return 0;
}
-/*
- * xfs_setattr
- */
int
xfs_setattr(
- xfs_inode_t *ip,
- bhv_vattr_t *vap,
+ struct xfs_inode *ip,
+ struct iattr *iattr,
int flags,
cred_t *credp)
{
xfs_mount_t *mp = ip->i_mount;
+ struct inode *inode = XFS_ITOV(ip);
+ int mask = iattr->ia_valid;
xfs_trans_t *tp;
- int mask;
int code;
uint lock_flags;
uint commit_flags=0;
uid_t uid=0, iuid=0;
gid_t gid=0, igid=0;
int timeflags = 0;
- xfs_prid_t projid=0, iprojid=0;
struct xfs_dquot *udqp, *gdqp, *olddquot1, *olddquot2;
int file_owner;
int need_iolock = 1;
if (mp->m_flags & XFS_MOUNT_RDONLY)
return XFS_ERROR(EROFS);
- /*
- * Cannot set certain attributes.
- */
- mask = vap->va_mask;
- if (mask & XFS_AT_NOSET) {
- return XFS_ERROR(EINVAL);
- }
-
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- /*
- * Timestamps do not need to be logged and hence do not
- * need to be done within a transaction.
- */
- if (mask & XFS_AT_UPDTIMES) {
- ASSERT((mask & ~XFS_AT_UPDTIMES) == 0);
- timeflags = ((mask & XFS_AT_UPDATIME) ? XFS_ICHGTIME_ACC : 0) |
- ((mask & XFS_AT_UPDCTIME) ? XFS_ICHGTIME_CHG : 0) |
- ((mask & XFS_AT_UPDMTIME) ? XFS_ICHGTIME_MOD : 0);
- xfs_ichgtime(ip, timeflags);
- return 0;
- }
-
olddquot1 = olddquot2 = NULL;
udqp = gdqp = NULL;
* 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|XFS_AT_PROJID))) {
+ if (XFS_IS_QUOTA_ON(mp) && (mask & (ATTR_UID|ATTR_GID))) {
uint qflags = 0;
- if ((mask & XFS_AT_UID) && XFS_IS_UQUOTA_ON(mp)) {
- uid = vap->va_uid;
+ if ((mask & ATTR_UID) && XFS_IS_UQUOTA_ON(mp)) {
+ uid = iattr->ia_uid;
qflags |= XFS_QMOPT_UQUOTA;
} else {
uid = ip->i_d.di_uid;
}
- if ((mask & XFS_AT_GID) && XFS_IS_GQUOTA_ON(mp)) {
- gid = vap->va_gid;
+ if ((mask & ATTR_GID) && XFS_IS_GQUOTA_ON(mp)) {
+ gid = iattr->ia_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, projid, qflags,
- &udqp, &gdqp);
+ code = XFS_QM_DQVOPALLOC(mp, ip, uid, gid, ip->i_d.di_projid,
+ qflags, &udqp, &gdqp);
if (code)
return code;
}
*/
tp = NULL;
lock_flags = XFS_ILOCK_EXCL;
- if (flags & ATTR_NOLOCK)
+ if (flags & XFS_ATTR_NOLOCK)
need_iolock = 0;
- if (!(mask & XFS_AT_SIZE)) {
- if ((mask != (XFS_AT_CTIME|XFS_AT_ATIME|XFS_AT_MTIME)) ||
+ if (!(mask & ATTR_SIZE)) {
+ if ((mask != (ATTR_CTIME|ATTR_ATIME|ATTR_MTIME)) ||
(mp->m_flags & XFS_MOUNT_WSYNC)) {
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
commit_flags = 0;
}
} else {
if (DM_EVENT_ENABLED(ip, DM_EVENT_TRUNCATE) &&
- !(flags & ATTR_DMI)) {
+ !(flags & XFS_ATTR_DMI)) {
int dmflags = AT_DELAY_FLAG(flags) | DM_SEM_FLAG_WR;
code = XFS_SEND_DATA(mp, DM_EVENT_TRUNCATE, ip,
- vap->va_size, 0, dmflags, NULL);
+ iattr->ia_size, 0, dmflags, NULL);
if (code) {
lock_flags = 0;
goto error_return;
* Only the owner or users with CAP_FOWNER
* capability may do these things.
*/
- if (mask &
- (XFS_AT_MODE|XFS_AT_XFLAGS|XFS_AT_EXTSIZE|XFS_AT_UID|
- XFS_AT_GID|XFS_AT_PROJID)) {
+ if (mask & (ATTR_MODE|ATTR_UID|ATTR_GID)) {
/*
* CAP_FOWNER overrides the following restrictions:
*
* IDs of the calling process shall match the group owner of
* the file when setting the set-group-ID bit on that file
*/
- if (mask & XFS_AT_MODE) {
+ if (mask & ATTR_MODE) {
mode_t m = 0;
- if ((vap->va_mode & S_ISUID) && !file_owner)
+ if ((iattr->ia_mode & S_ISUID) && !file_owner)
m |= S_ISUID;
- if ((vap->va_mode & S_ISGID) &&
+ if ((iattr->ia_mode & S_ISGID) &&
!in_group_p((gid_t)ip->i_d.di_gid))
m |= S_ISGID;
#if 0
/* Linux allows this, Irix doesn't. */
- if ((vap->va_mode & S_ISVTX) && !S_ISDIR(ip->i_d.di_mode))
+ if ((iattr->ia_mode & S_ISVTX) && !S_ISDIR(ip->i_d.di_mode))
m |= S_ISVTX;
#endif
if (m && !capable(CAP_FSETID))
- vap->va_mode &= ~m;
+ iattr->ia_mode &= ~m;
}
}
* and can change the group id only to a group of which he
* or she is a member.
*/
- if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) {
+ if (mask & (ATTR_UID|ATTR_GID)) {
/*
* These IDs could have changed since we last looked at them.
* But, we're assured that if the ownership did change
* would have changed also.
*/
iuid = ip->i_d.di_uid;
- iprojid = ip->i_d.di_projid;
igid = ip->i_d.di_gid;
- gid = (mask & XFS_AT_GID) ? vap->va_gid : igid;
- uid = (mask & XFS_AT_UID) ? vap->va_uid : iuid;
- projid = (mask & XFS_AT_PROJID) ? (xfs_prid_t)vap->va_projid :
- iprojid;
+ gid = (mask & ATTR_GID) ? iattr->ia_gid : igid;
+ uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
/*
* CAP_CHOWN overrides the following restrictions:
goto error_return;
}
/*
- * Do a quota reservation only if uid/projid/gid is actually
+ * Do a quota reservation only if uid/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)) {
ASSERT(tp);
code = XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, udqp, gdqp,
/*
* Truncate file. Must have write permission and not be a directory.
*/
- if (mask & XFS_AT_SIZE) {
+ if (mask & ATTR_SIZE) {
/* Short circuit the truncate case for zero length files */
- if ((vap->va_size == 0) &&
- (ip->i_size == 0) && (ip->i_d.di_nextents == 0)) {
+ if (iattr->ia_size == 0 &&
+ ip->i_size == 0 && ip->i_d.di_nextents == 0) {
xfs_iunlock(ip, XFS_ILOCK_EXCL);
lock_flags &= ~XFS_ILOCK_EXCL;
- if (mask & XFS_AT_CTIME)
+ if (mask & ATTR_CTIME)
xfs_ichgtime(ip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
code = 0;
goto error_return;
/*
* Change file access or modified times.
*/
- if (mask & (XFS_AT_ATIME|XFS_AT_MTIME)) {
+ if (mask & (ATTR_ATIME|ATTR_MTIME)) {
if (!file_owner) {
- if ((flags & ATTR_UTIME) &&
+ if ((mask & (ATTR_MTIME_SET|ATTR_ATIME_SET)) &&
!capable(CAP_FOWNER)) {
code = XFS_ERROR(EPERM);
goto error_return;
}
}
- /*
- * Change extent size or realtime flag.
- */
- if (mask & (XFS_AT_EXTSIZE|XFS_AT_XFLAGS)) {
- /*
- * Can't change extent size if any extents are allocated.
- */
- 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 change realtime flag if any extents are allocated.
- */
- if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
- (mask & XFS_AT_XFLAGS) &&
- (XFS_IS_REALTIME_INODE(ip)) !=
- (vap->va_xflags & XFS_XFLAG_REALTIME)) {
- code = XFS_ERROR(EINVAL); /* EFBIG? */
- goto error_return;
- }
- /*
- * Extent size must be a multiple of the appropriate block
- * size, if set at all.
- */
- if ((mask & XFS_AT_EXTSIZE) && vap->va_extsize != 0) {
- xfs_extlen_t size;
-
- if (XFS_IS_REALTIME_INODE(ip) ||
- ((mask & XFS_AT_XFLAGS) &&
- (vap->va_xflags & XFS_XFLAG_REALTIME))) {
- size = mp->m_sb.sb_rextsize <<
- mp->m_sb.sb_blocklog;
- } else {
- size = mp->m_sb.sb_blocksize;
- }
- if (vap->va_extsize % size) {
- code = XFS_ERROR(EINVAL);
- goto error_return;
- }
- }
- /*
- * If realtime flag is set then must have realtime data.
- */
- if ((mask & XFS_AT_XFLAGS) &&
- (vap->va_xflags & XFS_XFLAG_REALTIME)) {
- if ((mp->m_sb.sb_rblocks == 0) ||
- (mp->m_sb.sb_rextsize == 0) ||
- (ip->i_d.di_extsize % mp->m_sb.sb_rextsize)) {
- code = XFS_ERROR(EINVAL);
- goto error_return;
- }
- }
-
- /*
- * Can't modify an immutable/append-only file unless
- * we have appropriate permission.
- */
- if ((mask & XFS_AT_XFLAGS) &&
- (ip->i_d.di_flags &
- (XFS_DIFLAG_IMMUTABLE|XFS_DIFLAG_APPEND) ||
- (vap->va_xflags &
- (XFS_XFLAG_IMMUTABLE | XFS_XFLAG_APPEND))) &&
- !capable(CAP_LINUX_IMMUTABLE)) {
- code = XFS_ERROR(EPERM);
- goto error_return;
- }
- }
-
/*
* Now we can make the changes. Before we join the inode
- * to the transaction, if XFS_AT_SIZE is set then take care of
+ * to the transaction, if ATTR_SIZE is set then take care of
* the part of the truncation that must be done without the
* inode lock. This needs to be done before joining the inode
* to the transaction, because the inode cannot be unlocked
* once it is a part of the transaction.
*/
- if (mask & XFS_AT_SIZE) {
+ if (mask & ATTR_SIZE) {
code = 0;
- if ((vap->va_size > ip->i_size) &&
- (flags & ATTR_NOSIZETOK) == 0) {
- code = xfs_igrow_start(ip, vap->va_size, credp);
+ if (iattr->ia_size > ip->i_size) {
+ /*
+ * Do the first part of growing a file: zero any data
+ * in the last block that is beyond the old EOF. We
+ * need to do this before the inode is joined to the
+ * transaction to modify the i_size.
+ */
+ code = xfs_zero_eof(ip, iattr->ia_size, ip->i_size);
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
* not within the range we care about here.
*/
if (!code &&
- (ip->i_size != ip->i_d.di_size) &&
- (vap->va_size > ip->i_d.di_size)) {
+ ip->i_size != ip->i_d.di_size &&
+ iattr->ia_size > ip->i_d.di_size) {
code = xfs_flush_pages(ip,
- ip->i_d.di_size, vap->va_size,
+ ip->i_d.di_size, iattr->ia_size,
XFS_B_ASYNC, FI_NONE);
}
vn_iowait(ip);
if (!code)
- code = xfs_itruncate_data(ip, vap->va_size);
+ code = xfs_itruncate_data(ip, iattr->ia_size);
if (code) {
ASSERT(tp == NULL);
lock_flags &= ~XFS_ILOCK_EXCL;
/*
* Truncate file. Must have write permission and not be a directory.
*/
- if (mask & XFS_AT_SIZE) {
+ if (mask & ATTR_SIZE) {
/*
* Only change the c/mtime if we are changing the size
* or we are explicitly asked to change it. This handles
* the semantic difference between truncate() and ftruncate()
* as implemented in the VFS.
*/
- if (vap->va_size != ip->i_size || (mask & XFS_AT_CTIME))
+ if (iattr->ia_size != ip->i_size || (mask & ATTR_CTIME))
timeflags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG;
- if (vap->va_size > ip->i_size) {
- xfs_igrow_finish(tp, ip, vap->va_size,
- !(flags & ATTR_DMI));
- } else if ((vap->va_size <= ip->i_size) ||
- ((vap->va_size == 0) && ip->i_d.di_nextents)) {
+ if (iattr->ia_size > ip->i_size) {
+ ip->i_d.di_size = iattr->ia_size;
+ ip->i_size = iattr->ia_size;
+ if (!(flags & XFS_ATTR_DMI))
+ xfs_ichgtime(ip, XFS_ICHGTIME_CHG);
+ xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+ } else if (iattr->ia_size <= ip->i_size ||
+ (iattr->ia_size == 0 && ip->i_d.di_nextents)) {
/*
* signal a sync transaction unless
* we're truncating an already unlinked
* file on a wsync filesystem
*/
- code = xfs_itruncate_finish(&tp, ip,
- (xfs_fsize_t)vap->va_size,
+ code = xfs_itruncate_finish(&tp, ip, iattr->ia_size,
XFS_DATA_FORK,
((ip->i_d.di_nlink != 0 ||
!(mp->m_flags & XFS_MOUNT_WSYNC))
/*
* Change file access modes.
*/
- if (mask & XFS_AT_MODE) {
+ if (mask & ATTR_MODE) {
ip->i_d.di_mode &= S_IFMT;
- ip->i_d.di_mode |= vap->va_mode & ~S_IFMT;
+ ip->i_d.di_mode |= iattr->ia_mode & ~S_IFMT;
+
+ inode->i_mode &= S_IFMT;
+ inode->i_mode |= iattr->ia_mode & ~S_IFMT;
xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
timeflags |= XFS_ICHGTIME_CHG;
* and can change the group id only to a group of which he
* or she is a member.
*/
- if (mask & (XFS_AT_UID|XFS_AT_GID|XFS_AT_PROJID)) {
+ if (mask & (ATTR_UID|ATTR_GID)) {
/*
* CAP_FSETID overrides the following restrictions:
*
*/
if (iuid != uid) {
if (XFS_IS_UQUOTA_ON(mp)) {
- ASSERT(mask & XFS_AT_UID);
+ ASSERT(mask & ATTR_UID);
ASSERT(udqp);
olddquot1 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
&ip->i_udquot, udqp);
}
ip->i_d.di_uid = uid;
+ inode->i_uid = uid;
}
if (igid != gid) {
if (XFS_IS_GQUOTA_ON(mp)) {
ASSERT(!XFS_IS_PQUOTA_ON(mp));
- ASSERT(mask & XFS_AT_GID);
+ ASSERT(mask & ATTR_GID);
ASSERT(gdqp);
olddquot2 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
&ip->i_gdquot, gdqp);
}
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
- * the superblock version number since projids didn't
- * exist before DINODE_VERSION_2 and SB_VERSION_NLINK.
- */
- if (ip->i_d.di_version == XFS_DINODE_VERSION_1)
- xfs_bump_ino_vers2(tp, ip);
+ inode->i_gid = gid;
}
xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
/*
* Change file access or modified times.
*/
- if (mask & (XFS_AT_ATIME|XFS_AT_MTIME)) {
- if (mask & XFS_AT_ATIME) {
- ip->i_d.di_atime.t_sec = vap->va_atime.tv_sec;
- ip->i_d.di_atime.t_nsec = vap->va_atime.tv_nsec;
+ if (mask & (ATTR_ATIME|ATTR_MTIME)) {
+ if (mask & ATTR_ATIME) {
+ inode->i_atime = iattr->ia_atime;
+ ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec;
+ ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec;
ip->i_update_core = 1;
timeflags &= ~XFS_ICHGTIME_ACC;
}
- if (mask & XFS_AT_MTIME) {
- ip->i_d.di_mtime.t_sec = vap->va_mtime.tv_sec;
- ip->i_d.di_mtime.t_nsec = vap->va_mtime.tv_nsec;
+ if (mask & ATTR_MTIME) {
+ inode->i_mtime = iattr->ia_mtime;
+ ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec;
+ ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec;
timeflags &= ~XFS_ICHGTIME_MOD;
timeflags |= XFS_ICHGTIME_CHG;
}
- if (tp && (flags & ATTR_UTIME))
+ if (tp && (mask & (ATTR_MTIME_SET|ATTR_ATIME_SET)))
xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
}
/*
- * Change XFS-added attributes.
- */
- if (mask & (XFS_AT_EXTSIZE|XFS_AT_XFLAGS)) {
- if (mask & XFS_AT_EXTSIZE) {
- /*
- * Converting bytes to fs blocks.
- */
- ip->i_d.di_extsize = vap->va_extsize >>
- mp->m_sb.sb_blocklog;
- }
- if (mask & XFS_AT_XFLAGS) {
- uint di_flags;
-
- /* can't set PREALLOC this way, just preserve it */
- di_flags = (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
- if (vap->va_xflags & XFS_XFLAG_IMMUTABLE)
- di_flags |= XFS_DIFLAG_IMMUTABLE;
- if (vap->va_xflags & XFS_XFLAG_APPEND)
- di_flags |= XFS_DIFLAG_APPEND;
- if (vap->va_xflags & XFS_XFLAG_SYNC)
- di_flags |= XFS_DIFLAG_SYNC;
- if (vap->va_xflags & XFS_XFLAG_NOATIME)
- 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 (vap->va_xflags & XFS_XFLAG_NODEFRAG)
- di_flags |= XFS_DIFLAG_NODEFRAG;
- if (vap->va_xflags & XFS_XFLAG_FILESTREAM)
- di_flags |= XFS_DIFLAG_FILESTREAM;
- 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;
- 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;
- if (vap->va_xflags & XFS_XFLAG_EXTSIZE)
- di_flags |= XFS_DIFLAG_EXTSIZE;
- }
- ip->i_d.di_flags = di_flags;
- }
- xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
- timeflags |= XFS_ICHGTIME_CHG;
- }
-
- /*
- * Change file inode change time only if XFS_AT_CTIME set
+ * Change file inode change time only if ATTR_CTIME set
* AND we have been called by a DMI function.
*/
- if ( (flags & ATTR_DMI) && (mask & XFS_AT_CTIME) ) {
- ip->i_d.di_ctime.t_sec = vap->va_ctime.tv_sec;
- ip->i_d.di_ctime.t_nsec = vap->va_ctime.tv_nsec;
+ if ((flags & XFS_ATTR_DMI) && (mask & ATTR_CTIME)) {
+ inode->i_ctime = iattr->ia_ctime;
+ ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec;
+ ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec;
ip->i_update_core = 1;
timeflags &= ~XFS_ICHGTIME_CHG;
}
* Send out timestamp changes that need to be set to the
* current time. Not done when called by a DMI function.
*/
- if (timeflags && !(flags & ATTR_DMI))
+ if (timeflags && !(flags & XFS_ATTR_DMI))
xfs_ichgtime(ip, timeflags);
XFS_STATS_INC(xs_ig_attrchg);
}
if (DM_EVENT_ENABLED(ip, DM_EVENT_ATTRIBUTE) &&
- !(flags & ATTR_DMI)) {
+ !(flags & XFS_ATTR_DMI)) {
(void) XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, ip, DM_RIGHT_NULL,
NULL, DM_RIGHT_NULL, NULL, NULL,
0, 0, AT_DELAY_FLAG(flags));
/*
* xfs_fsync
*
- * This is called to sync the inode and its data out to disk.
- * We need to hold the I/O lock while flushing the data, and
- * the inode lock while flushing the inode. The inode lock CANNOT
- * be held while flushing the data, so acquire after we're done
- * with that.
+ * This is called to sync the inode and its data out to disk. We need to hold
+ * the I/O lock while flushing the data, and the inode lock while flushing the
+ * inode. The inode lock CANNOT be held while flushing the data, so acquire
+ * after we're done with that.
*/
int
xfs_fsync(
- xfs_inode_t *ip,
- int flag,
- xfs_off_t start,
- xfs_off_t stop)
+ xfs_inode_t *ip)
{
xfs_trans_t *tp;
int error;
xfs_itrace_entry(ip);
- ASSERT(start >= 0 && stop >= -1);
-
if (XFS_FORCED_SHUTDOWN(ip->i_mount))
return XFS_ERROR(EIO);
- if (flag & FSYNC_DATA)
- filemap_fdatawait(vn_to_inode(XFS_ITOV(ip))->i_mapping);
+ /* capture size updates in I/O completion before writing the inode. */
+ error = filemap_fdatawait(vn_to_inode(XFS_ITOV(ip))->i_mapping);
+ if (error)
+ return XFS_ERROR(error);
/*
- * We always need to make sure that the required inode state
- * is safe on disk. The vnode might be clean but because
- * of committed transactions that haven't hit the disk yet.
- * Likewise, there could be unflushed non-transactional
- * changes to the inode core that have to go to disk.
+ * We always need to make sure that the required inode state is safe on
+ * disk. The vnode might be clean but we still might need to force the
+ * log because of committed transactions that haven't hit the disk yet.
+ * Likewise, there could be unflushed non-transactional changes to the
+ * inode core that have to go to disk and this requires us to issue
+ * a synchronous transaction to capture these changes correctly.
*
- * The following code depends on one assumption: that
- * any transaction that changes an inode logs the core
- * because it has to change some field in the inode core
- * (typically nextents or nblocks). That assumption
- * implies that any transactions against an inode will
- * catch any non-transactional updates. If inode-altering
- * transactions exist that violate this assumption, the
- * code breaks. Right now, it figures that if the involved
- * update_* field is clear and the inode is unpinned, the
- * inode is clean. Either it's been flushed or it's been
- * committed and the commit has hit the disk unpinning the inode.
- * (Note that xfs_inode_item_format() called at commit clears
- * the update_* fields.)
+ * This code relies on the assumption that if the update_* fields
+ * of the inode are clear and the inode is unpinned then it is clean
+ * and no action is required.
*/
xfs_ilock(ip, XFS_ILOCK_SHARED);
- /* If we are flushing data then we care about update_size
- * being set, otherwise we care about update_core
- */
- if ((flag & FSYNC_DATA) ?
- (ip->i_update_size == 0) :
- (ip->i_update_core == 0)) {
+ if (!(ip->i_update_size || ip->i_update_core)) {
/*
- * Timestamps/size haven't changed since last inode
- * flush or inode transaction commit. That means
- * either nothing got written or a transaction
- * committed which caught the updates. If the
- * latter happened and the transaction hasn't
- * hit the disk yet, the inode will be still
- * be pinned. If it is, force the log.
+ * Timestamps/size haven't changed since last inode flush or
+ * inode transaction commit. That means either nothing got
+ * written or a transaction committed which caught the updates.
+ * If the latter happened and the transaction hasn't hit the
+ * disk yet, the inode will be still be pinned. If it is,
+ * force the log.
*/
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (xfs_ipincount(ip)) {
- _xfs_log_force(ip->i_mount, (xfs_lsn_t)0,
- XFS_LOG_FORCE |
- ((flag & FSYNC_WAIT)
- ? XFS_LOG_SYNC : 0),
+ error = _xfs_log_force(ip->i_mount, (xfs_lsn_t)0,
+ XFS_LOG_FORCE | XFS_LOG_SYNC,
&log_flushed);
} else {
/*
- * If the inode is not pinned and nothing
- * has changed we don't need to flush the
- * cache.
+ * If the inode is not pinned and nothing has changed
+ * we don't need to flush the cache.
*/
changed = 0;
}
- error = 0;
} else {
/*
- * Kick off a transaction to log the inode
- * core to get the updates. Make it
- * sync if FSYNC_WAIT is passed in (which
- * is done by everybody but specfs). The
- * sync transaction will also force the log.
+ * Kick off a transaction to log the inode core to get the
+ * updates. The sync transaction will also force the log.
*/
xfs_iunlock(ip, XFS_ILOCK_SHARED);
tp = xfs_trans_alloc(ip->i_mount, XFS_TRANS_FSYNC_TS);
- if ((error = xfs_trans_reserve(tp, 0,
- XFS_FSYNC_TS_LOG_RES(ip->i_mount),
- 0, 0, 0))) {
+ error = xfs_trans_reserve(tp, 0,
+ XFS_FSYNC_TS_LOG_RES(ip->i_mount), 0, 0, 0);
+ if (error) {
xfs_trans_cancel(tp, 0);
return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
/*
- * Note - it's possible that we might have pushed
- * ourselves out of the way during trans_reserve
- * which would flush the inode. But there's no
- * guarantee that the inode buffer has actually
- * gone out yet (it's delwri). Plus the buffer
- * could be pinned anyway if it's part of an
- * inode in another recent transaction. So we
- * play it safe and fire off the transaction anyway.
+ * Note - it's possible that we might have pushed ourselves out
+ * of the way during trans_reserve which would flush the inode.
+ * But there's no guarantee that the inode buffer has actually
+ * gone out yet (it's delwri). Plus the buffer could be pinned
+ * anyway if it's part of an inode in another recent
+ * transaction. So we play it safe and fire off the
+ * transaction anyway.
*/
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
- if (flag & FSYNC_WAIT)
- xfs_trans_set_sync(tp);
+ xfs_trans_set_sync(tp);
error = _xfs_trans_commit(tp, 0, &log_flushed);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
return VN_INACTIVE_CACHE;
}
-
+/*
+ * Lookups up an inode from "name". If ci_name is not NULL, then a CI match
+ * is allowed, otherwise it has to be an exact match. If a CI match is found,
+ * ci_name->name will point to a the actual name (caller must free) or
+ * will be set to NULL if an exact match is found.
+ */
int
xfs_lookup(
xfs_inode_t *dp,
struct xfs_name *name,
- xfs_inode_t **ipp)
+ xfs_inode_t **ipp,
+ struct xfs_name *ci_name)
{
xfs_ino_t inum;
int error;
return XFS_ERROR(EIO);
lock_mode = xfs_ilock_map_shared(dp);
- error = xfs_dir_lookup(NULL, dp, name, &inum);
+ error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
xfs_iunlock_map_shared(dp, lock_mode);
if (error)
error = xfs_iget(dp->i_mount, NULL, inum, 0, 0, ipp, 0);
if (error)
- goto out;
+ goto out_free_name;
xfs_itrace_ref(*ipp);
return 0;
- out:
+out_free_name:
+ if (ci_name)
+ kmem_free(ci_name->name);
+out:
*ipp = NULL;
return error;
}
#endif
}
-#ifdef DEBUG
-#define REMOVE_DEBUG_TRACE(x) {remove_which_error_return = (x);}
-int remove_which_error_return = 0;
-#else /* ! DEBUG */
-#define REMOVE_DEBUG_TRACE(x)
-#endif /* ! DEBUG */
-
int
xfs_remove(
xfs_inode_t *dp,
{
xfs_mount_t *mp = dp->i_mount;
xfs_trans_t *tp = NULL;
+ int is_dir = S_ISDIR(ip->i_d.di_mode);
int error = 0;
xfs_bmap_free_t free_list;
xfs_fsblock_t first_block;
int committed;
int link_zero;
uint resblks;
+ uint log_count;
xfs_itrace_entry(dp);
+ xfs_itrace_entry(ip);
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
return error;
}
- xfs_itrace_entry(ip);
- xfs_itrace_ref(ip);
-
error = XFS_QM_DQATTACH(mp, dp, 0);
- if (!error)
- error = XFS_QM_DQATTACH(mp, ip, 0);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
+ if (error)
goto std_return;
- }
- tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE);
+ error = XFS_QM_DQATTACH(mp, ip, 0);
+ if (error)
+ goto std_return;
+
+ if (is_dir) {
+ tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR);
+ log_count = XFS_DEFAULT_LOG_COUNT;
+ } else {
+ tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE);
+ log_count = XFS_REMOVE_LOG_COUNT;
+ }
cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
+
/*
* We try to get the real space reservation first,
* allowing for directory btree deletion(s) implying
*/
resblks = XFS_REMOVE_SPACE_RES(mp);
error = xfs_trans_reserve(tp, resblks, XFS_REMOVE_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_REMOVE_LOG_COUNT);
+ XFS_TRANS_PERM_LOG_RES, log_count);
if (error == ENOSPC) {
resblks = 0;
error = xfs_trans_reserve(tp, 0, XFS_REMOVE_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_REMOVE_LOG_COUNT);
+ XFS_TRANS_PERM_LOG_RES, log_count);
}
if (error) {
ASSERT(error != ENOSPC);
- REMOVE_DEBUG_TRACE(__LINE__);
- xfs_trans_cancel(tp, 0);
- return error;
+ cancel_flags = 0;
+ goto out_trans_cancel;
}
error = xfs_lock_dir_and_entry(dp, ip);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
- }
+ if (error)
+ goto out_trans_cancel;
/*
* At this point, we've gotten both the directory and the entry
IHOLD(dp);
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+ /*
+ * If we're removing a directory perform some additional validation.
+ */
+ if (is_dir) {
+ ASSERT(ip->i_d.di_nlink >= 2);
+ if (ip->i_d.di_nlink != 2) {
+ error = XFS_ERROR(ENOTEMPTY);
+ goto out_trans_cancel;
+ }
+ if (!xfs_dir_isempty(ip)) {
+ error = XFS_ERROR(ENOTEMPTY);
+ goto out_trans_cancel;
+ }
+ }
+
/*
* Entry must exist since we did a lookup in xfs_lock_dir_and_entry.
*/
&first_block, &free_list, resblks);
if (error) {
ASSERT(error != ENOENT);
- REMOVE_DEBUG_TRACE(__LINE__);
- goto error1;
+ goto out_bmap_cancel;
}
xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+ /*
+ * Bump the in memory generation count on the parent
+ * directory so that other can know that it has changed.
+ */
dp->i_gen++;
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
- error = xfs_droplink(tp, ip);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- goto error1;
+ if (is_dir) {
+ /*
+ * Drop the link from ip's "..".
+ */
+ error = xfs_droplink(tp, dp);
+ if (error)
+ goto out_bmap_cancel;
+
+ /*
+ * Drop the link from dp to ip.
+ */
+ error = xfs_droplink(tp, ip);
+ if (error)
+ goto out_bmap_cancel;
+ } else {
+ /*
+ * When removing a non-directory we need to log the parent
+ * inode here for the i_gen update. For a directory this is
+ * done implicitly by the xfs_droplink call for the ".." entry.
+ */
+ xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
}
- /* Determine if this is the last link while
+ /*
+ * Drop the "." link from ip to self.
+ */
+ error = xfs_droplink(tp, ip);
+ if (error)
+ goto out_bmap_cancel;
+
+ /*
+ * Determine if this is the last link while
* we are in the transaction.
*/
- link_zero = (ip)->i_d.di_nlink==0;
+ link_zero = (ip->i_d.di_nlink == 0);
/*
* If this is a synchronous mount, make sure that the
* remove transaction goes to disk before returning to
* the user.
*/
- if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
+ if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
xfs_trans_set_sync(tp);
- }
error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- goto error_rele;
- }
+ if (error)
+ goto out_bmap_cancel;
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
if (error)
* will get killed on last close in xfs_close() so we don't
* have to worry about that.
*/
- if (link_zero && xfs_inode_is_filestream(ip))
+ if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
xfs_filestream_deassociate(ip);
xfs_itrace_exit(ip);
+ xfs_itrace_exit(dp);
-/* Fall through to std_return with error = 0 */
std_return:
if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE,
- dp, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL,
- name->name, NULL, ip->i_d.di_mode, error, 0);
+ XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dp, DM_RIGHT_NULL,
+ NULL, DM_RIGHT_NULL, name->name, NULL,
+ ip->i_d.di_mode, error, 0);
}
- return error;
- error1:
- xfs_bmap_cancel(&free_list);
- cancel_flags |= XFS_TRANS_ABORT;
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
+ return error;
- error_rele:
- /*
- * In this case make sure to not release the inode until after
- * the current transaction is aborted. Releasing it beforehand
- * can cause us to go to xfs_inactive and start a recursive
- * transaction which can easily deadlock with the current one.
- */
+ out_bmap_cancel:
xfs_bmap_cancel(&free_list);
cancel_flags |= XFS_TRANS_ABORT;
+ out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
-
goto std_return;
}
goto std_return;
}
-int
-xfs_rmdir(
- xfs_inode_t *dp,
- struct xfs_name *name,
- xfs_inode_t *cdp)
-{
- xfs_mount_t *mp = dp->i_mount;
- xfs_trans_t *tp;
- int error;
- xfs_bmap_free_t free_list;
- xfs_fsblock_t first_block;
- int cancel_flags;
- int committed;
- int last_cdp_link;
- uint resblks;
-
- xfs_itrace_entry(dp);
-
- if (XFS_FORCED_SHUTDOWN(mp))
- return XFS_ERROR(EIO);
-
- if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE,
- dp, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL, name->name,
- NULL, cdp->i_d.di_mode, 0, 0);
- if (error)
- return XFS_ERROR(error);
- }
-
- /*
- * Get the dquots for the inodes.
- */
- error = XFS_QM_DQATTACH(mp, dp, 0);
- if (!error)
- error = XFS_QM_DQATTACH(mp, cdp, 0);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- goto std_return;
- }
-
- tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR);
- cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
- /*
- * We try to get the real space reservation first,
- * allowing for directory btree deletion(s) implying
- * possible bmap insert(s). If we can't get the space
- * reservation then we use 0 instead, and avoid the bmap
- * btree insert(s) in the directory code by, if the bmap
- * insert tries to happen, instead trimming the LAST
- * block from the directory.
- */
- resblks = XFS_REMOVE_SPACE_RES(mp);
- error = xfs_trans_reserve(tp, resblks, XFS_REMOVE_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_DEFAULT_LOG_COUNT);
- if (error == ENOSPC) {
- resblks = 0;
- error = xfs_trans_reserve(tp, 0, XFS_REMOVE_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_DEFAULT_LOG_COUNT);
- }
- if (error) {
- ASSERT(error != ENOSPC);
- cancel_flags = 0;
- goto error_return;
- }
- XFS_BMAP_INIT(&free_list, &first_block);
-
- /*
- * Now lock the child directory inode and the parent directory
- * inode in the proper order. This will take care of validating
- * that the directory entry for the child directory inode has
- * not changed while we were obtaining a log reservation.
- */
- error = xfs_lock_dir_and_entry(dp, cdp);
- if (error) {
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
- }
-
- IHOLD(dp);
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
-
- IHOLD(cdp);
- xfs_trans_ijoin(tp, cdp, XFS_ILOCK_EXCL);
-
- ASSERT(cdp->i_d.di_nlink >= 2);
- if (cdp->i_d.di_nlink != 2) {
- error = XFS_ERROR(ENOTEMPTY);
- goto error_return;
- }
- if (!xfs_dir_isempty(cdp)) {
- error = XFS_ERROR(ENOTEMPTY);
- goto error_return;
- }
-
- error = xfs_dir_removename(tp, dp, name, cdp->i_ino,
- &first_block, &free_list, resblks);
- if (error)
- goto error1;
-
- xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
-
- /*
- * Bump the in memory generation count on the parent
- * directory so that other can know that it has changed.
- */
- dp->i_gen++;
-
- /*
- * Drop the link from cdp's "..".
- */
- error = xfs_droplink(tp, dp);
- if (error) {
- goto error1;
- }
-
- /*
- * Drop the link from dp to cdp.
- */
- error = xfs_droplink(tp, cdp);
- if (error) {
- goto error1;
- }
-
- /*
- * Drop the "." link from cdp to self.
- */
- error = xfs_droplink(tp, cdp);
- if (error) {
- goto error1;
- }
-
- /* Determine these before committing transaction */
- last_cdp_link = (cdp)->i_d.di_nlink==0;
-
- /*
- * If this is a synchronous mount, make sure that the
- * rmdir transaction goes to disk before returning to
- * the user.
- */
- if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
- xfs_trans_set_sync(tp);
- }
-
- error = xfs_bmap_finish (&tp, &free_list, &committed);
- if (error) {
- xfs_bmap_cancel(&free_list);
- xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES |
- XFS_TRANS_ABORT));
- goto std_return;
- }
-
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error) {
- goto std_return;
- }
-
-
- /* Fall through to std_return with error = 0 or the errno
- * from xfs_trans_commit. */
- std_return:
- if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE,
- dp, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL,
- name->name, NULL, cdp->i_d.di_mode,
- error, 0);
- }
- return error;
-
- error1:
- xfs_bmap_cancel(&free_list);
- cancel_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
-
- error_return:
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
-}
-
int
xfs_symlink(
xfs_inode_t *dp,
{
xfs_perag_t *pag = xfs_get_perag(ip->i_mount, ip->i_ino);
bhv_vnode_t *vp = XFS_ITOV_NULL(ip);
- int error;
if (vp && VN_BAD(vp))
goto reclaim;
xfs_iflock(ip);
}
- if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
- if (ip->i_update_core ||
- ((ip->i_itemp != NULL) &&
- (ip->i_itemp->ili_format.ilf_fields != 0))) {
- error = xfs_iflush(ip, sync_mode);
- /*
- * If we hit an error, typically because of filesystem
- * shutdown, we don't need to let vn_reclaim to know
- * because we're gonna reclaim the inode anyway.
- */
- if (error) {
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- goto reclaim;
- }
- xfs_iflock(ip); /* synchronize with xfs_iflush_done */
- }
-
- ASSERT(ip->i_update_core == 0);
- ASSERT(ip->i_itemp == NULL ||
- ip->i_itemp->ili_format.ilf_fields == 0);
+ /*
+ * In the case of a forced shutdown we rely on xfs_iflush() to
+ * wait for the inode to be unpinned before returning an error.
+ */
+ if (xfs_iflush(ip, sync_mode) == 0) {
+ /* synchronize with xfs_iflush_done */
+ xfs_iflock(ip);
+ xfs_ifunlock(ip);
}
- xfs_ifunlock(ip);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
reclaim:
/* Generate a DMAPI event if needed. */
if (alloc_type != 0 && offset < ip->i_size &&
- (attr_flags&ATTR_DMI) == 0 &&
+ (attr_flags & XFS_ATTR_DMI) == 0 &&
DM_EVENT_ENABLED(ip, DM_EVENT_WRITE)) {
xfs_off_t end_dmi_offset;
allocatesize_fsb -= allocated_fsb;
}
dmapi_enospc_check:
- if (error == ENOSPC && (attr_flags & ATTR_DMI) == 0 &&
+ if (error == ENOSPC && (attr_flags & XFS_ATTR_DMI) == 0 &&
DM_EVENT_ENABLED(ip, DM_EVENT_NOSPACE)) {
error = XFS_SEND_NAMESP(mp, DM_EVENT_NOSPACE,
ip, DM_RIGHT_NULL,
end_dmi_offset = offset + len;
endoffset_fsb = XFS_B_TO_FSBT(mp, end_dmi_offset);
- if (offset < ip->i_size && (attr_flags & ATTR_DMI) == 0 &&
+ if (offset < ip->i_size && (attr_flags & XFS_ATTR_DMI) == 0 &&
DM_EVENT_ENABLED(ip, DM_EVENT_WRITE)) {
if (end_dmi_offset > ip->i_size)
end_dmi_offset = ip->i_size;
return error;
}
- if (attr_flags & ATTR_NOLOCK)
+ if (attr_flags & XFS_ATTR_NOLOCK)
need_iolock = 0;
if (need_iolock) {
xfs_ilock(ip, XFS_IOLOCK_EXCL);
xfs_off_t startoffset;
xfs_off_t llen;
xfs_trans_t *tp;
- bhv_vattr_t va;
+ struct iattr iattr;
xfs_itrace_entry(ip);
break;
}
- va.va_mask = XFS_AT_SIZE;
- va.va_size = startoffset;
+ iattr.ia_valid = ATTR_SIZE;
+ iattr.ia_size = startoffset;
- error = xfs_setattr(ip, &va, attr_flags, credp);
+ error = xfs_setattr(ip, &iattr, attr_flags, credp);
if (error)
return error;
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, ip);
- if ((attr_flags & ATTR_DMI) == 0) {
+ if ((attr_flags & XFS_ATTR_DMI) == 0) {
ip->i_d.di_mode &= ~S_ISUID;
/*