = 0;
	cur = NULL;
	agbp = NULL;
	while (left > 0 && agno < mp->m_sb.sb_agcount) {
		if (agbp == NULL) {
			down_read(&mp->m_peraglock);
			error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp);
			up_read(&mp->m_peraglock);
			if (error) {
				/*
				 * If we can't read the AGI of this ag,
				 * then just skip to the next one.
				 */
				ASSERT(cur == NULL);
				agbp = NULL;
				agno++;
				agino = 0;
				continue;
			}
			cur = xfs_btree_init_cursor(mp, NULL, agbp, agno,
				XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
			error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp);
			if (error) {
				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
				cur = NULL;
				xfs_buf_relse(agbp);
				agbp = NULL;
				/*
854
				 * Move up the last inode in the current
L
Linus Torvalds 已提交
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879
				 * chunk.  The lookup_ge will always get
				 * us the first inode in the next chunk.
				 */
				agino += XFS_INODES_PER_CHUNK - 1;
				continue;
			}
		}
		if ((error = xfs_inobt_get_rec(cur, &gino, &gcnt, &gfree,
			&i)) ||
		    i == 0) {
			xfs_buf_relse(agbp);
			agbp = NULL;
			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
			cur = NULL;
			agno++;
			agino = 0;
			continue;
		}
		agino = gino + XFS_INODES_PER_CHUNK - 1;
		buffer[bufidx].xi_startino = XFS_AGINO_TO_INO(mp, agno, gino);
		buffer[bufidx].xi_alloccount = XFS_INODES_PER_CHUNK - gcnt;
		buffer[bufidx].xi_allocmask = ~gfree;
		bufidx++;
		left--;
		if (bufidx == bcount) {
880 881
			long written;
			if (formatter(ubuffer, buffer, bufidx, &written)) {
L
Linus Torvalds 已提交
882 883 884
				error = XFS_ERROR(EFAULT);
				break;
			}
885
			ubuffer += written;
L
Linus Torvalds 已提交
886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906
			*count += bufidx;
			bufidx = 0;
		}
		if (left) {
			error = xfs_inobt_increment(cur, 0, &tmp);
			if (error) {
				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
				cur = NULL;
				xfs_buf_relse(agbp);
				agbp = NULL;
				/*
				 * The agino value has already been bumped.
				 * Just try to skip up to it.
				 */
				agino += XFS_INODES_PER_CHUNK;
				continue;
			}
		}
	}
	if (!error) {
		if (bufidx) {
907 908
			long written;
			if (formatter(ubuffer, buffer, bufidx, &written))
L
Linus Torvalds 已提交
909 910 911 912 913 914
				error = XFS_ERROR(EFAULT);
			else
				*count += bufidx;
		}
		*lastino = XFS_AGINO_TO_INO(mp, agno, agino);
	}
915
	kmem_free(buffer);
L
Linus Torvalds 已提交
916 917 918 919 920 921 922
	if (cur)
		xfs_btree_del_cursor(cur, (error ? XFS_BTREE_ERROR :
					   XFS_BTREE_NOERROR));
	if (agbp)
		xfs_buf_relse(agbp);
	return error;
}