xfs_qm_syscalls.c 35.1 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2 3
 * Copyright (c) 2000-2005 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 19 20

#include <linux/capability.h>

L
Linus Torvalds 已提交
21 22
#include "xfs.h"
#include "xfs_fs.h"
23
#include "xfs_bit.h"
L
Linus Torvalds 已提交
24
#include "xfs_log.h"
25
#include "xfs_inum.h"
L
Linus Torvalds 已提交
26 27
#include "xfs_trans.h"
#include "xfs_sb.h"
28
#include "xfs_ag.h"
L
Linus Torvalds 已提交
29 30 31 32 33 34
#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"
35
#include "xfs_alloc_btree.h"
L
Linus Torvalds 已提交
36 37
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
38
#include "xfs_attr_sf.h"
L
Linus Torvalds 已提交
39 40
#include "xfs_dinode.h"
#include "xfs_inode.h"
41 42
#include "xfs_ialloc.h"
#include "xfs_itable.h"
L
Linus Torvalds 已提交
43
#include "xfs_bmap.h"
44
#include "xfs_btree.h"
L
Linus Torvalds 已提交
45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
#include "xfs_qm.h"

#ifdef DEBUG
# define qdprintk(s, args...)	cmn_err(CE_DEBUG, s, ## args)
#else
# define qdprintk(s, args...)	do { } while (0)
#endif

STATIC int	xfs_qm_log_quotaoff(xfs_mount_t *, xfs_qoff_logitem_t **, uint);
STATIC int	xfs_qm_log_quotaoff_end(xfs_mount_t *, xfs_qoff_logitem_t *,
					uint);
STATIC uint	xfs_qm_export_flags(uint);
STATIC uint	xfs_qm_export_qtype_flags(uint);
STATIC void	xfs_qm_export_dquot(xfs_mount_t *, xfs_disk_dquot_t *,
					fs_disk_quota_t *);


/*
 * Turn off quota accounting and/or enforcement for all udquots and/or
 * gdquots. Called only at unmount time.
 *
 * This assumes that there are no dquots of this file system cached
 * incore, and modifies the ondisk dquot directly. Therefore, for example,
 * it is an error to call this twice, without purging the cache.
 */
77
int
L
Linus Torvalds 已提交
78 79
xfs_qm_scall_quotaoff(
	xfs_mount_t		*mp,
80
	uint			flags)
L
Linus Torvalds 已提交
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
{
	uint			dqtype;
	int			error;
	uint			inactivate_flags;
	xfs_qoff_logitem_t	*qoffstart;
	int			nculprits;

	/*
	 * No file system can have quotas enabled on disk but not in core.
	 * Note that quota utilities (like quotaoff) _expect_
	 * errno == EEXIST here.
	 */
	if ((mp->m_qflags & flags) == 0)
		return XFS_ERROR(EEXIST);
	error = 0;

	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);

	/*
	 * We don't want to deal with two quotaoffs messing up each other,
	 * so we're going to serialize it. quotaoff isn't exactly a performance
	 * critical thing.
	 * If quotaoff, then we must be dealing with the root filesystem.
	 */
	ASSERT(mp->m_quotainfo);
	if (mp->m_quotainfo)
107
		mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
L
Linus Torvalds 已提交
108 109 110 111 112 113 114 115 116

	ASSERT(mp->m_quotainfo);

	/*
	 * If we're just turning off quota enforcement, change mp and go.
	 */
	if ((flags & XFS_ALL_QUOTA_ACCT) == 0) {
		mp->m_qflags &= ~(flags);

E
Eric Sandeen 已提交
117
		spin_lock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
118
		mp->m_sb.sb_qflags = mp->m_qflags;
E
Eric Sandeen 已提交
119
		spin_unlock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));

		/* XXX what to do if error ? Revert back to old vals incore ? */
		error = xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS);
		return (error);
	}

	dqtype = 0;
	inactivate_flags = 0;
	/*
	 * If accounting is off, we must turn enforcement off, clear the
	 * quota 'CHKD' certificate to make it known that we have to
	 * do a quotacheck the next time this quota is turned on.
	 */
	if (flags & XFS_UQUOTA_ACCT) {
		dqtype |= XFS_QMOPT_UQUOTA;
		flags |= (XFS_UQUOTA_CHKD | XFS_UQUOTA_ENFD);
		inactivate_flags |= XFS_UQUOTA_ACTIVE;
	}
	if (flags & XFS_GQUOTA_ACCT) {
		dqtype |= XFS_QMOPT_GQUOTA;
141
		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
L
Linus Torvalds 已提交
142
		inactivate_flags |= XFS_GQUOTA_ACTIVE;
143 144 145 146
	} else if (flags & XFS_PQUOTA_ACCT) {
		dqtype |= XFS_QMOPT_PQUOTA;
		flags |= (XFS_OQUOTA_CHKD | XFS_OQUOTA_ENFD);
		inactivate_flags |= XFS_PQUOTA_ACTIVE;
L
Linus Torvalds 已提交
147 148 149 150 151 152 153 154 155 156 157 158 159
	}

	/*
	 * Nothing to do?  Don't complain. This happens when we're just
	 * turning off quota enforcement.
	 */
	if ((mp->m_qflags & flags) == 0) {
		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
		return (0);
	}

	/*
	 * Write the LI_QUOTAOFF log record, and do SB changes atomically,
160 161
	 * and synchronously. If we fail to write, we should abort the
	 * operation as it cannot be recovered safely if we crash.
L
Linus Torvalds 已提交
162
	 */
163 164 165
	error = xfs_qm_log_quotaoff(mp, &qoffstart, flags);
	if (error)
		goto out_error;
L
Linus Torvalds 已提交
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220

	/*
	 * Next we clear the XFS_MOUNT_*DQ_ACTIVE bit(s) in the mount struct
	 * to take care of the race between dqget and quotaoff. We don't take
	 * any special locks to reset these bits. All processes need to check
	 * these bits *after* taking inode lock(s) to see if the particular
	 * quota type is in the process of being turned off. If *ACTIVE, it is
	 * guaranteed that all dquot structures and all quotainode ptrs will all
	 * stay valid as long as that inode is kept locked.
	 *
	 * There is no turning back after this.
	 */
	mp->m_qflags &= ~inactivate_flags;

	/*
	 * Give back all the dquot reference(s) held by inodes.
	 * Here we go thru every single incore inode in this file system, and
	 * do a dqrele on the i_udquot/i_gdquot that it may have.
	 * Essentially, as long as somebody has an inode locked, this guarantees
	 * that quotas will not be turned off. This is handy because in a
	 * transaction once we lock the inode(s) and check for quotaon, we can
	 * depend on the quota inodes (and other things) being valid as long as
	 * we keep the lock(s).
	 */
	xfs_qm_dqrele_all_inodes(mp, flags);

	/*
	 * Next we make the changes in the quota flag in the mount struct.
	 * This isn't protected by a particular lock directly, because we
	 * don't want to take a mrlock everytime we depend on quotas being on.
	 */
	mp->m_qflags &= ~(flags);

	/*
	 * Go through all the dquots of this file system and purge them,
	 * according to what was turned off. We may not be able to get rid
	 * of all dquots, because dquots can have temporary references that
	 * are not attached to inodes. eg. xfs_setattr, xfs_create.
	 * So, if we couldn't purge all the dquots from the filesystem,
	 * we can't get rid of the incore data structures.
	 */
	while ((nculprits = xfs_qm_dqpurge_all(mp, dqtype|XFS_QMOPT_QUOTAOFF)))
		delay(10 * nculprits);

	/*
	 * Transactions that had started before ACTIVE state bit was cleared
	 * could have logged many dquots, so they'd have higher LSNs than
	 * the first QUOTAOFF log record does. If we happen to crash when
	 * the tail of the log has gone past the QUOTAOFF record, but
	 * before the last dquot modification, those dquots __will__
	 * recover, and that's not good.
	 *
	 * So, we have QUOTAOFF start and end logitems; the start
	 * logitem won't get overwritten until the end logitem appears...
	 */
221 222 223 224 225 226
	error = xfs_qm_log_quotaoff_end(mp, qoffstart, flags);
	if (error) {
		/* We're screwed now. Shutdown is the only option. */
		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
		goto out_error;
	}
L
Linus Torvalds 已提交
227 228 229 230

	/*
	 * If quotas is completely disabled, close shop.
	 */
231 232
	if (((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET1) ||
	    ((flags & XFS_MOUNT_QUOTA_ALL) == XFS_MOUNT_QUOTA_SET2)) {
L
Linus Torvalds 已提交
233 234 235 236 237 238 239 240 241 242
		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
		xfs_qm_destroy_quotainfo(mp);
		return (0);
	}

	/*
	 * Release our quotainode references, and vn_purge them,
	 * if we don't need them anymore.
	 */
	if ((dqtype & XFS_QMOPT_UQUOTA) && XFS_QI_UQIP(mp)) {
243
		IRELE(XFS_QI_UQIP(mp));
L
Linus Torvalds 已提交
244 245
		XFS_QI_UQIP(mp) = NULL;
	}
246
	if ((dqtype & (XFS_QMOPT_GQUOTA|XFS_QMOPT_PQUOTA)) && XFS_QI_GQIP(mp)) {
247
		IRELE(XFS_QI_GQIP(mp));
L
Linus Torvalds 已提交
248 249
		XFS_QI_GQIP(mp) = NULL;
	}
250
out_error:
L
Linus Torvalds 已提交
251 252 253 254 255
	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));

	return (error);
}

256
int
L
Linus Torvalds 已提交
257 258 259 260
xfs_qm_scall_trunc_qfiles(
	xfs_mount_t	*mp,
	uint		flags)
{
261
	int		error = 0, error2 = 0;
L
Linus Torvalds 已提交
262 263
	xfs_inode_t	*qip;

264
	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) {
L
Linus Torvalds 已提交
265 266 267 268 269 270
		qdprintk("qtrunc flags=%x m_qflags=%x\n", flags, mp->m_qflags);
		return XFS_ERROR(EINVAL);
	}

	if ((flags & XFS_DQ_USER) && mp->m_sb.sb_uquotino != NULLFSINO) {
		error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, 0, 0, &qip, 0);
271 272
		if (!error) {
			error = xfs_truncate_file(mp, qip);
273
			IRELE(qip);
L
Linus Torvalds 已提交
274 275 276
		}
	}

277 278
	if ((flags & (XFS_DQ_GROUP|XFS_DQ_PROJ)) &&
	    mp->m_sb.sb_gquotino != NULLFSINO) {
279 280 281
		error2 = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &qip, 0);
		if (!error2) {
			error2 = xfs_truncate_file(mp, qip);
282
			IRELE(qip);
L
Linus Torvalds 已提交
283 284 285
		}
	}

286
	return error ? error : error2;
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294
}


/*
 * Switch on (a given) quota enforcement for a filesystem.  This takes
 * effect immediately.
 * (Switching on quota accounting must be done at mount time.)
 */
295
int
L
Linus Torvalds 已提交
296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
xfs_qm_scall_quotaon(
	xfs_mount_t	*mp,
	uint		flags)
{
	int		error;
	uint		qf;
	uint		accflags;
	__int64_t	sbflags;

	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
	/*
	 * Switching on quota accounting must be done at mount time.
	 */
	accflags = flags & XFS_ALL_QUOTA_ACCT;
	flags &= ~(XFS_ALL_QUOTA_ACCT);

	sbflags = 0;

	if (flags == 0) {
		qdprintk("quotaon: zero flags, m_qflags=%x\n", mp->m_qflags);
		return XFS_ERROR(EINVAL);
	}

	/* No fs can turn on quotas with a delayed effect */
	ASSERT((flags & XFS_ALL_QUOTA_ACCT) == 0);

	/*
	 * Can't enforce without accounting. We check the superblock
	 * qflags here instead of m_qflags because rootfs can have
	 * quota acct on ondisk without m_qflags' knowing.
	 */
	if (((flags & XFS_UQUOTA_ACCT) == 0 &&
	    (mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) == 0 &&
	    (flags & XFS_UQUOTA_ENFD))
	    ||
331 332
	    ((flags & XFS_PQUOTA_ACCT) == 0 &&
	    (mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) == 0 &&
333
	    (flags & XFS_GQUOTA_ACCT) == 0 &&
L
Linus Torvalds 已提交
334
	    (mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) == 0 &&
335
	    (flags & XFS_OQUOTA_ENFD))) {
L
Linus Torvalds 已提交
336 337 338 339 340 341 342 343 344 345 346 347 348 349
		qdprintk("Can't enforce without acct, flags=%x sbflags=%x\n",
			flags, mp->m_sb.sb_qflags);
		return XFS_ERROR(EINVAL);
	}
	/*
	 * If everything's upto-date incore, then don't waste time.
	 */
	if ((mp->m_qflags & flags) == flags)
		return XFS_ERROR(EEXIST);

	/*
	 * Change sb_qflags on disk but not incore mp->qflags
	 * if this is the root filesystem.
	 */
E
Eric Sandeen 已提交
350
	spin_lock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
351 352
	qf = mp->m_sb.sb_qflags;
	mp->m_sb.sb_qflags = qf | flags;
E
Eric Sandeen 已提交
353
	spin_unlock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368

	/*
	 * There's nothing to change if it's the same.
	 */
	if ((qf & flags) == flags && sbflags == 0)
		return XFS_ERROR(EEXIST);
	sbflags |= XFS_SB_QFLAGS;

	if ((error = xfs_qm_write_sb_changes(mp, sbflags)))
		return (error);
	/*
	 * If we aren't trying to switch on quota enforcement, we are done.
	 */
	if  (((mp->m_sb.sb_qflags & XFS_UQUOTA_ACCT) !=
	     (mp->m_qflags & XFS_UQUOTA_ACCT)) ||
369 370 371 372
	     ((mp->m_sb.sb_qflags & XFS_PQUOTA_ACCT) !=
	     (mp->m_qflags & XFS_PQUOTA_ACCT)) ||
	     ((mp->m_sb.sb_qflags & XFS_GQUOTA_ACCT) !=
	     (mp->m_qflags & XFS_GQUOTA_ACCT)) ||
L
Linus Torvalds 已提交
373 374 375 376 377 378 379 380 381
	    (flags & XFS_ALL_QUOTA_ENFD) == 0)
		return (0);

	if (! XFS_IS_QUOTA_RUNNING(mp))
		return XFS_ERROR(ESRCH);

	/*
	 * Switch on quota enforcement in core.
	 */
382
	mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
L
Linus Torvalds 已提交
383 384 385 386 387 388 389 390 391 392
	mp->m_qflags |= (flags & XFS_ALL_QUOTA_ENFD);
	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));

	return (0);
}


/*
 * Return quota status information, such as uquota-off, enforcements, etc.
 */
393
int
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401 402 403 404 405
xfs_qm_scall_getqstat(
	xfs_mount_t	*mp,
	fs_quota_stat_t *out)
{
	xfs_inode_t	*uip, *gip;
	boolean_t	tempuqip, tempgqip;

	uip = gip = NULL;
	tempuqip = tempgqip = B_FALSE;
	memset(out, 0, sizeof(fs_quota_stat_t));

	out->qs_version = FS_QSTAT_VERSION;
406
	if (!xfs_sb_version_hasquota(&mp->m_sb)) {
L
Linus Torvalds 已提交
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
		out->qs_uquota.qfs_ino = NULLFSINO;
		out->qs_gquota.qfs_ino = NULLFSINO;
		return (0);
	}
	out->qs_flags = (__uint16_t) xfs_qm_export_flags(mp->m_qflags &
							(XFS_ALL_QUOTA_ACCT|
							 XFS_ALL_QUOTA_ENFD));
	out->qs_pad = 0;
	out->qs_uquota.qfs_ino = mp->m_sb.sb_uquotino;
	out->qs_gquota.qfs_ino = mp->m_sb.sb_gquotino;

	if (mp->m_quotainfo) {
		uip = mp->m_quotainfo->qi_uquotaip;
		gip = mp->m_quotainfo->qi_gquotaip;
	}
	if (!uip && mp->m_sb.sb_uquotino != NULLFSINO) {
		if (xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
					0, 0, &uip, 0) == 0)
			tempuqip = B_TRUE;
	}
	if (!gip && mp->m_sb.sb_gquotino != NULLFSINO) {
		if (xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
					0, 0, &gip, 0) == 0)
			tempgqip = B_TRUE;
	}
	if (uip) {
		out->qs_uquota.qfs_nblks = uip->i_d.di_nblocks;
		out->qs_uquota.qfs_nextents = uip->i_d.di_nextents;
		if (tempuqip)
436
			IRELE(uip);
L
Linus Torvalds 已提交
437 438 439 440 441
	}
	if (gip) {
		out->qs_gquota.qfs_nblks = gip->i_d.di_nblocks;
		out->qs_gquota.qfs_nextents = gip->i_d.di_nextents;
		if (tempgqip)
442
			IRELE(gip);
L
Linus Torvalds 已提交
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
	}
	if (mp->m_quotainfo) {
		out->qs_incoredqs = XFS_QI_MPLNDQUOTS(mp);
		out->qs_btimelimit = XFS_QI_BTIMELIMIT(mp);
		out->qs_itimelimit = XFS_QI_ITIMELIMIT(mp);
		out->qs_rtbtimelimit = XFS_QI_RTBTIMELIMIT(mp);
		out->qs_bwarnlimit = XFS_QI_BWARNLIMIT(mp);
		out->qs_iwarnlimit = XFS_QI_IWARNLIMIT(mp);
	}
	return (0);
}

/*
 * Adjust quota limits, and start/stop timers accordingly.
 */
458
int
L
Linus Torvalds 已提交
459 460 461 462 463 464 465 466 467 468 469 470
xfs_qm_scall_setqlim(
	xfs_mount_t		*mp,
	xfs_dqid_t		id,
	uint			type,
	fs_disk_quota_t		*newlim)
{
	xfs_disk_dquot_t	*ddq;
	xfs_dquot_t		*dqp;
	xfs_trans_t		*tp;
	int			error;
	xfs_qcnt_t		hard, soft;

471 472
	if ((newlim->d_fieldmask &
	    (FS_DQ_LIMIT_MASK|FS_DQ_TIMER_MASK|FS_DQ_WARNS_MASK)) == 0)
L
Linus Torvalds 已提交
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487
		return (0);

	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SETQLIM);
	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_disk_dquot_t) + 128,
				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
		xfs_trans_cancel(tp, 0);
		return (error);
	}

	/*
	 * We don't want to race with a quotaoff so take the quotaoff lock.
	 * (We don't hold an inode lock, so there's nothing else to stop
	 * a quotaoff from happening). (XXXThis doesn't currently happen
	 * because we take the vfslock before calling xfs_qm_sysent).
	 */
488
	mutex_lock(&(XFS_QI_QOFFLOCK(mp)));
L
Linus Torvalds 已提交
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508

	/*
	 * Get the dquot (locked), and join it to the transaction.
	 * Allocate the dquot if this doesn't exist.
	 */
	if ((error = xfs_qm_dqget(mp, NULL, id, type, XFS_QMOPT_DQALLOC, &dqp))) {
		xfs_trans_cancel(tp, XFS_TRANS_ABORT);
		mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));
		ASSERT(error != ENOENT);
		return (error);
	}
	xfs_dqtrace_entry(dqp, "Q_SETQLIM: AFT DQGET");
	xfs_trans_dqjoin(tp, dqp);
	ddq = &dqp->q_core;

	/*
	 * Make sure that hardlimits are >= soft limits before changing.
	 */
	hard = (newlim->d_fieldmask & FS_DQ_BHARD) ?
		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_hardlimit) :
509
			be64_to_cpu(ddq->d_blk_hardlimit);
L
Linus Torvalds 已提交
510 511
	soft = (newlim->d_fieldmask & FS_DQ_BSOFT) ?
		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_blk_softlimit) :
512
			be64_to_cpu(ddq->d_blk_softlimit);
L
Linus Torvalds 已提交
513
	if (hard == 0 || hard >= soft) {
514 515
		ddq->d_blk_hardlimit = cpu_to_be64(hard);
		ddq->d_blk_softlimit = cpu_to_be64(soft);
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524
		if (id == 0) {
			mp->m_quotainfo->qi_bhardlimit = hard;
			mp->m_quotainfo->qi_bsoftlimit = soft;
		}
	} else {
		qdprintk("blkhard %Ld < blksoft %Ld\n", hard, soft);
	}
	hard = (newlim->d_fieldmask & FS_DQ_RTBHARD) ?
		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_hardlimit) :
525
			be64_to_cpu(ddq->d_rtb_hardlimit);
L
Linus Torvalds 已提交
526 527
	soft = (newlim->d_fieldmask & FS_DQ_RTBSOFT) ?
		(xfs_qcnt_t) XFS_BB_TO_FSB(mp, newlim->d_rtb_softlimit) :
528
			be64_to_cpu(ddq->d_rtb_softlimit);
L
Linus Torvalds 已提交
529
	if (hard == 0 || hard >= soft) {
530 531
		ddq->d_rtb_hardlimit = cpu_to_be64(hard);
		ddq->d_rtb_softlimit = cpu_to_be64(soft);
L
Linus Torvalds 已提交
532 533 534 535 536 537 538 539 540 541
		if (id == 0) {
			mp->m_quotainfo->qi_rtbhardlimit = hard;
			mp->m_quotainfo->qi_rtbsoftlimit = soft;
		}
	} else {
		qdprintk("rtbhard %Ld < rtbsoft %Ld\n", hard, soft);
	}

	hard = (newlim->d_fieldmask & FS_DQ_IHARD) ?
		(xfs_qcnt_t) newlim->d_ino_hardlimit :
542
			be64_to_cpu(ddq->d_ino_hardlimit);
L
Linus Torvalds 已提交
543 544
	soft = (newlim->d_fieldmask & FS_DQ_ISOFT) ?
		(xfs_qcnt_t) newlim->d_ino_softlimit :
545
			be64_to_cpu(ddq->d_ino_softlimit);
L
Linus Torvalds 已提交
546
	if (hard == 0 || hard >= soft) {
547 548
		ddq->d_ino_hardlimit = cpu_to_be64(hard);
		ddq->d_ino_softlimit = cpu_to_be64(soft);
L
Linus Torvalds 已提交
549 550 551 552 553 554 555 556
		if (id == 0) {
			mp->m_quotainfo->qi_ihardlimit = hard;
			mp->m_quotainfo->qi_isoftlimit = soft;
		}
	} else {
		qdprintk("ihard %Ld < isoft %Ld\n", hard, soft);
	}

557 558 559 560
	/*
	 * Update warnings counter(s) if requested
	 */
	if (newlim->d_fieldmask & FS_DQ_BWARNS)
561
		ddq->d_bwarns = cpu_to_be16(newlim->d_bwarns);
562
	if (newlim->d_fieldmask & FS_DQ_IWARNS)
563
		ddq->d_iwarns = cpu_to_be16(newlim->d_iwarns);
564
	if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
565
		ddq->d_rtbwarns = cpu_to_be16(newlim->d_rtbwarns);
566

L
Linus Torvalds 已提交
567 568 569 570 571
	if (id == 0) {
		/*
		 * Timelimits for the super user set the relative time
		 * the other users can be over quota for this file system.
		 * If it is zero a default is used.  Ditto for the default
572 573
		 * soft and hard limit values (already done, above), and
		 * for warnings.
L
Linus Torvalds 已提交
574 575 576
		 */
		if (newlim->d_fieldmask & FS_DQ_BTIMER) {
			mp->m_quotainfo->qi_btimelimit = newlim->d_btimer;
577
			ddq->d_btimer = cpu_to_be32(newlim->d_btimer);
L
Linus Torvalds 已提交
578 579 580
		}
		if (newlim->d_fieldmask & FS_DQ_ITIMER) {
			mp->m_quotainfo->qi_itimelimit = newlim->d_itimer;
581
			ddq->d_itimer = cpu_to_be32(newlim->d_itimer);
L
Linus Torvalds 已提交
582 583 584
		}
		if (newlim->d_fieldmask & FS_DQ_RTBTIMER) {
			mp->m_quotainfo->qi_rtbtimelimit = newlim->d_rtbtimer;
585
			ddq->d_rtbtimer = cpu_to_be32(newlim->d_rtbtimer);
L
Linus Torvalds 已提交
586
		}
587 588 589 590 591 592 593
		if (newlim->d_fieldmask & FS_DQ_BWARNS)
			mp->m_quotainfo->qi_bwarnlimit = newlim->d_bwarns;
		if (newlim->d_fieldmask & FS_DQ_IWARNS)
			mp->m_quotainfo->qi_iwarnlimit = newlim->d_iwarns;
		if (newlim->d_fieldmask & FS_DQ_RTBWARNS)
			mp->m_quotainfo->qi_rtbwarnlimit = newlim->d_rtbwarns;
	} else {
L
Linus Torvalds 已提交
594 595 596 597 598 599 600 601 602 603 604 605 606
		/*
		 * If the user is now over quota, start the timelimit.
		 * The user will not be 'warned'.
		 * Note that we keep the timers ticking, whether enforcement
		 * is on or off. We don't really want to bother with iterating
		 * over all ondisk dquots and turning the timers on/off.
		 */
		xfs_qm_adjust_dqtimers(mp, ddq);
	}
	dqp->dq_flags |= XFS_DQ_DIRTY;
	xfs_trans_log_dquot(tp, dqp);

	xfs_dqtrace_entry(dqp, "Q_SETQLIM: COMMIT");
607
	error = xfs_trans_commit(tp, 0);
L
Linus Torvalds 已提交
608 609 610 611
	xfs_qm_dqprint(dqp);
	xfs_qm_dqrele(dqp);
	mutex_unlock(&(XFS_QI_QOFFLOCK(mp)));

612
	return error;
L
Linus Torvalds 已提交
613 614
}

615
int
L
Linus Torvalds 已提交
616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658
xfs_qm_scall_getquota(
	xfs_mount_t	*mp,
	xfs_dqid_t	id,
	uint		type,
	fs_disk_quota_t *out)
{
	xfs_dquot_t	*dqp;
	int		error;

	/*
	 * Try to get the dquot. We don't want it allocated on disk, so
	 * we aren't passing the XFS_QMOPT_DOALLOC flag. If it doesn't
	 * exist, we'll get ENOENT back.
	 */
	if ((error = xfs_qm_dqget(mp, NULL, id, type, 0, &dqp))) {
		return (error);
	}

	xfs_dqtrace_entry(dqp, "Q_GETQUOTA SUCCESS");
	/*
	 * If everything's NULL, this dquot doesn't quite exist as far as
	 * our utility programs are concerned.
	 */
	if (XFS_IS_DQUOT_UNINITIALIZED(dqp)) {
		xfs_qm_dqput(dqp);
		return XFS_ERROR(ENOENT);
	}
	/* xfs_qm_dqprint(dqp); */
	/*
	 * Convert the disk dquot to the exportable format
	 */
	xfs_qm_export_dquot(mp, &dqp->q_core, out);
	xfs_qm_dqput(dqp);
	return (error ? XFS_ERROR(EFAULT) : 0);
}


STATIC int
xfs_qm_log_quotaoff_end(
	xfs_mount_t		*mp,
	xfs_qoff_logitem_t	*startqoff,
	uint			flags)
{
659
	xfs_trans_t		*tp;
L
Linus Torvalds 已提交
660
	int			error;
661
	xfs_qoff_logitem_t	*qoffi;
L
Linus Torvalds 已提交
662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680

	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF_END);

	if ((error = xfs_trans_reserve(tp, 0, sizeof(xfs_qoff_logitem_t) * 2,
				      0, 0, XFS_DEFAULT_LOG_COUNT))) {
		xfs_trans_cancel(tp, 0);
		return (error);
	}

	qoffi = xfs_trans_get_qoff_item(tp, startqoff,
					flags & XFS_ALL_QUOTA_ACCT);
	xfs_trans_log_quotaoff_item(tp, qoffi);

	/*
	 * We have to make sure that the transaction is secure on disk before we
	 * return and actually stop quota accounting. So, make it synchronous.
	 * We don't care about quotoff's performance.
	 */
	xfs_trans_set_sync(tp);
681
	error = xfs_trans_commit(tp, 0);
L
Linus Torvalds 已提交
682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709
	return (error);
}


STATIC int
xfs_qm_log_quotaoff(
	xfs_mount_t	       *mp,
	xfs_qoff_logitem_t     **qoffstartp,
	uint		       flags)
{
	xfs_trans_t	       *tp;
	int			error;
	xfs_qoff_logitem_t     *qoffi=NULL;
	uint			oldsbqflag=0;

	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
	if ((error = xfs_trans_reserve(tp, 0,
				      sizeof(xfs_qoff_logitem_t) * 2 +
				      mp->m_sb.sb_sectsize + 128,
				      0,
				      0,
				      XFS_DEFAULT_LOG_COUNT))) {
		goto error0;
	}

	qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
	xfs_trans_log_quotaoff_item(tp, qoffi);

E
Eric Sandeen 已提交
710
	spin_lock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
711 712
	oldsbqflag = mp->m_sb.sb_qflags;
	mp->m_sb.sb_qflags = (mp->m_qflags & ~(flags)) & XFS_MOUNT_QUOTA_ALL;
E
Eric Sandeen 已提交
713
	spin_unlock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
714 715 716 717 718 719 720 721 722

	xfs_mod_sb(tp, XFS_SB_QFLAGS);

	/*
	 * We have to make sure that the transaction is secure on disk before we
	 * return and actually stop quota accounting. So, make it synchronous.
	 * We don't care about quotoff's performance.
	 */
	xfs_trans_set_sync(tp);
723
	error = xfs_trans_commit(tp, 0);
L
Linus Torvalds 已提交
724 725 726 727 728 729 730 731

error0:
	if (error) {
		xfs_trans_cancel(tp, 0);
		/*
		 * No one else is modifying sb_qflags, so this is OK.
		 * We still hold the quotaofflock.
		 */
E
Eric Sandeen 已提交
732
		spin_lock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
733
		mp->m_sb.sb_qflags = oldsbqflag;
E
Eric Sandeen 已提交
734
		spin_unlock(&mp->m_sb_lock);
L
Linus Torvalds 已提交
735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
	}
	*qoffstartp = qoffi;
	return (error);
}


/*
 * Translate an internal style on-disk-dquot to the exportable format.
 * The main differences are that the counters/limits are all in Basic
 * Blocks (BBs) instead of the internal FSBs, and all on-disk data has
 * to be converted to the native endianness.
 */
STATIC void
xfs_qm_export_dquot(
	xfs_mount_t		*mp,
	xfs_disk_dquot_t	*src,
	struct fs_disk_quota	*dst)
{
	memset(dst, 0, sizeof(*dst));
	dst->d_version = FS_DQUOT_VERSION;  /* different from src->d_version */
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
	dst->d_flags = xfs_qm_export_qtype_flags(src->d_flags);
	dst->d_id = be32_to_cpu(src->d_id);
	dst->d_blk_hardlimit =
		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_hardlimit));
	dst->d_blk_softlimit =
		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_blk_softlimit));
	dst->d_ino_hardlimit = be64_to_cpu(src->d_ino_hardlimit);
	dst->d_ino_softlimit = be64_to_cpu(src->d_ino_softlimit);
	dst->d_bcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_bcount));
	dst->d_icount = be64_to_cpu(src->d_icount);
	dst->d_btimer = be32_to_cpu(src->d_btimer);
	dst->d_itimer = be32_to_cpu(src->d_itimer);
	dst->d_iwarns = be16_to_cpu(src->d_iwarns);
	dst->d_bwarns = be16_to_cpu(src->d_bwarns);
	dst->d_rtb_hardlimit =
		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_hardlimit));
	dst->d_rtb_softlimit =
		XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtb_softlimit));
	dst->d_rtbcount = XFS_FSB_TO_BB(mp, be64_to_cpu(src->d_rtbcount));
	dst->d_rtbtimer = be32_to_cpu(src->d_rtbtimer);
	dst->d_rtbwarns = be16_to_cpu(src->d_rtbwarns);
L
Linus Torvalds 已提交
776 777 778

	/*
	 * Internally, we don't reset all the timers when quota enforcement
779
	 * gets turned off. No need to confuse the user level code,
L
Linus Torvalds 已提交
780 781
	 * so return zeroes in that case.
	 */
782 783 784
	if ((!XFS_IS_UQUOTA_ENFORCED(mp) && src->d_flags == XFS_DQ_USER) ||
	    (!XFS_IS_OQUOTA_ENFORCED(mp) &&
			(src->d_flags & (XFS_DQ_PROJ | XFS_DQ_GROUP)))) {
L
Linus Torvalds 已提交
785 786 787 788 789 790
		dst->d_btimer = 0;
		dst->d_itimer = 0;
		dst->d_rtbtimer = 0;
	}

#ifdef DEBUG
791 792 793 794
	if (((XFS_IS_UQUOTA_ENFORCED(mp) && dst->d_flags == XFS_USER_QUOTA) ||
	     (XFS_IS_OQUOTA_ENFORCED(mp) &&
			(dst->d_flags & (XFS_PROJ_QUOTA | XFS_GROUP_QUOTA)))) &&
	    dst->d_id != 0) {
L
Linus Torvalds 已提交
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811
		if (((int) dst->d_bcount >= (int) dst->d_blk_softlimit) &&
		    (dst->d_blk_softlimit > 0)) {
			ASSERT(dst->d_btimer != 0);
		}
		if (((int) dst->d_icount >= (int) dst->d_ino_softlimit) &&
		    (dst->d_ino_softlimit > 0)) {
			ASSERT(dst->d_itimer != 0);
		}
	}
#endif
}

STATIC uint
xfs_qm_export_qtype_flags(
	uint flags)
{
	/*
812
	 * Can't be more than one, or none.
L
Linus Torvalds 已提交
813
	 */
814 815 816 817 818 819 820
	ASSERT((flags & (XFS_PROJ_QUOTA | XFS_USER_QUOTA)) !=
		(XFS_PROJ_QUOTA | XFS_USER_QUOTA));
	ASSERT((flags & (XFS_PROJ_QUOTA | XFS_GROUP_QUOTA)) !=
		(XFS_PROJ_QUOTA | XFS_GROUP_QUOTA));
	ASSERT((flags & (XFS_USER_QUOTA | XFS_GROUP_QUOTA)) !=
		(XFS_USER_QUOTA | XFS_GROUP_QUOTA));
	ASSERT((flags & (XFS_PROJ_QUOTA|XFS_USER_QUOTA|XFS_GROUP_QUOTA)) != 0);
L
Linus Torvalds 已提交
821 822

	return (flags & XFS_DQ_USER) ?
823 824
		XFS_USER_QUOTA : (flags & XFS_DQ_PROJ) ?
			XFS_PROJ_QUOTA : XFS_GROUP_QUOTA;
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832 833 834 835
}

STATIC uint
xfs_qm_export_flags(
	uint flags)
{
	uint uflags;

	uflags = 0;
	if (flags & XFS_UQUOTA_ACCT)
		uflags |= XFS_QUOTA_UDQ_ACCT;
836 837
	if (flags & XFS_PQUOTA_ACCT)
		uflags |= XFS_QUOTA_PDQ_ACCT;
L
Linus Torvalds 已提交
838 839 840 841
	if (flags & XFS_GQUOTA_ACCT)
		uflags |= XFS_QUOTA_GDQ_ACCT;
	if (flags & XFS_UQUOTA_ENFD)
		uflags |= XFS_QUOTA_UDQ_ENFD;
842 843 844 845
	if (flags & (XFS_OQUOTA_ENFD)) {
		uflags |= (flags & XFS_GQUOTA_ACCT) ?
			XFS_QUOTA_GDQ_ENFD : XFS_QUOTA_PDQ_ENFD;
	}
L
Linus Torvalds 已提交
846 847 848 849 850
	return (uflags);
}


/*
851
 * Release all the dquots on the inodes in an AG.
L
Linus Torvalds 已提交
852
 */
853 854 855 856 857
STATIC void
xfs_qm_dqrele_inodes_ag(
	xfs_mount_t	*mp,
	int		ag,
	uint		flags)
L
Linus Torvalds 已提交
858
{
859 860 861 862
	xfs_inode_t	*ip = NULL;
	xfs_perag_t	*pag = &mp->m_perag[ag];
	int		first_index = 0;
	int		nr_found;
L
Linus Torvalds 已提交
863 864

	do {
865 866 867 868 869 870 871 872 873 874 875 876
		/*
		 * use a gang lookup to find the next inode in the tree
		 * as the tree is sparse and a gang lookup walks to find
		 * the number of objects requested.
		 */
		read_lock(&pag->pag_ici_lock);
		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
				(void**)&ip, first_index, 1);

		if (!nr_found) {
			read_unlock(&pag->pag_ici_lock);
			break;
L
Linus Torvalds 已提交
877
		}
878

879 880 881 882 883 884
		/*
		 * Update the index for the next lookup. Catch overflows
		 * into the next AG range which can occur if we have inodes
		 * in the last block of the AG and we are currently
		 * pointing to the last inode.
		 */
885
		first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
886 887 888 889
		if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino)) {
			read_unlock(&pag->pag_ici_lock);
			break;
		}
890

891 892
		/* skip quota inodes */
		if (ip == XFS_QI_UQIP(mp) || ip == XFS_QI_GQIP(mp)) {
L
Linus Torvalds 已提交
893 894
			ASSERT(ip->i_udquot == NULL);
			ASSERT(ip->i_gdquot == NULL);
895
			read_unlock(&pag->pag_ici_lock);
L
Linus Torvalds 已提交
896 897
			continue;
		}
898 899 900 901 902 903

		/*
		 * If we can't get a reference on the inode, it must be
		 * in reclaim. Leave it for the reclaim code to flush.
		 */
		if (!igrab(VFS_I(ip))) {
904
			read_unlock(&pag->pag_ici_lock);
905
			continue;
L
Linus Torvalds 已提交
906
		}
907 908
		read_unlock(&pag->pag_ici_lock);

909 910 911 912 913 914
		/* avoid new inodes though we shouldn't find any here */
		if (xfs_iflags_test(ip, XFS_INEW)) {
			IRELE(ip);
			continue;
		}

915
		xfs_ilock(ip, XFS_ILOCK_EXCL);
L
Linus Torvalds 已提交
916 917 918 919
		if ((flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
			xfs_qm_dqrele(ip->i_udquot);
			ip->i_udquot = NULL;
		}
920 921
		if (flags & (XFS_PQUOTA_ACCT|XFS_GQUOTA_ACCT) &&
		    ip->i_gdquot) {
L
Linus Torvalds 已提交
922 923 924
			xfs_qm_dqrele(ip->i_gdquot);
			ip->i_gdquot = NULL;
		}
925 926
		xfs_iput(ip, XFS_ILOCK_EXCL);

927 928 929 930 931 932 933 934 935 936 937 938 939 940 941
	} while (nr_found);
}

/*
 * Go thru all the inodes in the file system, releasing their dquots.
 * Note that the mount structure gets modified to indicate that quotas are off
 * AFTER this, in the case of quotaoff. This also gets called from
 * xfs_rootumount.
 */
void
xfs_qm_dqrele_all_inodes(
	struct xfs_mount *mp,
	uint		 flags)
{
	int		i;
L
Linus Torvalds 已提交
942

943 944 945 946 947 948
	ASSERT(mp->m_quotainfo);
	for (i = 0; i < mp->m_sb.sb_agcount; i++) {
		if (!mp->m_perag[i].pag_ici_init)
			continue;
		xfs_qm_dqrele_inodes_ag(mp, i, flags);
	}
L
Linus Torvalds 已提交
949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 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
}

/*------------------------------------------------------------------------*/
#ifdef DEBUG
/*
 * This contains all the test functions for XFS disk quotas.
 * Currently it does a quota accounting check. ie. it walks through
 * all inodes in the file system, calculating the dquot accounting fields,
 * and prints out any inconsistencies.
 */
xfs_dqhash_t *qmtest_udqtab;
xfs_dqhash_t *qmtest_gdqtab;
int	      qmtest_hashmask;
int	      qmtest_nfails;
mutex_t	      qcheck_lock;

#define DQTEST_HASHVAL(mp, id) (((__psunsigned_t)(mp) + \
				 (__psunsigned_t)(id)) & \
				(qmtest_hashmask - 1))

#define DQTEST_HASH(mp, id, type)   ((type & XFS_DQ_USER) ? \
				     (qmtest_udqtab + \
				      DQTEST_HASHVAL(mp, id)) : \
				     (qmtest_gdqtab + \
				      DQTEST_HASHVAL(mp, id)))

#define DQTEST_LIST_PRINT(l, NXT, title) \
{ \
	  xfs_dqtest_t	*dqp; int i = 0;\
	  cmn_err(CE_DEBUG, "%s (#%d)", title, (int) (l)->qh_nelems); \
	  for (dqp = (xfs_dqtest_t *)(l)->qh_next; dqp != NULL; \
	       dqp = (xfs_dqtest_t *)dqp->NXT) { \
		cmn_err(CE_DEBUG, "  %d. \"%d (%s)\"  bcnt = %d, icnt = %d", \
			 ++i, dqp->d_id, DQFLAGTO_TYPESTR(dqp),	     \
			 dqp->d_bcount, dqp->d_icount); } \
}

typedef struct dqtest {
	xfs_dqmarker_t	q_lists;
	xfs_dqhash_t	*q_hash;	/* the hashchain header */
	xfs_mount_t	*q_mount;	/* filesystem this relates to */
	xfs_dqid_t	d_id;		/* user id or group id */
	xfs_qcnt_t	d_bcount;	/* # disk blocks owned by the user */
	xfs_qcnt_t	d_icount;	/* # inodes owned by the user */
} xfs_dqtest_t;

STATIC void
xfs_qm_hashinsert(xfs_dqhash_t *h, xfs_dqtest_t *dqp)
{
	xfs_dquot_t *d;
	if (((d) = (h)->qh_next))
		(d)->HL_PREVP = &((dqp)->HL_NEXT);
	(dqp)->HL_NEXT = d;
	(dqp)->HL_PREVP = &((h)->qh_next);
	(h)->qh_next = (xfs_dquot_t *)dqp;
	(h)->qh_version++;
	(h)->qh_nelems++;
}
STATIC void
xfs_qm_dqtest_print(
	xfs_dqtest_t	*d)
{
	cmn_err(CE_DEBUG, "-----------DQTEST DQUOT----------------");
	cmn_err(CE_DEBUG, "---- dquot ID = %d", d->d_id);
	cmn_err(CE_DEBUG, "---- fs       = 0x%p", d->q_mount);
	cmn_err(CE_DEBUG, "---- bcount   = %Lu (0x%x)",
		d->d_bcount, (int)d->d_bcount);
	cmn_err(CE_DEBUG, "---- icount   = %Lu (0x%x)",
		d->d_icount, (int)d->d_icount);
	cmn_err(CE_DEBUG, "---------------------------");
}

STATIC void
xfs_qm_dqtest_failed(
	xfs_dqtest_t	*d,
	xfs_dquot_t	*dqp,
	char		*reason,
	xfs_qcnt_t	a,
	xfs_qcnt_t	b,
	int		error)
{
	qmtest_nfails++;
	if (error)
		cmn_err(CE_DEBUG, "quotacheck failed id=%d, err=%d\nreason: %s",
1033
		       d->d_id, error, reason);
L
Linus Torvalds 已提交
1034 1035
	else
		cmn_err(CE_DEBUG, "quotacheck failed id=%d (%s) [%d != %d]",
1036
		       d->d_id, reason, (int)a, (int)b);
L
Linus Torvalds 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
	xfs_qm_dqtest_print(d);
	if (dqp)
		xfs_qm_dqprint(dqp);
}

STATIC int
xfs_dqtest_cmp2(
	xfs_dqtest_t	*d,
	xfs_dquot_t	*dqp)
{
	int err = 0;
1048
	if (be64_to_cpu(dqp->q_core.d_icount) != d->d_icount) {
L
Linus Torvalds 已提交
1049
		xfs_qm_dqtest_failed(d, dqp, "icount mismatch",
1050
			be64_to_cpu(dqp->q_core.d_icount),
L
Linus Torvalds 已提交
1051 1052 1053
			d->d_icount, 0);
		err++;
	}
1054
	if (be64_to_cpu(dqp->q_core.d_bcount) != d->d_bcount) {
L
Linus Torvalds 已提交
1055
		xfs_qm_dqtest_failed(d, dqp, "bcount mismatch",
1056
			be64_to_cpu(dqp->q_core.d_bcount),
L
Linus Torvalds 已提交
1057 1058 1059
			d->d_bcount, 0);
		err++;
	}
1060 1061 1062
	if (dqp->q_core.d_blk_softlimit &&
	    be64_to_cpu(dqp->q_core.d_bcount) >=
	    be64_to_cpu(dqp->q_core.d_blk_softlimit)) {
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069
		if (!dqp->q_core.d_btimer && dqp->q_core.d_id) {
			cmn_err(CE_DEBUG,
				"%d [%s] [0x%p] BLK TIMER NOT STARTED",
				d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
			err++;
		}
	}
1070 1071 1072
	if (dqp->q_core.d_ino_softlimit &&
	    be64_to_cpu(dqp->q_core.d_icount) >=
	    be64_to_cpu(dqp->q_core.d_ino_softlimit)) {
L
Linus Torvalds 已提交
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
		if (!dqp->q_core.d_itimer && dqp->q_core.d_id) {
			cmn_err(CE_DEBUG,
				"%d [%s] [0x%p] INO TIMER NOT STARTED",
				d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
			err++;
		}
	}
#ifdef QUOTADEBUG
	if (!err) {
		cmn_err(CE_DEBUG, "%d [%s] [0x%p] qchecked",
1083
			d->d_id, DQFLAGTO_TYPESTR(d), d->q_mount);
L
Linus Torvalds 已提交
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 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
	}
#endif
	return (err);
}

STATIC void
xfs_dqtest_cmp(
	xfs_dqtest_t	*d)
{
	xfs_dquot_t	*dqp;
	int		error;

	/* xfs_qm_dqtest_print(d); */
	if ((error = xfs_qm_dqget(d->q_mount, NULL, d->d_id, d->dq_flags, 0,
				 &dqp))) {
		xfs_qm_dqtest_failed(d, NULL, "dqget failed", 0, 0, error);
		return;
	}
	xfs_dqtest_cmp2(d, dqp);
	xfs_qm_dqput(dqp);
}

STATIC int
xfs_qm_internalqcheck_dqget(
	xfs_mount_t	*mp,
	xfs_dqid_t	id,
	uint		type,
	xfs_dqtest_t	**O_dq)
{
	xfs_dqtest_t	*d;
	xfs_dqhash_t	*h;

	h = DQTEST_HASH(mp, id, type);
	for (d = (xfs_dqtest_t *) h->qh_next; d != NULL;
	     d = (xfs_dqtest_t *) d->HL_NEXT) {
		/* DQTEST_LIST_PRINT(h, HL_NEXT, "@@@@@ dqtestlist @@@@@"); */
		if (d->d_id == id && mp == d->q_mount) {
			*O_dq = d;
			return (0);
		}
	}
	d = kmem_zalloc(sizeof(xfs_dqtest_t), KM_SLEEP);
	d->dq_flags = type;
	d->d_id = id;
	d->q_mount = mp;
	d->q_hash = h;
	xfs_qm_hashinsert(h, d);
	*O_dq = d;
	return (0);
}

STATIC void
xfs_qm_internalqcheck_get_dquots(
	xfs_mount_t	*mp,
	xfs_dqid_t	uid,
1139
	xfs_dqid_t	projid,
L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146 1147
	xfs_dqid_t	gid,
	xfs_dqtest_t	**ud,
	xfs_dqtest_t	**gd)
{
	if (XFS_IS_UQUOTA_ON(mp))
		xfs_qm_internalqcheck_dqget(mp, uid, XFS_DQ_USER, ud);
	if (XFS_IS_GQUOTA_ON(mp))
		xfs_qm_internalqcheck_dqget(mp, gid, XFS_DQ_GROUP, gd);
1148 1149
	else if (XFS_IS_PQUOTA_ON(mp))
		xfs_qm_internalqcheck_dqget(mp, projid, XFS_DQ_PROJ, gd);
L
Linus Torvalds 已提交
1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 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 1210
}


STATIC void
xfs_qm_internalqcheck_dqadjust(
	xfs_inode_t		*ip,
	xfs_dqtest_t		*d)
{
	d->d_icount++;
	d->d_bcount += (xfs_qcnt_t)ip->i_d.di_nblocks;
}

STATIC int
xfs_qm_internalqcheck_adjust(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	ino,		/* inode number to get data for */
	void		__user *buffer,	/* not used */
	int		ubsize,		/* not used */
	void		*private_data,	/* not used */
	xfs_daddr_t	bno,		/* starting block of inode cluster */
	int		*ubused,	/* not used */
	void		*dip,		/* not used */
	int		*res)		/* bulkstat result code */
{
	xfs_inode_t		*ip;
	xfs_dqtest_t		*ud, *gd;
	uint			lock_flags;
	boolean_t		ipreleased;
	int			error;

	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) {
		*res = BULKSTAT_RV_NOTHING;
		qdprintk("internalqcheck: ino=%llu, uqino=%llu, gqino=%llu\n",
			(unsigned long long) ino,
			(unsigned long long) mp->m_sb.sb_uquotino,
			(unsigned long long) mp->m_sb.sb_gquotino);
		return XFS_ERROR(EINVAL);
	}
	ipreleased = B_FALSE;
 again:
	lock_flags = XFS_ILOCK_SHARED;
	if ((error = xfs_iget(mp, NULL, ino, 0, lock_flags, &ip, bno))) {
		*res = BULKSTAT_RV_NOTHING;
		return (error);
	}

	/*
	 * This inode can have blocks after eof which can get released
	 * when we send it to inactive. Since we don't check the dquot
	 * until the after all our calculations are done, we must get rid
	 * of those now.
	 */
	if (! ipreleased) {
		xfs_iput(ip, lock_flags);
		ipreleased = B_TRUE;
		goto again;
	}
	xfs_qm_internalqcheck_get_dquots(mp,
					(xfs_dqid_t) ip->i_d.di_uid,
1211
					(xfs_dqid_t) ip->i_d.di_projid,
L
Linus Torvalds 已提交
1212 1213 1214 1215 1216 1217
					(xfs_dqid_t) ip->i_d.di_gid,
					&ud, &gd);
	if (XFS_IS_UQUOTA_ON(mp)) {
		ASSERT(ud);
		xfs_qm_internalqcheck_dqadjust(ip, ud);
	}
1218
	if (XFS_IS_OQUOTA_ON(mp)) {
L
Linus Torvalds 已提交
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
		ASSERT(gd);
		xfs_qm_internalqcheck_dqadjust(ip, gd);
	}
	xfs_iput(ip, lock_flags);
	*res = BULKSTAT_RV_DIDONE;
	return (0);
}


/* PRIVATE, debugging */
int
xfs_qm_internalqcheck(
	xfs_mount_t	*mp)
{
	xfs_ino_t	lastino;
	int		done, count;
	int		i;
	xfs_dqtest_t	*d, *e;
	xfs_dqhash_t	*h1;
	int		error;

	lastino = 0;
	qmtest_hashmask = 32;
	count = 5;
	done = 0;
	qmtest_nfails = 0;

	if (! XFS_IS_QUOTA_ON(mp))
		return XFS_ERROR(ESRCH);

	xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC);
	XFS_bflush(mp->m_ddev_targp);
	xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC);
	XFS_bflush(mp->m_ddev_targp);

1254
	mutex_lock(&qcheck_lock);
L
Linus Torvalds 已提交
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
	/* There should be absolutely no quota activity while this
	   is going on. */
	qmtest_udqtab = kmem_zalloc(qmtest_hashmask *
				    sizeof(xfs_dqhash_t), KM_SLEEP);
	qmtest_gdqtab = kmem_zalloc(qmtest_hashmask *
				    sizeof(xfs_dqhash_t), KM_SLEEP);
	do {
		/*
		 * Iterate thru all the inodes in the file system,
		 * adjusting the corresponding dquot counters
		 */
		if ((error = xfs_bulkstat(mp, &lastino, &count,
				 xfs_qm_internalqcheck_adjust, NULL,
				 0, NULL, BULKSTAT_FG_IGET, &done))) {
			break;
		}
	} while (! done);
	if (error) {
		cmn_err(CE_DEBUG, "Bulkstat returned error 0x%x", error);
	}
	cmn_err(CE_DEBUG, "Checking results against system dquots");
	for (i = 0; i < qmtest_hashmask; i++) {
		h1 = &qmtest_udqtab[i];
		for (d = (xfs_dqtest_t *) h1->qh_next; d != NULL; ) {
			xfs_dqtest_cmp(d);
			e = (xfs_dqtest_t *) d->HL_NEXT;
1281
			kmem_free(d);
L
Linus Torvalds 已提交
1282 1283 1284 1285 1286 1287
			d = e;
		}
		h1 = &qmtest_gdqtab[i];
		for (d = (xfs_dqtest_t *) h1->qh_next; d != NULL; ) {
			xfs_dqtest_cmp(d);
			e = (xfs_dqtest_t *) d->HL_NEXT;
1288
			kmem_free(d);
L
Linus Torvalds 已提交
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
			d = e;
		}
	}

	if (qmtest_nfails) {
		cmn_err(CE_DEBUG, "******** quotacheck failed  ********");
		cmn_err(CE_DEBUG, "failures = %d", qmtest_nfails);
	} else {
		cmn_err(CE_DEBUG, "******** quotacheck successful! ********");
	}
1299 1300
	kmem_free(qmtest_udqtab);
	kmem_free(qmtest_gdqtab);
L
Linus Torvalds 已提交
1301 1302 1303 1304 1305
	mutex_unlock(&qcheck_lock);
	return (qmtest_nfails);
}

#endif /* DEBUG */