/*
- * Copyright (c) 2000-2004 Silicon Graphics, Inc. All Rights Reserved.
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
*
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
+ * 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.
+ * 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.
*
- * Further, this software is distributed without any warranty that it is
- * free of the rightful claim of any third person regarding infringement
- * or the like. Any license provided herein, whether implied or
- * otherwise, applies only to this software file. Patent licenses, if
- * any, provided herein do not apply to combinations of this program with
- * other software, or any other product whatsoever.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc., 59
- * Temple Place - Suite 330, Boston MA 02111-1307, USA.
- *
- * Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
- * Mountain View, CA 94043, or:
- *
- * http://www.sgi.com
- *
- * For further information regarding this notice, see:
- *
- * http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
+ * 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 <linux/capability.h>
+
#include "xfs.h"
-#include "xfs_macros.h"
+#include "xfs_fs.h"
#include "xfs_types.h"
-#include "xfs_inum.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_dmapi.h"
#include "xfs_mount.h"
-#include "xfs_alloc_btree.h"
+#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
+#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_itable.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_alloc.h"
-#include "xfs_attr_sf.h"
#include "xfs_dir_sf.h"
#include "xfs_dir2_sf.h"
+#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
-#include "xfs_inode_item.h"
#include "xfs_inode.h"
+#include "xfs_inode_item.h"
+#include "xfs_dir_leaf.h"
+#include "xfs_itable.h"
+#include "xfs_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_alloc.h"
#include "xfs_bmap.h"
-#include "xfs_da_btree.h"
#include "xfs_attr.h"
#include "xfs_rw.h"
-#include "xfs_refcache.h"
#include "xfs_error.h"
-#include "xfs_bit.h"
-#include "xfs_rtalloc.h"
#include "xfs_quota.h"
#include "xfs_utils.h"
+#include "xfs_rtalloc.h"
+#include "xfs_refcache.h"
#include "xfs_trans_space.h"
-#include "xfs_dir_leaf.h"
-#include "xfs_mac.h"
#include "xfs_log_priv.h"
+#include "xfs_mac.h"
/*
* If it's a directory with any blocks, read-ahead block 0
* as we're almost certain to have the next operation be a read there.
*/
- if (vp->v_type == VDIR && ip->i_d.di_nextents > 0) {
+ if (VN_ISDIR(vp) && ip->i_d.di_nextents > 0) {
mode = xfs_ilock_map_shared(ip);
if (ip->i_d.di_nextents > 0)
(void)xfs_da_reada_buf(NULL, ip, 0, XFS_DATA_FORK);
/*
* Copy from in-core inode.
*/
- vap->va_type = vp->v_type;
- vap->va_mode = ip->i_d.di_mode & MODEMASK;
+ vap->va_mode = ip->i_d.di_mode;
vap->va_uid = ip->i_d.di_uid;
vap->va_gid = ip->i_d.di_gid;
vap->va_xid = ip->i_d.di_xid;
/*
* Check vnode type block/char vs. everything else.
- * Do it with bitmask because that's faster than looking
- * for multiple values individually.
*/
- if (((1 << vp->v_type) & ((1<<VBLK) | (1<<VCHR))) == 0) {
+ switch (ip->i_d.di_mode & S_IFMT) {
+ case S_IFBLK:
+ case S_IFCHR:
+ vap->va_rdev = ip->i_df.if_u2.if_rdev;
+ vap->va_blocksize = BLKDEV_IOSIZE;
+ break;
+ default:
vap->va_rdev = 0;
if (!(ip->i_d.di_flags & XFS_DIFLAG_REALTIME)) {
-
-#if 0
- /* Large block sizes confuse various
- * user space programs, so letting the
- * stripe size through is not a good
- * idea for now.
- */
- vap->va_blocksize = mp->m_swidth ?
- /*
- * If the underlying volume is a stripe, then
- * return the stripe width in bytes as the
- * recommended I/O size.
- */
- (mp->m_swidth << mp->m_sb.sb_blocklog) :
- /*
- * Return the largest of the preferred buffer
- * sizes since doing small I/Os into larger
- * buffers causes buffers to be decommissioned.
- * The value returned is in bytes.
- */
- (1 << (int)MAX(mp->m_readio_log,
- mp->m_writeio_log));
-
-#else
- vap->va_blocksize =
- /*
- * Return the largest of the preferred buffer
- * sizes since doing small I/Os into larger
- * buffers causes buffers to be decommissioned.
- * The value returned is in bytes.
- */
- 1 << (int)MAX(mp->m_readio_log,
- mp->m_writeio_log);
-#endif
+ vap->va_blocksize = xfs_preferred_iosize(mp);
} else {
/*
(ip->i_d.di_extsize << mp->m_sb.sb_blocklog) :
(mp->m_sb.sb_rextsize << mp->m_sb.sb_blocklog);
}
- } else {
- vap->va_rdev = ip->i_df.if_u2.if_rdev;
- vap->va_blocksize = BLKDEV_IOSIZE;
+ break;
}
- vap->va_atime.tv_sec = ip->i_d.di_atime.t_sec;
- vap->va_atime.tv_nsec = ip->i_d.di_atime.t_nsec;
+ vn_atime_to_timespec(vp, &vap->va_atime);
vap->va_mtime.tv_sec = ip->i_d.di_mtime.t_sec;
vap->va_mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
vap->va_ctime.tv_sec = ip->i_d.di_ctime.t_sec;
* If the IDs do change before we take the ilock, we're covered
* because the i_*dquot fields will get updated anyway.
*/
- if (XFS_IS_QUOTA_ON(mp) && (mask & (XFS_AT_UID|XFS_AT_GID))) {
+ if (XFS_IS_QUOTA_ON(mp) &&
+ (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID))) {
uint qflags = 0;
/* FIXME: handle xid? */
- if (mask & XFS_AT_UID) {
+ if ((mask & XFS_AT_UID) && XFS_IS_UQUOTA_ON(mp)) {
uid = vap->va_uid;
qflags |= XFS_QMOPT_UQUOTA;
} else {
uid = ip->i_d.di_uid;
}
- if (mask & XFS_AT_GID) {
+ if ((mask & XFS_AT_GID) && XFS_IS_GQUOTA_ON(mp)) {
gid = vap->va_gid;
qflags |= XFS_QMOPT_GQUOTA;
} else {
gid = ip->i_d.di_gid;
}
+ if ((mask & XFS_AT_PROJID) && XFS_IS_PQUOTA_ON(mp)) {
+ projid = vap->va_projid;
+ qflags |= XFS_QMOPT_PQUOTA;
+ } else {
+ projid = ip->i_d.di_projid;
+ }
/*
* We take a reference when we initialize udqp and gdqp,
* so it is important that we never blindly double trip on
*/
ASSERT(udqp == NULL);
ASSERT(gdqp == NULL);
- code = XFS_QM_DQVOPALLOC(mp, ip, uid,gid, qflags, &udqp, &gdqp);
+ code = XFS_QM_DQVOPALLOC(mp, ip, uid, gid, projid, qflags,
+ &udqp, &gdqp);
if (code)
- return (code);
+ return code;
}
/*
m |= S_ISGID;
#if 0
/* Linux allows this, Irix doesn't. */
- if ((vap->va_mode & S_ISVTX) && vp->v_type != VDIR)
+ if ((vap->va_mode & S_ISVTX) && !VN_ISDIR(vp))
m |= S_ISVTX;
#endif
if (m && !capable(CAP_FSETID))
* that the group ID supplied to the chown() function
* shall be equal to either the group ID or one of the
* supplementary group IDs of the calling process.
- *
- * XXX: How does restricted_chown affect projid?
*/
if (restricted_chown &&
(iuid != uid || (igid != gid &&
goto error_return;
}
/*
- * Do a quota reservation only if uid or gid is actually
+ * Do a quota reservation only if uid/projid/gid is actually
* going to change.
*/
if ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
+ (XFS_IS_PQUOTA_ON(mp) && iprojid != projid) ||
(XFS_IS_GQUOTA_ON(mp) && igid != gid)) {
/* FIXME: handle xid? */
ASSERT(tp);
goto error_return;
}
- if (vp->v_type == VDIR) {
+ if (VN_ISDIR(vp)) {
code = XFS_ERROR(EISDIR);
goto error_return;
- } else if (vp->v_type != VREG) {
+ } else if (!VN_ISREG(vp)) {
code = XFS_ERROR(EINVAL);
goto error_return;
}
/*
* Can't change extent size if any extents are allocated.
*/
- if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
- (mask & XFS_AT_EXTSIZE) &&
+ if (ip->i_d.di_nextents && (mask & XFS_AT_EXTSIZE) &&
((ip->i_d.di_extsize << mp->m_sb.sb_blocklog) !=
vap->va_extsize) ) {
code = XFS_ERROR(EINVAL); /* EFBIG? */
goto error_return;
}
- /*
- * Can't set extent size unless the file is marked, or
- * about to be marked as a realtime file.
- *
- * This check will be removed when fixed size extents
- * with buffered data writes is implemented.
- *
- */
- if ((mask & XFS_AT_EXTSIZE) &&
- ((ip->i_d.di_extsize << mp->m_sb.sb_blocklog) !=
- vap->va_extsize) &&
- (!((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) ||
- ((mask & XFS_AT_XFLAGS) &&
- (vap->va_xflags & XFS_XFLAG_REALTIME))))) {
- code = XFS_ERROR(EINVAL);
- goto error_return;
- }
-
/*
* Can't change realtime flag if any extents are allocated.
*/
- if (ip->i_d.di_nextents && (mask & XFS_AT_XFLAGS) &&
+ if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
+ (mask & XFS_AT_XFLAGS) &&
(ip->i_d.di_flags & XFS_DIFLAG_REALTIME) !=
(vap->va_xflags & XFS_XFLAG_REALTIME)) {
code = XFS_ERROR(EINVAL); /* EFBIG? */
*/
if (mask & XFS_AT_SIZE) {
code = 0;
- if (vap->va_size > ip->i_d.di_size)
+ if ((vap->va_size > ip->i_d.di_size) &&
+ (flags & ATTR_NOSIZETOK) == 0) {
code = xfs_igrow_start(ip, vap->va_size, credp);
+ }
xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ vn_iowait(vp); /* wait for the completion of any pending DIOs */
if (!code)
code = xfs_itruncate_data(ip, vap->va_size);
if (code) {
}
if (igid != gid) {
if (XFS_IS_GQUOTA_ON(mp)) {
+ ASSERT(!XFS_IS_PQUOTA_ON(mp));
ASSERT(mask & XFS_AT_GID);
ASSERT(gdqp);
olddquot2 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
ip->i_d.di_gid = gid;
}
if (iprojid != projid) {
+ if (XFS_IS_PQUOTA_ON(mp)) {
+ ASSERT(!XFS_IS_GQUOTA_ON(mp));
+ ASSERT(mask & XFS_AT_PROJID);
+ ASSERT(gdqp);
+ olddquot2 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
+ &ip->i_gdquot, gdqp);
+ }
ip->i_d.di_projid = projid;
/*
* We may have to rev the inode as well as
di_flags |= XFS_DIFLAG_NOATIME;
if (vap->va_xflags & XFS_XFLAG_NODUMP)
di_flags |= XFS_DIFLAG_NODUMP;
+ if (vap->va_xflags & XFS_XFLAG_PROJINHERIT)
+ di_flags |= XFS_DIFLAG_PROJINHERIT;
if ((ip->i_d.di_mode & S_IFMT) == S_IFDIR) {
if (vap->va_xflags & XFS_XFLAG_RTINHERIT)
di_flags |= XFS_DIFLAG_RTINHERIT;
if (vap->va_xflags & XFS_XFLAG_NOSYMLINKS)
di_flags |= XFS_DIFLAG_NOSYMLINKS;
- } else {
+ if (vap->va_xflags & XFS_XFLAG_EXTSZINHERIT)
+ di_flags |= XFS_DIFLAG_EXTSZINHERIT;
+ } else if ((ip->i_d.di_mode & S_IFMT) == S_IFREG) {
if (vap->va_xflags & XFS_XFLAG_REALTIME) {
di_flags |= XFS_DIFLAG_REALTIME;
ip->i_iocore.io_flags |= XFS_IOCORE_RT;
} else {
ip->i_iocore.io_flags &= ~XFS_IOCORE_RT;
}
+ if (vap->va_xflags & XFS_XFLAG_EXTSIZE)
+ di_flags |= XFS_DIFLAG_EXTSIZE;
}
ip->i_d.di_flags = di_flags;
}
* If this is a synchronous mount, make sure that the
* transaction goes to disk before returning to the user.
* This is slightly sub-optimal in that truncates require
- * two sync transactions instead of one for wsync filesytems.
+ * two sync transactions instead of one for wsync filesystems.
* One for the truncate and one for the timestamps since we
* don't want to change the timestamps unless we're sure the
* truncate worked. Truncates are less than 1% of the laddis
goto error_return;
}
- if (!(ioflags & IO_INVIS)) {
- xfs_ichgtime(ip, XFS_ICHGTIME_ACC);
- }
-
/*
* See if the symlink is stored inline.
*/
}
-
error_return:
-
xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
return error;
}
xfs_inode_t *ip;
xfs_trans_t *tp;
int error;
+ int log_flushed = 0, changed = 1;
vn_trace_entry(BHV_TO_VNODE(bdp),
__FUNCTION__, (inst_t *)__return_address);
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (xfs_ipincount(ip)) {
- xfs_log_force(ip->i_mount, (xfs_lsn_t)0,
+ _xfs_log_force(ip->i_mount, (xfs_lsn_t)0,
XFS_LOG_FORCE |
((flag & FSYNC_WAIT)
- ? XFS_LOG_SYNC : 0));
+ ? XFS_LOG_SYNC : 0),
+ &log_flushed);
+ } else {
+ /*
+ * If the inode is not pinned and nothing
+ * has changed we don't need to flush the
+ * cache.
+ */
+ changed = 0;
}
error = 0;
} else {
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
if (flag & FSYNC_WAIT)
xfs_trans_set_sync(tp);
- error = xfs_trans_commit(tp, 0, NULL);
+ error = _xfs_trans_commit(tp, 0, NULL, &log_flushed);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
+
+ if ((ip->i_mount->m_flags & XFS_MOUNT_BARRIER) && changed) {
+ /*
+ * If the log write didn't issue an ordered tag we need
+ * to flush the disk cache for the data device now.
+ */
+ if (!log_flushed)
+ xfs_blkdev_issue_flush(ip->i_mount->m_ddev_targp);
+
+ /*
+ * If this inode is on the RT dev we need to flush that
+ * cache as well.
+ */
+ if (ip->i_d.di_flags & XFS_DIFLAG_REALTIME)
+ xfs_blkdev_issue_flush(ip->i_mount->m_rtdev_targp);
+ }
+
return error;
}
last_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp));
map_len = last_fsb - end_fsb;
if (map_len <= 0)
- return (0);
+ return 0;
nimaps = 1;
xfs_ilock(ip, XFS_ILOCK_SHARED);
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (!error && (nimaps != 0) &&
- (imap.br_startblock != HOLESTARTBLOCK)) {
+ (imap.br_startblock != HOLESTARTBLOCK ||
+ ip->i_delayed_blks)) {
/*
* Attach the dquots to the inode up front.
*/
if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
- return (error);
+ return error;
/*
* There are blocks after the end of file.
ASSERT(XFS_FORCED_SHUTDOWN(mp));
xfs_trans_cancel(tp, 0);
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- return (error);
+ return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
}
xfs_iunlock(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
}
- return (error);
+ return error;
}
/*
*/
done = 0;
XFS_BMAP_INIT(&free_list, &first_block);
- nmaps = sizeof(mval) / sizeof(mval[0]);
+ nmaps = ARRAY_SIZE(mval);
if ((error = xfs_bmapi(tp, ip, 0, XFS_B_TO_FSB(mp, size),
XFS_BMAPI_METADATA, &first_block, 0, mval, &nmaps,
&free_list)))
*/
ntp = xfs_trans_dup(tp);
/*
- * Commit the transaction containing extent freeing and EFD's.
+ * Commit the transaction containing extent freeing and EFDs.
* If we get an error on the commit here or on the reserve below,
* we need to unlock the inode since the new transaction doesn't
* have the inode attached.
if (error) {
xfs_trans_cancel(*tpp, 0);
*tpp = NULL;
- return (error);
+ return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
XFS_DATA_FORK);
ASSERT(ip->i_df.if_bytes == 0);
}
- return (0);
+ return 0;
}
/*
if (error) {
*tpp = NULL;
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- return (error); /* goto out*/
+ return error; /* goto out */
}
tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
xfs_trans_cancel(tp, 0);
*tpp = NULL;
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- return (error);
+ return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
ASSERT(ip->i_d.di_anextents == 0);
*tpp = tp;
- return (0);
+ return 0;
}
STATIC int
vp = BHV_TO_VNODE(bdp);
ip = XFS_BHVTOI(bdp);
- if ((vp->v_type != VREG) || (ip->i_d.di_mode == 0)) {
+ if (!VN_ISREG(vp) || (ip->i_d.di_mode == 0)) {
return 0;
}
if (ip->i_d.di_nlink != 0) {
if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
- ((ip->i_d.di_size > 0) || (VN_CACHED(vp) > 0)) &&
+ ((ip->i_d.di_size > 0) || (VN_CACHED(vp) > 0 ||
+ ip->i_delayed_blks > 0)) &&
(ip->i_df.if_flags & XFS_IFEXTENTS)) &&
- (!(ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC|XFS_DIFLAG_APPEND)))) {
+ (!(ip->i_d.di_flags &
+ (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)))) {
if ((error = xfs_inactive_free_eofblocks(mp, ip)))
- return (error);
+ return error;
/* Update linux inode block count after free above */
- LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+ vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
ip->i_d.di_nblocks + ip->i_delayed_blks);
}
}
* only one with a reference to the inode.
*/
truncate = ((ip->i_d.di_nlink == 0) &&
- ((ip->i_d.di_size != 0) || (ip->i_d.di_nextents > 0)) &&
+ ((ip->i_d.di_size != 0) || (ip->i_d.di_nextents > 0) ||
+ (ip->i_delayed_blks > 0)) &&
((ip->i_d.di_mode & S_IFMT) == S_IFREG));
mp = ip->i_mount;
if (ip->i_d.di_nlink != 0) {
if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
- ((ip->i_d.di_size > 0) || (VN_CACHED(vp) > 0)) &&
- (ip->i_df.if_flags & XFS_IFEXTENTS)) &&
- (!(ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC|XFS_DIFLAG_APPEND)) ||
- (ip->i_delayed_blks != 0))) {
+ ((ip->i_d.di_size > 0) || (VN_CACHED(vp) > 0 ||
+ ip->i_delayed_blks > 0)) &&
+ (ip->i_df.if_flags & XFS_IFEXTENTS) &&
+ (!(ip->i_d.di_flags &
+ (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)) ||
+ (ip->i_delayed_blks != 0)))) {
if ((error = xfs_inactive_free_eofblocks(mp, ip)))
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
/* Update linux inode block count after free above */
- LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+ vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
ip->i_d.di_nblocks + ip->i_delayed_blks);
}
goto out;
ASSERT(ip->i_d.di_nlink == 0);
if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
if (truncate) {
ASSERT(XFS_FORCED_SHUTDOWN(mp));
xfs_trans_cancel(tp, 0);
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_cancel(tp,
XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
xfs_iunlock(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
}
} else if ((ip->i_d.di_mode & S_IFMT) == S_IFLNK) {
if (error) {
ASSERT(tp == NULL);
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
}
xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
if (error) {
ASSERT(XFS_FORCED_SHUTDOWN(mp));
xfs_trans_cancel(tp, 0);
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
}
xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
* cancelled, and the inode is unlocked. Just get out.
*/
if (error)
- return (VN_INACTIVE_CACHE);
+ return VN_INACTIVE_CACHE;
} else if (ip->i_afp) {
xfs_idestroy_fork(ip, XFS_ATTR_FORK);
}
dp = XFS_BHVTOI(dir_bdp);
mp = dp->i_mount;
- dm_di_mode = vap->va_mode|VTTOIF(vap->va_type);
+ dm_di_mode = vap->va_mode;
namelen = VNAMELEN(dentry);
if (DM_EVENT_ENABLED(dir_vp->v_vfsp, dp, DM_EVENT_CREATE)) {
/* Return through std_return after this point. */
udqp = gdqp = NULL;
- if (vap->va_mask & XFS_AT_PROJID)
+ if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
+ prid = dp->i_d.di_projid;
+ else if (vap->va_mask & XFS_AT_PROJID)
prid = (xfs_prid_t)vap->va_projid;
else
prid = (xfs_prid_t)dfltprid;
* Make sure that we have allocated dquot(s) on disk.
*/
error = XFS_QM_DQVOPALLOC(mp, dp,
- current_fsuid(credp), current_fsgid(credp),
+ current_fsuid(credp), current_fsgid(credp), prid,
XFS_QMOPT_QUOTALL|XFS_QMOPT_INHERIT, &udqp, &gdqp);
if (error)
goto std_return;
(error = XFS_DIR_CANENTER(mp, tp, dp, name, namelen)))
goto error_return;
rdev = (vap->va_mask & XFS_AT_RDEV) ? vap->va_rdev : 0;
- error = xfs_dir_ialloc(&tp, dp,
- MAKEIMODE(vap->va_type,vap->va_mode), 1,
+ error = xfs_dir_ialloc(&tp, dp, vap->va_mode, 1,
rdev, credp, prid, resblks > 0,
&ip, &committed);
if (error) {
XFS_QM_DQRELE(mp, gdqp);
/*
- * Propogate the fact that the vnode changed after the
+ * Propagate the fact that the vnode changed after the
* xfs_inode locks have been released.
*/
VOP_VNODE_CHANGE(vp, VCHANGE_FLAGS_TRUNCATED, 3);
abort_return:
cancel_flags |= XFS_TRANS_ABORT;
/* FALLTHROUGH */
- error_return:
+ error_return:
if (tp != NULL)
xfs_trans_cancel(tp, cancel_flags);
* for a log reservation. Since we'll have to wait for the
* inactive code to complete before returning from xfs_iget,
* we need to make sure that we don't have log space reserved
- * when we call xfs_iget. Instead we get an unlocked referece
+ * when we call xfs_iget. Instead we get an unlocked reference
* to the inode before getting our log reservation.
*/
error = xfs_get_dir_entry(dentry, &ip);
int cancel_flags;
int committed;
vnode_t *target_dir_vp;
- bhv_desc_t *src_bdp;
int resblks;
char *target_name = VNAME(dentry);
int target_namelen;
vn_trace_entry(src_vp, __FUNCTION__, (inst_t *)__return_address);
target_namelen = VNAMELEN(dentry);
- if (src_vp->v_type == VDIR)
+ if (VN_ISDIR(src_vp))
return XFS_ERROR(EPERM);
- /*
- * For now, manually find the XFS behavior descriptor for
- * the source vnode. If it doesn't exist then something
- * is wrong and we should just return an error.
- * Eventually we need to figure out how link is going to
- * work in the face of stacked vnodes.
- */
- src_bdp = vn_bhv_lookup_unlocked(VN_BHV_HEAD(src_vp), &xfs_vnodeops);
- if (src_bdp == NULL) {
- return XFS_ERROR(EXDEV);
- }
- sip = XFS_BHVTOI(src_bdp);
+ sip = xfs_vtoi(src_vp);
tdp = XFS_BHVTOI(target_dir_bdp);
mp = tdp->i_mount;
if (XFS_FORCED_SHUTDOWN(mp))
goto error_return;
}
+ /*
+ * If we are using project inheritance, we only allow hard link
+ * creation in our tree when the project IDs are the same; else
+ * the tree quota mechanism could be circumvented.
+ */
+ if (unlikely((tdp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
+ (tdp->i_d.di_projid != sip->i_d.di_projid))) {
+ error = XFS_ERROR(EXDEV);
+ goto error_return;
+ }
+
if (resblks == 0 &&
(error = XFS_DIR_CANENTER(mp, tp, tdp, target_name,
target_namelen)))
abort_return:
cancel_flags |= XFS_TRANS_ABORT;
/* FALLTHROUGH */
+
error_return:
xfs_trans_cancel(tp, cancel_flags);
-
goto std_return;
}
/*
tp = NULL;
dp_joined_to_trans = B_FALSE;
- dm_di_mode = vap->va_mode|VTTOIF(vap->va_type);
+ dm_di_mode = vap->va_mode;
if (DM_EVENT_ENABLED(dir_vp->v_vfsp, dp, DM_EVENT_CREATE)) {
error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE,
mp = dp->i_mount;
udqp = gdqp = NULL;
- if (vap->va_mask & XFS_AT_PROJID)
+ if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
+ prid = dp->i_d.di_projid;
+ else if (vap->va_mask & XFS_AT_PROJID)
prid = (xfs_prid_t)vap->va_projid;
else
prid = (xfs_prid_t)dfltprid;
* Make sure that we have allocated dquot(s) on disk.
*/
error = XFS_QM_DQVOPALLOC(mp, dp,
- current_fsuid(credp), current_fsgid(credp),
+ current_fsuid(credp), current_fsgid(credp), prid,
XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, &udqp, &gdqp);
if (error)
goto std_return;
/*
* create the directory inode.
*/
- error = xfs_dir_ialloc(&tp, dp,
- MAKEIMODE(vap->va_type,vap->va_mode), 2,
+ error = xfs_dir_ialloc(&tp, dp, vap->va_mode, 2,
0, credp, prid, resblks > 0,
&cdp, NULL);
if (error) {
* for a log reservation. Since we'll have to wait for the
* inactive code to complete before returning from xfs_iget,
* we need to make sure that we don't have log space reserved
- * when we call xfs_iget. Instead we get an unlocked referece
+ * when we call xfs_iget. Instead we get an unlocked reference
* to the inode before getting our log reservation.
*/
error = xfs_get_dir_entry(dentry, &cdp);
/* Fall through to std_return with error = 0 or the errno
* from xfs_trans_commit. */
-std_return:
+ std_return:
if (DM_EVENT_ENABLED(dir_vp->v_vfsp, dp, DM_EVENT_POSTREMOVE)) {
(void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE,
dir_vp, DM_RIGHT_NULL,
error1:
xfs_bmap_cancel(&free_list);
cancel_flags |= XFS_TRANS_ABORT;
+ /* FALLTHROUGH */
+
error_return:
xfs_trans_cancel(tp, cancel_flags);
goto std_return;
xfs_trans_t *tp = NULL;
int error = 0;
uint lock_mode;
- xfs_off_t start_offset;
vn_trace_entry(BHV_TO_VNODE(dir_bdp), __FUNCTION__,
(inst_t *)__return_address);
}
lock_mode = xfs_ilock_map_shared(dp);
- start_offset = uiop->uio_offset;
error = XFS_DIR_GETDENTS(dp->i_mount, tp, dp, uiop, eofp);
- if (start_offset != uiop->uio_offset) {
- xfs_ichgtime(dp, XFS_ICHGTIME_ACC);
- }
xfs_iunlock_map_shared(dp, lock_mode);
return error;
}
/* Return through std_return after this point. */
udqp = gdqp = NULL;
- if (vap->va_mask & XFS_AT_PROJID)
+ if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
+ prid = dp->i_d.di_projid;
+ else if (vap->va_mask & XFS_AT_PROJID)
prid = (xfs_prid_t)vap->va_projid;
else
prid = (xfs_prid_t)dfltprid;
* Make sure that we have allocated dquot(s) on disk.
*/
error = XFS_QM_DQVOPALLOC(mp, dp,
- current_fsuid(credp), current_fsgid(credp),
+ current_fsuid(credp), current_fsgid(credp), prid,
XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, &udqp, &gdqp);
if (error)
goto std_return;
vnode_t *vp;
vp = BHV_TO_VNODE(bdp);
- if (vp->v_type == VDIR)
+ if (VN_ISDIR(vp))
return 1;
ip = XFS_BHVTOI(bdp);
if (locktype == VRWLOCK_WRITE) {
xfs_ilock(ip, XFS_IOLOCK_EXCL);
} else if (locktype == VRWLOCK_TRY_READ) {
- return (xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED));
+ return xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED);
} else if (locktype == VRWLOCK_TRY_WRITE) {
- return (xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL));
+ return xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL);
} else {
ASSERT((locktype == VRWLOCK_READ) ||
(locktype == VRWLOCK_WRITE_DIRECT));
vnode_t *vp;
vp = BHV_TO_VNODE(bdp);
- if (vp->v_type == VDIR)
+ if (VN_ISDIR(vp))
return;
ip = XFS_BHVTOI(bdp);
if (locktype == VRWLOCK_WRITE) {
return 0;
}
- if ((ip->i_d.di_mode & S_IFMT) == S_IFREG) {
- if (ip->i_d.di_size > 0) {
- /*
- * Flush and invalidate any data left around that is
- * a part of this file.
- *
- * Get the inode's i/o lock so that buffers are pushed
- * out while holding the proper lock. We can't hold
- * the inode lock here since flushing out buffers may
- * cause us to try to get the lock in xfs_strategy().
- *
- * We don't have to call remapf() here, because there
- * cannot be any mapped file references to this vnode
- * since it is being reclaimed.
- */
- xfs_ilock(ip, XFS_IOLOCK_EXCL);
+ vn_iowait(vp);
- /*
- * If we hit an IO error, we need to make sure that the
- * buffer and page caches of file data for
- * the file are tossed away. We don't want to use
- * VOP_FLUSHINVAL_PAGES here because we don't want dirty
- * pages to stay attached to the vnode, but be
- * marked P_BAD. pdflush/vnode_pagebad
- * hates that.
- */
- if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
- VOP_FLUSHINVAL_PAGES(vp, 0, -1, FI_NONE);
- } else {
- VOP_TOSS_PAGES(vp, 0, -1, FI_NONE);
- }
+ ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
- ASSERT(VN_CACHED(vp) == 0);
- ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) ||
- ip->i_delayed_blks == 0);
- xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- } else if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
- /*
- * di_size field may not be quite accurate if we're
- * shutting down.
- */
- VOP_TOSS_PAGES(vp, 0, -1, FI_NONE);
- ASSERT(VN_CACHED(vp) == 0);
- }
- }
+ /*
+ * Make sure the atime in the XFS inode is correct before freeing the
+ * Linux inode.
+ */
+ xfs_synchronize_atime(ip);
/* If we have nothing to flush with this inode then complete the
* teardown now, otherwise break the link between the xfs inode
xfs_ifunlock(ip);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
- return(1);
+ return 1;
}
ip->i_flags |= XFS_IRECLAIM;
write_unlock(&ih->ih_lock);
}
}
XFS_MOUNT_IUNLOCK(mp);
- xfs_finish_reclaim(ip, noblock,
- XFS_IFLUSH_DELWRI_ELSE_ASYNC);
+ if (xfs_finish_reclaim(ip, noblock,
+ XFS_IFLUSH_DELWRI_ELSE_ASYNC))
+ delay(1);
purged = 1;
break;
}
* errno on error
*
*/
-int
+STATIC int
xfs_alloc_file_space(
xfs_inode_t *ip,
xfs_off_t offset,
int alloc_type,
int attr_flags)
{
+ xfs_mount_t *mp = ip->i_mount;
+ xfs_off_t count;
xfs_filblks_t allocated_fsb;
xfs_filblks_t allocatesize_fsb;
- int committed;
- xfs_off_t count;
- xfs_filblks_t datablocks;
- int error;
+ xfs_extlen_t extsz, temp;
+ xfs_fileoff_t startoffset_fsb;
xfs_fsblock_t firstfsb;
- xfs_bmap_free_t free_list;
- xfs_bmbt_irec_t *imapp;
- xfs_bmbt_irec_t imaps[1];
- xfs_mount_t *mp;
- int numrtextents;
- int reccount;
- uint resblks;
+ int nimaps;
+ int bmapi_flag;
+ int quota_flag;
int rt;
- int rtextsize;
- xfs_fileoff_t startoffset_fsb;
xfs_trans_t *tp;
- int xfs_bmapi_flags;
+ xfs_bmbt_irec_t imaps[1], *imapp;
+ xfs_bmap_free_t free_list;
+ uint qblocks, resblks, resrtextents;
+ int committed;
+ int error;
vn_trace_entry(XFS_ITOV(ip), __FUNCTION__, (inst_t *)__return_address);
- mp = ip->i_mount;
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- /*
- * determine if this is a realtime file
- */
- if ((rt = XFS_IS_REALTIME_INODE(ip)) != 0) {
- if (ip->i_d.di_extsize)
- rtextsize = ip->i_d.di_extsize;
- else
- rtextsize = mp->m_sb.sb_rextsize;
- } else
- rtextsize = 0;
+ rt = XFS_IS_REALTIME_INODE(ip);
+ if (unlikely(rt)) {
+ if (!(extsz = ip->i_d.di_extsize))
+ extsz = mp->m_sb.sb_rextsize;
+ } else {
+ extsz = ip->i_d.di_extsize;
+ }
if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
return error;
count = len;
error = 0;
imapp = &imaps[0];
- reccount = 1;
- xfs_bmapi_flags = XFS_BMAPI_WRITE | (alloc_type ? XFS_BMAPI_PREALLOC : 0);
+ nimaps = 1;
+ bmapi_flag = XFS_BMAPI_WRITE | (alloc_type ? XFS_BMAPI_PREALLOC : 0);
startoffset_fsb = XFS_B_TO_FSBT(mp, offset);
allocatesize_fsb = XFS_B_TO_FSB(mp, count);
offset, end_dmi_offset - offset,
0, NULL);
if (error)
- return(error);
+ return error;
}
/*
- * allocate file space until done or until there is an error
+ * Allocate file space until done or until there is an error
*/
retry:
while (allocatesize_fsb && !error) {
+ xfs_fileoff_t s, e;
+
/*
- * determine if reserving space on
- * the data or realtime partition.
+ * Determine space reservations for data/realtime.
*/
- if (rt) {
- xfs_fileoff_t s, e;
-
+ if (unlikely(extsz)) {
s = startoffset_fsb;
- do_div(s, rtextsize);
- s *= rtextsize;
- e = roundup_64(startoffset_fsb + allocatesize_fsb,
- rtextsize);
- numrtextents = (int)(e - s) / mp->m_sb.sb_rextsize;
- datablocks = 0;
+ do_div(s, extsz);
+ s *= extsz;
+ e = startoffset_fsb + allocatesize_fsb;
+ if ((temp = do_mod(startoffset_fsb, extsz)))
+ e += temp;
+ if ((temp = do_mod(e, extsz)))
+ e += extsz - temp;
} else {
- datablocks = allocatesize_fsb;
- numrtextents = 0;
+ s = 0;
+ e = allocatesize_fsb;
+ }
+
+ if (unlikely(rt)) {
+ resrtextents = qblocks = (uint)(e - s);
+ resrtextents /= mp->m_sb.sb_rextsize;
+ resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
+ quota_flag = XFS_QMOPT_RES_RTBLKS;
+ } else {
+ resrtextents = 0;
+ resblks = qblocks = \
+ XFS_DIOSTRAT_SPACE_RES(mp, (uint)(e - s));
+ quota_flag = XFS_QMOPT_RES_REGBLKS;
}
/*
- * allocate and setup the transaction
+ * Allocate and setup the transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT);
- resblks = XFS_DIOSTRAT_SPACE_RES(mp, datablocks);
- error = xfs_trans_reserve(tp,
- resblks,
- XFS_WRITE_LOG_RES(mp),
- numrtextents,
+ error = xfs_trans_reserve(tp, resblks,
+ XFS_WRITE_LOG_RES(mp), resrtextents,
XFS_TRANS_PERM_LOG_RES,
XFS_WRITE_LOG_COUNT);
-
/*
- * check for running out of space
+ * Check for running out of space
*/
if (error) {
/*
break;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- error = XFS_TRANS_RESERVE_QUOTA_BYDQUOTS(mp, tp,
- ip->i_udquot, ip->i_gdquot, resblks, 0, rt ?
- XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
+ error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, tp, ip,
+ qblocks, 0, quota_flag);
if (error)
goto error1;
xfs_trans_ihold(tp, ip);
/*
- * issue the bmapi() call to allocate the blocks
+ * Issue the xfs_bmapi() call to allocate the blocks
*/
XFS_BMAP_INIT(&free_list, &firstfsb);
error = xfs_bmapi(tp, ip, startoffset_fsb,
- allocatesize_fsb, xfs_bmapi_flags,
- &firstfsb, 0, imapp, &reccount,
+ allocatesize_fsb, bmapi_flag,
+ &firstfsb, 0, imapp, &nimaps,
&free_list);
if (error) {
goto error0;
}
/*
- * complete the transaction
+ * Complete the transaction
*/
error = xfs_bmap_finish(&tp, &free_list, firstfsb, &committed);
if (error) {
allocated_fsb = imapp->br_blockcount;
- if (reccount == 0) {
+ if (nimaps == 0) {
error = XFS_ERROR(ENOSPC);
break;
}
return error;
- error0:
+error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */
xfs_bmap_cancel(&free_list);
- error1:
+ XFS_TRANS_UNRESERVE_QUOTA_NBLKS(mp, tp, ip, qblocks, 0, quota_flag);
+
+error1: /* Just cancel transaction */
xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
goto dmapi_enospc_check;
xfs_off_t len,
int attr_flags)
{
+ vnode_t *vp;
int committed;
int done;
xfs_off_t end_dmi_offset;
xfs_trans_t *tp;
int need_iolock = 1;
- vn_trace_entry(XFS_ITOV(ip), __FUNCTION__, (inst_t *)__return_address);
+ vp = XFS_ITOV(ip);
mp = ip->i_mount;
+ vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
+
if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
return error;
DM_EVENT_ENABLED(XFS_MTOVFS(mp), ip, DM_EVENT_WRITE)) {
if (end_dmi_offset > ip->i_d.di_size)
end_dmi_offset = ip->i_d.di_size;
- error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, XFS_ITOV(ip),
+ error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp,
offset, end_dmi_offset - offset,
AT_DELAY_FLAG(attr_flags), NULL);
if (error)
- return(error);
+ return error;
}
ASSERT(attr_flags & ATTR_NOLOCK ? attr_flags & ATTR_DMI : 1);
if (attr_flags & ATTR_NOLOCK)
need_iolock = 0;
- if (need_iolock)
+ if (need_iolock) {
xfs_ilock(ip, XFS_IOLOCK_EXCL);
+ vn_iowait(vp); /* wait for the completion of any pending DIOs */
+ }
rounding = MAX((__uint8_t)(1 << mp->m_sb.sb_blocklog),
(__uint8_t)NBPP);
ioffset = offset & ~(rounding - 1);
if (ilen & (rounding - 1))
ilen = (ilen + rounding) & ~(rounding - 1);
- xfs_inval_cached_pages(XFS_ITOV(ip), &(ip->i_iocore), ioffset, 0, 0);
+
+ if (VN_CACHED(vp) != 0) {
+ xfs_inval_cached_trace(&ip->i_iocore, ioffset, -1,
+ ctooff(offtoct(ioffset)), -1);
+ VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(ioffset)),
+ -1, FI_REMAPF_LOCKED);
+ }
+
/*
* Need to zero the stuff we're not freeing, on disk.
* If its a realtime file & can't use unwritten extents then we
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
error = XFS_TRANS_RESERVE_QUOTA(mp, tp,
- ip->i_udquot, ip->i_gdquot, resblks, 0, rt ?
- XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS);
+ ip->i_udquot, ip->i_gdquot, resblks, 0,
+ XFS_QMOPT_RES_REGBLKS);
if (error)
goto error1;
/*
* must be a regular file and have write permission
*/
- if (vp->v_type != VREG)
+ if (!VN_ISREG(vp))
return XFS_ERROR(EINVAL);
xfs_ilock(ip, XFS_ILOCK_SHARED);
.vop_read = xfs_read,
#ifdef HAVE_SENDFILE
.vop_sendfile = xfs_sendfile,
+#endif
+#ifdef HAVE_SPLICE
+ .vop_splice_read = xfs_splice_read,
+ .vop_splice_write = xfs_splice_write,
#endif
.vop_write = xfs_write,
.vop_ioctl = xfs_ioctl,
git://git.onelab.eu / linux-2.6.git / blobdiff