xfs_dquot.c

linux-2.6.15.6

/*
 * Copyright (c) 2000-2003 Silicon Graphics, Inc.
 * All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * 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.
 *
 * 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
 */
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_bit.h"
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir.h"
#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"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dir_sf.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_cap.h"
#include "xfs_mac.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
#include "xfs_trans_priv.h"
#include "xfs_qm.h"


/*
   LOCK ORDER

   inode lock		    (ilock)
   dquot hash-chain lock    (hashlock)
   xqm dquot freelist lock  (freelistlock
   mount's dquot list lock  (mplistlock)
   user dquot lock - lock ordering among dquots is based on the uid or gid
   group dquot lock - similar to udquots. Between the two dquots, the udquot
		      has to be locked first.
   pin lock - the dquot lock must be held to take this lock.
   flush lock - ditto.
*/

STATIC void		xfs_qm_dqflush_done(xfs_buf_t *, xfs_dq_logitem_t *);

#ifdef DEBUG
xfs_buftarg_t *xfs_dqerror_target;
int xfs_do_dqerror;
int xfs_dqreq_num;
int xfs_dqerror_mod = 33;
#endif

/*
 * Allocate and initialize a dquot. We don't always allocate fresh memory;
 * we try to reclaim a free dquot if the number of incore dquots are above
 * a threshold.
 * The only field inside the core that gets initialized at this point
 * is the d_id field. The idea is to fill in the entire q_core
 * when we read in the on disk dquot.
 */
STATIC xfs_dquot_t *
xfs_qm_dqinit(
	xfs_mount_t  *mp,
	xfs_dqid_t   id,
	uint	     type)
{
	xfs_dquot_t	*dqp;
	boolean_t	brandnewdquot;

	brandnewdquot = xfs_qm_dqalloc_incore(&dqp);
	dqp->dq_flags = type;
	dqp->q_core.d_id = cpu_to_be32(id);
	dqp->q_mount = mp;

	/*
	 * No need to re-initialize these if this is a reclaimed dquot.
	 */
	if (brandnewdquot) {
		dqp->dq_flnext = dqp->dq_flprev = dqp;
		mutex_init(&dqp->q_qlock,  MUTEX_DEFAULT, "xdq");
		initnsema(&dqp->q_flock, 1, "fdq");
		sv_init(&dqp->q_pinwait, SV_DEFAULT, "pdq");

#ifdef XFS_DQUOT_TRACE
		dqp->q_trace = ktrace_alloc(DQUOT_TRACE_SIZE, KM_SLEEP);
		xfs_dqtrace_entry(dqp, "DQINIT");
#endif
	} else {
		/*
		 * Only the q_core portion was zeroed in dqreclaim_one().
		 * So, we need to reset others.
		 */
		 dqp->q_nrefs = 0;
		 dqp->q_blkno = 0;
		 dqp->MPL_NEXT = dqp->HL_NEXT = NULL;
		 dqp->HL_PREVP = dqp->MPL_PREVP = NULL;
		 dqp->q_bufoffset = 0;
		 dqp->q_fileoffset = 0;
		 dqp->q_transp = NULL;
		 dqp->q_gdquot = NULL;
		 dqp->q_res_bcount = 0;
		 dqp->q_res_icount = 0;
		 dqp->q_res_rtbcount = 0;
		 dqp->q_pincount = 0;
		 dqp->q_hash = NULL;
		 ASSERT(dqp->dq_flnext == dqp->dq_flprev);

#ifdef XFS_DQUOT_TRACE
		 ASSERT(dqp->q_trace);
		 xfs_dqtrace_entry(dqp, "DQRECLAIMED_INIT");
#endif
	 }

	/*
	 * log item gets initialized later
	 */
	return (dqp);
}

/*
 * This is called to free all the memory associated with a dquot
 */
void
xfs_qm_dqdestroy(
	xfs_dquot_t	*dqp)
{
	ASSERT(! XFS_DQ_IS_ON_FREELIST(dqp));

	mutex_destroy(&dqp->q_qlock);
	freesema(&dqp->q_flock);
	sv_destroy(&dqp->q_pinwait);

#ifdef XFS_DQUOT_TRACE
	if (dqp->q_trace)
	     ktrace_free(dqp->q_trace);
	dqp->q_trace = NULL;
#endif
	kmem_zone_free(xfs_Gqm->qm_dqzone, dqp);
	atomic_dec(&xfs_Gqm->qm_totaldquots);
}

/*
 * This is what a 'fresh' dquot inside a dquot chunk looks like on disk.
 */
STATIC void
xfs_qm_dqinit_core(
	xfs_dqid_t	id,
	uint		type,
	xfs_dqblk_t	*d)
{
	/*
	 * Caller has zero'd the entire dquot 'chunk' already.
	 */
	d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
	d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
	d->dd_diskdq.d_id = cpu_to_be32(id);
	d->dd_diskdq.d_flags = type;
}


#ifdef XFS_DQUOT_TRACE
/*
 * Dquot tracing for debugging.
 */
/* ARGSUSED */
void
__xfs_dqtrace_entry(
	xfs_dquot_t	*dqp,
	char		*func,
	void		*retaddr,
	xfs_inode_t	*ip)
{
	xfs_dquot_t	*udqp = NULL;
	xfs_ino_t	ino = 0;

	ASSERT(dqp->q_trace);
	if (ip) {
		ino = ip->i_ino;
		udqp = ip->i_udquot;
	}
	ktrace_enter(dqp->q_trace,
		     (void *)(__psint_t)DQUOT_KTRACE_ENTRY,
		     (void *)func,
		     (void *)(__psint_t)dqp->q_nrefs,
		     (void *)(__psint_t)dqp->dq_flags,
		     (void *)(__psint_t)dqp->q_res_bcount,
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_bcount),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_icount),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_hardlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_softlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_hardlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_softlimit),
		     (void *)(__psint_t)be32_to_cpu(dqp->q_core.d_id),
		     (void *)(__psint_t)current_pid(),
		     (void *)(__psint_t)ino,
		     (void *)(__psint_t)retaddr,
		     (void *)(__psint_t)udqp);
	return;
}
#endif


/*
 * If default limits are in force, push them into the dquot now.
 * We overwrite the dquot limits only if they are zero and this
 * is not the root dquot.
 */
void
xfs_qm_adjust_dqlimits(
	xfs_mount_t		*mp,
	xfs_disk_dquot_t	*d)
{
	xfs_quotainfo_t		*q = mp->m_quotainfo;

	ASSERT(d->d_id);

	if (q->qi_bsoftlimit && !d->d_blk_softlimit)
		d->d_blk_softlimit = cpu_to_be64(q->qi_bsoftlimit);
	if (q->qi_bhardlimit && !d->d_blk_hardlimit)
		d->d_blk_hardlimit = cpu_to_be64(q->qi_bhardlimit);
	if (q->qi_isoftlimit && !d->d_ino_softlimit)
		d->d_ino_softlimit = cpu_to_be64(q->qi_isoftlimit);
	if (q->qi_ihardlimit && !d->d_ino_hardlimit)
		d->d_ino_hardlimit = cpu_to_be64(q->qi_ihardlimit);
	if (q->qi_rtbsoftlimit && !d->d_rtb_softlimit)
		d->d_rtb_softlimit = cpu_to_be64(q->qi_rtbsoftlimit);
	if (q->qi_rtbhardlimit && !d->d_rtb_hardlimit)
		d->d_rtb_hardlimit = cpu_to_be64(q->qi_rtbhardlimit);
}

/*
 * Check the limits and timers of a dquot and start or reset timers
 * if necessary.
 * This gets called even when quota enforcement is OFF, which makes our
 * life a little less complicated. (We just don't reject any quota
 * reservations in that case, when enforcement is off).
 * We also return 0 as the values of the timers in Q_GETQUOTA calls, when
 * enforcement's off.
 * In contrast, warnings are a little different in that they don't
 * 'automatically' get started when limits get exceeded.  They do
 * get reset to zero, however, when we find the count to be under
 * the soft limit (they are only ever set non-zero via userspace).
 */
void
xfs_qm_adjust_dqtimers(
	xfs_mount_t		*mp,
	xfs_disk_dquot_t	*d)
{
	ASSERT(d->d_id);

#ifdef QUOTADEBUG
	if (d->d_blk_hardlimit)
		ASSERT(be64_to_cpu(d->d_blk_softlimit) <=
		       be64_to_cpu(d->d_blk_hardlimit));
	if (d->d_ino_hardlimit)
		ASSERT(be64_to_cpu(d->d_ino_softlimit) <=
		       be64_to_cpu(d->d_ino_hardlimit));
	if (d->d_rtb_hardlimit)
		ASSERT(be64_to_cpu(d->d_rtb_softlimit) <=
		       be64_to_cpu(d->d_rtb_hardlimit));
#endif
	if (!d->d_btimer) {
		if ((d->d_blk_softlimit &&
		     (be64_to_cpu(d->d_bcount) >=
		      be64_to_cpu(d->d_blk_softlimit))) ||
		    (d->d_blk_hardlimit &&
		     (be64_to_cpu(d->d_bcount) >=
		      be64_to_cpu(d->d_blk_hardlimit)))) {
			d->d_btimer = cpu_to_be32(get_seconds() +
					XFS_QI_BTIMELIMIT(mp));
		} else {
			d->d_bwarns = 0;
		}
	} else {
		if ((!d->d_blk_softlimit ||
		     (be64_to_cpu(d->d_bcount) <
		      be64_to_cpu(d->d_blk_softlimit))) &&
		    (!d->d_blk_hardlimit ||
		    (be64_to_cpu(d->d_bcount) <
		     be64_to_cpu(d->d_blk_hardlimit)))) {
			d->d_btimer = 0;
		}
	}

	if (!d->d_itimer) {
		if ((d->d_ino_softlimit &&
		     (be64_to_cpu(d->d_icount) >=
		      be64_to_cpu(d->d_ino_softlimit))) ||
		    (d->d_ino_hardlimit &&
		     (be64_to_cpu(d->d_icount) >=
		      be64_to_cpu(d->d_ino_hardlimit)))) {
			d->d_itimer = cpu_to_be32(get_seconds() +
					XFS_QI_ITIMELIMIT(mp));
		} else {
			d->d_iwarns = 0;
		}
	} else {
		if ((!d->d_ino_softlimit ||
		     (be64_to_cpu(d->d_icount) <
		      be64_to_cpu(d->d_ino_softlimit)))  &&
		    (!d->d_ino_hardlimit ||
		     (be64_to_cpu(d->d_icount) <
		      be64_to_cpu(d->d_ino_hardlimit)))) {
			d->d_itimer = 0;
		}
	}

	if (!d->d_rtbtimer) {
		if ((d->d_rtb_softlimit &&
		     (be64_to_cpu(d->d_rtbcount) >=
		      be64_to_cpu(d->d_rtb_softlimit))) ||
		    (d->d_rtb_hardlimit &&
		     (be64_to_cpu(d->d_rtbcount) >=
		      be64_to_cpu(d->d_rtb_hardlimit)))) {
			d->d_rtbtimer = cpu_to_be32(get_seconds() +
					XFS_QI_RTBTIMELIMIT(mp));
		} else {
			d->d_rtbwarns = 0;
		}
	} else {
		if ((!d->d_rtb_softlimit ||
		     (be64_to_cpu(d->d_rtbcount) <
		      be64_to_cpu(d->d_rtb_softlimit))) &&
		    (!d->d_rtb_hardlimit ||
		     (be64_to_cpu(d->d_rtbcount) <
		      be64_to_cpu(d->d_rtb_hardlimit)))) {
			d->d_rtbtimer = 0;
		}
	}
}

/*
 * initialize a buffer full of dquots and log the whole thing
 */
STATIC void
xfs_qm_init_dquot_blk(
	xfs_trans_t	*tp,
	xfs_mount_t	*mp,
	xfs_dqid_t	id,
	uint		type,
	xfs_buf_t	*bp)
{
	xfs_dqblk_t	*d;
	int		curid, i;

	ASSERT(tp);
	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);

	d = (xfs_dqblk_t *)XFS_BUF_PTR(bp);

	/*
	 * ID of the first dquot in the block - id's are zero based.
	 */
	curid = id - (id % XFS_QM_DQPERBLK(mp));
	ASSERT(curid >= 0);
	memset(d, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)));
	for (i = 0; i < XFS_QM_DQPERBLK(mp); i++, d++, curid++)
		xfs_qm_dqinit_core(curid, type, d);
	xfs_trans_dquot_buf(tp, bp,
			    (type & XFS_DQ_USER ? XFS_BLI_UDQUOT_BUF :
			    ((type & XFS_DQ_PROJ) ? XFS_BLI_PDQUOT_BUF :
			     XFS_BLI_GDQUOT_BUF)));
	xfs_trans_log_buf(tp, bp, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)) - 1);
}



/*
 * Allocate a block and fill it with dquots.
 * This is called when the bmapi finds a hole.
 */
STATIC int
xfs_qm_dqalloc(
	xfs_trans_t	**tpp,
	xfs_mount_t	*mp,
	xfs_dquot_t	*dqp,
	xfs_inode_t	*quotip,
	xfs_fileoff_t	offset_fsb,
	xfs_buf_t	**O_bpp)
{
	xfs_fsblock_t	firstblock;
	xfs_bmap_free_t flist;
	xfs_bmbt_irec_t map;
	int		nmaps, error, committed;
	xfs_buf_t	*bp;
	xfs_trans_t	*tp = *tpp;

	ASSERT(tp != NULL);
	xfs_dqtrace_entry(dqp, "DQALLOC");

	/*
	 * Initialize the bmap freelist prior to calling bmapi code.
	 */
	XFS_BMAP_INIT(&flist, &firstblock);
	xfs_ilock(quotip, XFS_ILOCK_EXCL);
	/*
	 * Return if this type of quotas is turned off while we didn't
	 * have an inode lock
	 */
	if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
		xfs_iunlock(quotip, XFS_ILOCK_EXCL);
		return (ESRCH);
	}

	/*
	 * xfs_trans_commit normally decrements the vnode ref count
	 * when it unlocks the inode. Since we want to keep the quota
	 * inode around, we bump the vnode ref count now.
	 */
	VN_HOLD(XFS_ITOV(quotip));

	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
	nmaps = 1;
	if ((error = xfs_bmapi(tp, quotip,
			      offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB,
			      XFS_BMAPI_METADATA | XFS_BMAPI_WRITE,
			      &firstblock,
			      XFS_QM_DQALLOC_SPACE_RES(mp),
			      &map, &nmaps, &flist))) {
		goto error0;
	}
	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
	ASSERT(nmaps == 1);
	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	       (map.br_startblock != HOLESTARTBLOCK));

	/*
	 * Keep track of the blkno to save a lookup later
	 */
	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);

	/* now we can just get the buffer (there's nothing to read yet) */
	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
			       dqp->q_blkno,
			       XFS_QI_DQCHUNKLEN(mp),
			       0);
	if (!bp || (error = XFS_BUF_GETERROR(bp)))
		goto error1;
	/*
	 * Make a chunk of dquots out of this buffer and log
	 * the entire thing.
	 */
	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);

	/*
	 * xfs_bmap_finish() may commit the current transaction and
	 * start a second transaction if the freelist is not empty.
	 *
	 * Since we still want to modify this buffer, we need to
	 * ensure that the buffer is not released on commit of
	 * the first transaction and ensure the buffer is added to the
	 * second transaction.
	 *
	 * If there is only one transaction then don't stop the buffer
	 * from being released when it commits later on.
	 */

	xfs_trans_bhold(tp, bp);

	if ((error = xfs_bmap_finish(tpp, &flist, firstblock, &committed))) {
		goto error1;
	}

	if (committed) {
		tp = *tpp;
		xfs_trans_bjoin(tp, bp);
	} else {
		xfs_trans_bhold_release(tp, bp);
	}

	*O_bpp = bp;
	return 0;

      error1:
	xfs_bmap_cancel(&flist);
      error0:
	xfs_iunlock(quotip, XFS_ILOCK_EXCL);

	return (error);
}

/*
 * Maps a dquot to the buffer containing its on-disk version.
 * This returns a ptr to the buffer containing the on-disk dquot
 * in the bpp param, and a ptr to the on-disk dquot within that buffer
 */
STATIC int
xfs_qm_dqtobp(
	xfs_trans_t		**tpp,
	xfs_dquot_t		*dqp,
	xfs_disk_dquot_t	**O_ddpp,
	xfs_buf_t		**O_bpp,
	uint			flags)
{
	xfs_bmbt_irec_t map;
	int		nmaps, error;
	xfs_buf_t	*bp;
	xfs_inode_t	*quotip;
	xfs_mount_t	*mp;
	xfs_disk_dquot_t *ddq;
	xfs_dqid_t	id;
	boolean_t	newdquot;
	xfs_trans_t	*tp = (tpp ? *tpp : NULL);

	mp = dqp->q_mount;