[net-next-2.6.git] / fs / xfs / xfs_utils.c

/*
 * Copyright (c) 2000-2002,2005 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_types.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_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
#include "xfs_error.h"
#include "xfs_quota.h"
#include "xfs_itable.h"
#include "xfs_utils.h"


/*
 * Allocates a new inode from disk and return a pointer to the
 * incore copy. This routine will internally commit the current
 * transaction and allocate a new one if the Space Manager needed
 * to do an allocation to replenish the inode free-list.
 *
 * This routine is designed to be called from xfs_create and
 * xfs_create_dir.
 *
 */
int
xfs_dir_ialloc(
	xfs_trans_t	**tpp,		/* input: current transaction;
					   output: may be a new transaction. */
	xfs_inode_t	*dp,		/* directory within whose allocate
					   the inode. */
	mode_t		mode,
	xfs_nlink_t	nlink,
	xfs_dev_t	rdev,
	cred_t		*credp,
	prid_t		prid,		/* project id */
	int		okalloc,	/* ok to allocate new space */
	xfs_inode_t	**ipp,		/* pointer to inode; it will be
					   locked. */
	int		*committed)

{
	xfs_trans_t	*tp;
	xfs_trans_t	*ntp;
	xfs_inode_t	*ip;
	xfs_buf_t	*ialloc_context = NULL;
	boolean_t	call_again = B_FALSE;
	int		code;
	uint		log_res;
	uint		log_count;
	void		*dqinfo;
	uint		tflags;

	tp = *tpp;
	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);

	/*
	 * xfs_ialloc will return a pointer to an incore inode if
	 * the Space Manager has an available inode on the free
	 * list. Otherwise, it will do an allocation and replenish
	 * the freelist.  Since we can only do one allocation per
	 * transaction without deadlocks, we will need to commit the
	 * current transaction and start a new one.  We will then
	 * need to call xfs_ialloc again to get the inode.
	 *
	 * If xfs_ialloc did an allocation to replenish the freelist,
	 * it returns the bp containing the head of the freelist as
	 * ialloc_context. We will hold a lock on it across the
	 * transaction commit so that no other process can steal
	 * the inode(s) that we've just allocated.
	 */
	code = xfs_ialloc(tp, dp, mode, nlink, rdev, credp, prid, okalloc,
			  &ialloc_context, &call_again, &ip);

	/*
	 * Return an error if we were unable to allocate a new inode.
	 * This should only happen if we run out of space on disk or
	 * encounter a disk error.
	 */
	if (code) {
		*ipp = NULL;
		return code;
	}
	if (!call_again && (ip == NULL)) {
		*ipp = NULL;
		return XFS_ERROR(ENOSPC);
	}

	/*
	 * If call_again is set, then we were unable to get an
	 * inode in one operation.  We need to commit the current
	 * transaction and call xfs_ialloc() again.  It is guaranteed
	 * to succeed the second time.
	 */
	if (call_again) {

		/*
		 * Normally, xfs_trans_commit releases all the locks.
		 * We call bhold to hang on to the ialloc_context across
		 * the commit.  Holding this buffer prevents any other
		 * processes from doing any allocations in this
		 * allocation group.
		 */
		xfs_trans_bhold(tp, ialloc_context);
		/*
		 * Save the log reservation so we can use
		 * them in the next transaction.
		 */
		log_res = xfs_trans_get_log_res(tp);
		log_count = xfs_trans_get_log_count(tp);

		/*
		 * We want the quota changes to be associated with the next
		 * transaction, NOT this one. So, detach the dqinfo from this
		 * and attach it to the next transaction.
		 */
		dqinfo = NULL;
		tflags = 0;
		if (tp->t_dqinfo) {
			dqinfo = (void *)tp->t_dqinfo;
			tp->t_dqinfo = NULL;
			tflags = tp->t_flags & XFS_TRANS_DQ_DIRTY;
			tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY);
		}

		ntp = xfs_trans_dup(tp);
		code = xfs_trans_commit(tp, 0);
		tp = ntp;
		if (committed != NULL) {
			*committed = 1;
		}
		/*
		 * If we get an error during the commit processing,
		 * release the buffer that is still held and return
		 * to the caller.
		 */
		if (code) {
			xfs_buf_relse(ialloc_context);
			if (dqinfo) {
				tp->t_dqinfo = dqinfo;
				xfs_trans_free_dqinfo(tp);
			}
			*tpp = ntp;
			*ipp = NULL;
			return code;
		}

		/*
		 * transaction commit worked ok so we can drop the extra ticket
		 * reference that we gained in xfs_trans_dup()
		 */
		xfs_log_ticket_put(tp->t_ticket);
		code = xfs_trans_reserve(tp, 0, log_res, 0,
					 XFS_TRANS_PERM_LOG_RES, log_count);
		/*
		 * Re-attach the quota info that we detached from prev trx.
		 */
		if (dqinfo) {
			tp->t_dqinfo = dqinfo;
			tp->t_flags |= tflags;
		}

		if (code) {
			xfs_buf_relse(ialloc_context);
			*tpp = ntp;
			*ipp = NULL;
			return code;
		}
		xfs_trans_bjoin(tp, ialloc_context);

		/*
		 * Call ialloc again. Since we've locked out all
		 * other allocations in this allocation group,
		 * this call should always succeed.
		 */
		code = xfs_ialloc(tp, dp, mode, nlink, rdev, credp, prid,
				  okalloc, &ialloc_context, &call_again, &ip);

		/*
		 * If we get an error at this point, return to the caller
		 * so that the current transaction can be aborted.
		 */
		if (code) {
			*tpp = tp;
			*ipp = NULL;
			return code;
		}
		ASSERT ((!call_again) && (ip != NULL));

	} else {
		if (committed != NULL) {
			*committed = 0;
		}
	}

	*ipp = ip;
	*tpp = tp;

	return 0;
}

/*
 * Decrement the link count on an inode & log the change.
 * If this causes the link count to go to zero, initiate the
 * logging activity required to truncate a file.
 */
int				/* error */
xfs_droplink(
	xfs_trans_t *tp,
	xfs_inode_t *ip)
{
	int	error;

	xfs_ichgtime(ip, XFS_ICHGTIME_CHG);

	ASSERT (ip->i_d.di_nlink > 0);
	ip->i_d.di_nlink--;
	drop_nlink(VFS_I(ip));
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	error = 0;
	if (ip->i_d.di_nlink == 0) {
		/*
		 * We're dropping the last link to this file.
		 * Move the on-disk inode to the AGI unlinked list.
		 * From xfs_inactive() we will pull the inode from
		 * the list and free it.
		 */
		error = xfs_iunlink(tp, ip);
	}
	return error;
}

/*
 * This gets called when the inode's version needs to be changed from 1 to 2.
 * Currently this happens when the nlink field overflows the old 16-bit value
 * or when chproj is called to change the project for the first time.
 * As a side effect the superblock version will also get rev'd
 * to contain the NLINK bit.
 */
void
xfs_bump_ino_vers2(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip)
{
	xfs_mount_t	*mp;

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(ip->i_d.di_version == 1);

	ip->i_d.di_version = 2;
	ip->i_d.di_onlink = 0;
	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
	mp = tp->t_mountp;
	if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
		spin_lock(&mp->m_sb_lock);
		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
			xfs_sb_version_addnlink(&mp->m_sb);
			spin_unlock(&mp->m_sb_lock);
			xfs_mod_sb(tp, XFS_SB_VERSIONNUM);
		} else {
			spin_unlock(&mp->m_sb_lock);
		}
	}
	/* Caller must log the inode */
}

/*
 * Increment the link count on an inode & log the change.
 */
int
xfs_bumplink(
	xfs_trans_t *tp,
	xfs_inode_t *ip)
{
	if (ip->i_d.di_nlink >= XFS_MAXLINK)
		return XFS_ERROR(EMLINK);
	xfs_ichgtime(ip, XFS_ICHGTIME_CHG);

	ASSERT(ip->i_d.di_nlink > 0);
	ip->i_d.di_nlink++;
	inc_nlink(VFS_I(ip));
	if ((ip->i_d.di_version == 1) &&
	    (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
		/*
		 * The inode has increased its number of links beyond
		 * what can fit in an old format inode.  It now needs
		 * to be converted to a version 2 inode with a 32 bit
		 * link count.  If this is the first inode in the file
		 * system to do this, then we need to bump the superblock
		 * version number as well.
		 */
		xfs_bump_ino_vers2(tp, ip);
	}

	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
	return 0;
}
Commit	Line	Data
1da177e4	1	/*
7b718769 NS	2	* Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
7b718769 NS	3	* All Rights Reserved.
1da177e4	4	*
7b718769 NS	5	* This program is free software; you can redistribute it and/or
7b718769 NS	6	* modify it under the terms of the GNU General Public License as
1da177e4 LT	7	* published by the Free Software Foundation.
1da177e4 LT	8	*
7b718769 NS	9	* This program is distributed in the hope that it would be useful,
	10	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	12	* GNU General Public License for more details.
1da177e4	13	*
7b718769 NS	14	* You should have received a copy of the GNU General Public License
	15	* along with this program; if not, write the Free Software Foundation,
	16	* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
1da177e4	17	*/
1da177e4	18	#include "xfs.h"
a844f451	19	#include "xfs_fs.h"
1da177e4	20	#include "xfs_types.h"
a844f451	21	#include "xfs_bit.h"
1da177e4	22	#include "xfs_log.h"
a844f451	23	#include "xfs_inum.h"
1da177e4 LT	24	#include "xfs_trans.h"
1da177e4 LT	25	#include "xfs_sb.h"
a844f451	26	#include "xfs_ag.h"
1da177e4	27	#include "xfs_dir2.h"
1da177e4 LT	28	#include "xfs_mount.h"
1da177e4 LT	29	#include "xfs_bmap_btree.h"
1da177e4	30	#include "xfs_dinode.h"
1da177e4	31	#include "xfs_inode.h"
a844f451	32	#include "xfs_inode_item.h"
1da177e4 LT	33	#include "xfs_bmap.h"
	34	#include "xfs_error.h"
	35	#include "xfs_quota.h"
1da177e4 LT	36	#include "xfs_itable.h"
	37	#include "xfs_utils.h"
	38
1da177e4	39
1da177e4 LT	40	/*
	41	* Allocates a new inode from disk and return a pointer to the
	42	* incore copy. This routine will internally commit the current
	43	* transaction and allocate a new one if the Space Manager needed
	44	* to do an allocation to replenish the inode free-list.
	45	*
	46	* This routine is designed to be called from xfs_create and
	47	* xfs_create_dir.
	48	*
	49	*/
	50	int
	51	xfs_dir_ialloc(
	52	xfs_trans_t *tpp, / input: current transaction;
	53	output: may be a new transaction. */
	54	xfs_inode_t dp, / directory within whose allocate
	55	the inode. */
	56	mode_t mode,
31b084ae	57	xfs_nlink_t nlink,
1da177e4 LT	58	xfs_dev_t rdev,
	59	cred_t *credp,
	60	prid_t prid, /* project id */
	61	int okalloc, /* ok to allocate new space */
	62	xfs_inode_t *ipp, / pointer to inode; it will be
	63	locked. */
	64	int *committed)
	65
	66	{
	67	xfs_trans_t *tp;
	68	xfs_trans_t *ntp;
	69	xfs_inode_t *ip;
	70	xfs_buf_t *ialloc_context = NULL;
	71	boolean_t call_again = B_FALSE;
	72	int code;
	73	uint log_res;
	74	uint log_count;
	75	void *dqinfo;
	76	uint tflags;
	77
	78	tp = *tpp;
	79	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
	80
	81	/*
	82	* xfs_ialloc will return a pointer to an incore inode if
	83	* the Space Manager has an available inode on the free
	84	* list. Otherwise, it will do an allocation and replenish
	85	* the freelist. Since we can only do one allocation per
	86	* transaction without deadlocks, we will need to commit the
	87	* current transaction and start a new one. We will then
	88	* need to call xfs_ialloc again to get the inode.
	89	*
	90	* If xfs_ialloc did an allocation to replenish the freelist,
	91	* it returns the bp containing the head of the freelist as
	92	* ialloc_context. We will hold a lock on it across the
	93	* transaction commit so that no other process can steal
	94	* the inode(s) that we've just allocated.
	95	*/
	96	code = xfs_ialloc(tp, dp, mode, nlink, rdev, credp, prid, okalloc,
	97	&ialloc_context, &call_again, &ip);
	98
	99	/*
	100	* Return an error if we were unable to allocate a new inode.
	101	* This should only happen if we run out of space on disk or
	102	* encounter a disk error.
	103	*/
	104	if (code) {
	105	*ipp = NULL;
	106	return code;
	107	}
	108	if (!call_again && (ip == NULL)) {
	109	*ipp = NULL;
	110	return XFS_ERROR(ENOSPC);
	111	}
	112
	113	/*
	114	* If call_again is set, then we were unable to get an
	115	* inode in one operation. We need to commit the current
	116	* transaction and call xfs_ialloc() again. It is guaranteed
	117	* to succeed the second time.
	118	*/
	119	if (call_again) {
	120
	121	/*
122	* Normally, xfs_trans_commit releases all the locks.
123	* We call bhold to hang on to the ialloc_context across
124	* the commit. Holding this buffer prevents any other
125	* processes from doing any allocations in this
126	* allocation group.
127	*/
128	xfs_trans_bhold(tp, ialloc_context);
129	/*
130	* Save the log reservation so we can use
131	* them in the next transaction.
132	*/
133	log_res = xfs_trans_get_log_res(tp);
134	log_count = xfs_trans_get_log_count(tp);
135
136	/*
137	* We want the quota changes to be associated with the next
138	* transaction, NOT this one. So, detach the dqinfo from this
139	* and attach it to the next transaction.
140	*/
141	dqinfo = NULL;
142	tflags = 0;
143	if (tp->t_dqinfo) {
144	dqinfo = (void *)tp->t_dqinfo;
145	tp->t_dqinfo = NULL;
146	tflags = tp->t_flags & XFS_TRANS_DQ_DIRTY;
147	tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY);
148	}
149
150	ntp = xfs_trans_dup(tp);
1c72bf90	151	code = xfs_trans_commit(tp, 0);
1da177e4 LT	152	tp = ntp;
	153	if (committed != NULL) {
	154	*committed = 1;
	155	}
	156	/*
	157	* If we get an error during the commit processing,
	158	* release the buffer that is still held and return
	159	* to the caller.
	160	*/
	161	if (code) {
	162	xfs_buf_relse(ialloc_context);
	163	if (dqinfo) {
	164	tp->t_dqinfo = dqinfo;
7d095257	165	xfs_trans_free_dqinfo(tp);
1da177e4 LT	166	}
	167	*tpp = ntp;
	168	*ipp = NULL;
	169	return code;
	170	}
cc09c0dc DC	171
	172	/*
	173	* transaction commit worked ok so we can drop the extra ticket
	174	* reference that we gained in xfs_trans_dup()
	175	*/
	176	xfs_log_ticket_put(tp->t_ticket);
1da177e4 LT	177	code = xfs_trans_reserve(tp, 0, log_res, 0,
	178	XFS_TRANS_PERM_LOG_RES, log_count);
	179	/*
	180	* Re-attach the quota info that we detached from prev trx.
	181	*/
	182	if (dqinfo) {
	183	tp->t_dqinfo = dqinfo;
	184	tp->t_flags \|= tflags;
	185	}
	186
	187	if (code) {
	188	xfs_buf_relse(ialloc_context);
	189	*tpp = ntp;
	190	*ipp = NULL;
	191	return code;
	192	}
	193	xfs_trans_bjoin(tp, ialloc_context);
	194
	195	/*
	196	* Call ialloc again. Since we've locked out all
	197	* other allocations in this allocation group,
	198	* this call should always succeed.
	199	*/
	200	code = xfs_ialloc(tp, dp, mode, nlink, rdev, credp, prid,
	201	okalloc, &ialloc_context, &call_again, &ip);
	202
	203	/*
	204	* If we get an error at this point, return to the caller
	205	* so that the current transaction can be aborted.
	206	*/
	207	if (code) {
	208	*tpp = tp;
	209	*ipp = NULL;
	210	return code;
	211	}
	212	ASSERT ((!call_again) && (ip != NULL));
	213
	214	} else {
	215	if (committed != NULL) {
	216	*committed = 0;
	217	}
	218	}
	219
	220	*ipp = ip;
	221	*tpp = tp;
	222
	223	return 0;
	224	}
	225
	226	/*
	227	* Decrement the link count on an inode & log the change.
	228	* If this causes the link count to go to zero, initiate the
	229	* logging activity required to truncate a file.
	230	*/
	231	int /* error */
	232	xfs_droplink(
	233	xfs_trans_t *tp,
	234	xfs_inode_t *ip)
	235	{
	236	int error;
	237
	238	xfs_ichgtime(ip, XFS_ICHGTIME_CHG);
	239
	240	ASSERT (ip->i_d.di_nlink > 0);
241	ip->i_d.di_nlink--;
01651646	242	drop_nlink(VFS_I(ip));
1da177e4 LT	243	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
	244
	245	error = 0;
	246	if (ip->i_d.di_nlink == 0) {
	247	/*
	248	* We're dropping the last link to this file.
	249	* Move the on-disk inode to the AGI unlinked list.
	250	* From xfs_inactive() we will pull the inode from
	251	* the list and free it.
	252	*/
	253	error = xfs_iunlink(tp, ip);
	254	}
	255	return error;
	256	}
	257
	258	/*
	259	* This gets called when the inode's version needs to be changed from 1 to 2.
	260	* Currently this happens when the nlink field overflows the old 16-bit value
	261	* or when chproj is called to change the project for the first time.
	262	* As a side effect the superblock version will also get rev'd
	263	* to contain the NLINK bit.
	264	*/
	265	void
	266	xfs_bump_ino_vers2(
	267	xfs_trans_t *tp,
	268	xfs_inode_t *ip)
	269	{
	270	xfs_mount_t *mp;
1da177e4	271
579aa9ca	272	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
51ce16d5	273	ASSERT(ip->i_d.di_version == 1);
1da177e4	274
51ce16d5	275	ip->i_d.di_version = 2;
1da177e4 LT	276	ip->i_d.di_onlink = 0;
	277	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
	278	mp = tp->t_mountp;
62118709	279	if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
3685c2a1	280	spin_lock(&mp->m_sb_lock);
62118709 ES	281	if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
62118709 ES	282	xfs_sb_version_addnlink(&mp->m_sb);
3685c2a1	283	spin_unlock(&mp->m_sb_lock);
1da177e4 LT	284	xfs_mod_sb(tp, XFS_SB_VERSIONNUM);
1da177e4 LT	285	} else {
3685c2a1	286	spin_unlock(&mp->m_sb_lock);
1da177e4 LT	287	}
	288	}
	289	/* Caller must log the inode */
	290	}
	291
	292	/*
	293	* Increment the link count on an inode & log the change.
	294	*/
	295	int
	296	xfs_bumplink(
	297	xfs_trans_t *tp,
	298	xfs_inode_t *ip)
	299	{
	300	if (ip->i_d.di_nlink >= XFS_MAXLINK)
	301	return XFS_ERROR(EMLINK);
	302	xfs_ichgtime(ip, XFS_ICHGTIME_CHG);
	303
	304	ASSERT(ip->i_d.di_nlink > 0);
	305	ip->i_d.di_nlink++;
01651646	306	inc_nlink(VFS_I(ip));
51ce16d5	307	if ((ip->i_d.di_version == 1) &&
1da177e4 LT	308	(ip->i_d.di_nlink > XFS_MAXLINK_1)) {
	309	/*
	310	* The inode has increased its number of links beyond
	311	* what can fit in an old format inode. It now needs
	312	* to be converted to a version 2 inode with a 32 bit
	313	* link count. If this is the first inode in the file
	314	* system to do this, then we need to bump the superblock
	315	* version number as well.
	316	*/
	317	xfs_bump_ino_vers2(tp, ip);
	318	}
	319
	320	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
	321	return 0;
	322	}