#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_bmap.h"
+#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_rw.h"
#include "xfs_error.h"
#include "xfs_vnodeops.h"
int
-xfs_open(
- xfs_inode_t *ip)
-{
- int mode;
-
- if (XFS_FORCED_SHUTDOWN(ip->i_mount))
- return XFS_ERROR(EIO);
-
- /*
- * 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 (S_ISDIR(ip->i_d.di_mode) && 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);
- xfs_iunlock(ip, mode);
- }
- return 0;
-}
-
-/*
- * xfs_setattr
- */
-int
xfs_setattr(
- xfs_inode_t *ip,
- bhv_vattr_t *vap,
- int flags,
- cred_t *credp)
+ struct xfs_inode *ip,
+ struct iattr *iattr,
+ int flags)
{
xfs_mount_t *mp = ip->i_mount;
+ struct inode *inode = VFS_I(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;
xfs_itrace_entry(ip);
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;
- }
+ code = -inode_change_ok(inode, iattr);
+ if (code)
+ return code;
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_vop_dqalloc(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;
xfs_ilock(ip, lock_flags);
- /* boolean: are we the file owner? */
- file_owner = (current_fsuid(credp) == ip->i_d.di_uid);
-
- /*
- * Change various properties of a file.
- * 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)) {
- /*
- * CAP_FOWNER overrides the following restrictions:
- *
- * The user ID of the calling process must be equal
- * to the file owner ID, except in cases where the
- * CAP_FSETID capability is applicable.
- */
- if (!file_owner && !capable(CAP_FOWNER)) {
- code = XFS_ERROR(EPERM);
- goto error_return;
- }
-
- /*
- * CAP_FSETID overrides the following restrictions:
- *
- * The effective user ID of the calling process shall match
- * the file owner when setting the set-user-ID and
- * set-group-ID bits on that file.
- *
- * The effective group ID or one of the supplementary group
- * 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) {
- mode_t m = 0;
-
- if ((vap->va_mode & S_ISUID) && !file_owner)
- m |= S_ISUID;
- if ((vap->va_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))
- m |= S_ISVTX;
-#endif
- if (m && !capable(CAP_FSETID))
- vap->va_mode &= ~m;
- }
- }
-
/*
* Change file ownership. Must be the owner or privileged.
- * If the system was configured with the "restricted_chown"
- * option, the owner is not permitted to give away the file,
- * 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:
- *
- * If _POSIX_CHOWN_RESTRICTED is defined, this capability
- * shall override the restriction that a process cannot
- * change the user ID of a file it owns and the restriction
- * 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.
- */
- if (restricted_chown &&
- (iuid != uid || (igid != gid &&
- !in_group_p((gid_t)gid))) &&
- !capable(CAP_CHOWN)) {
- code = XFS_ERROR(EPERM);
- 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)) {
+ if (XFS_IS_QUOTA_RUNNING(mp) &&
+ ((XFS_IS_UQUOTA_ON(mp) && iuid != uid) ||
+ (XFS_IS_GQUOTA_ON(mp) && igid != gid))) {
ASSERT(tp);
- code = XFS_QM_DQVOPCHOWNRESV(mp, tp, ip, udqp, gdqp,
+ code = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp,
capable(CAP_FOWNER) ?
XFS_QMOPT_FORCE_RES : 0);
if (code) /* out of quota */
/*
* 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;
code = XFS_ERROR(EINVAL);
goto error_return;
}
- /*
- * Make sure that the dquots are attached to the inode.
- */
- if ((code = XFS_QM_DQATTACH(mp, ip, XFS_QMOPT_ILOCKED)))
- goto error_return;
- }
-
- /*
- * Change file access or modified times.
- */
- if (mask & (XFS_AT_ATIME|XFS_AT_MTIME)) {
- if (!file_owner) {
- if ((flags & ATTR_UTIME) &&
- !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.
+ * Make sure that the dquots are attached to the inode.
*/
- 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? */
+ code = xfs_qm_dqattach_locked(ip, 0);
+ if (code)
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.
+ * Now we can make the changes. Before we join the inode
+ * 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_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
- * 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) {
- 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);
}
/* wait for all I/O to complete */
- vn_iowait(ip);
+ xfs_ioend_wait(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;
}
commit_flags = XFS_TRANS_RELEASE_LOG_RES;
xfs_ilock(ip, XFS_ILOCK_EXCL);
- }
- if (tp) {
xfs_trans_ijoin(tp, ip, lock_flags);
xfs_trans_ihold(tp, ip);
- }
- /*
- * Truncate file. Must have write permission and not be a directory.
- */
- if (mask & XFS_AT_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))
*/
xfs_iflags_set(ip, XFS_ITRUNCATED);
}
- }
-
- /*
- * Change file access modes.
- */
- if (mask & XFS_AT_MODE) {
- ip->i_d.di_mode &= S_IFMT;
- ip->i_d.di_mode |= vap->va_mode & ~S_IFMT;
-
- xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
- timeflags |= XFS_ICHGTIME_CHG;
+ } else if (tp) {
+ xfs_trans_ijoin(tp, ip, lock_flags);
+ xfs_trans_ihold(tp, ip);
}
/*
* Change file ownership. Must be the owner or privileged.
- * If the system was configured with the "restricted_chown"
- * option, the owner is not permitted to give away the file,
- * 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:
*
* in the transaction.
*/
if (iuid != uid) {
- if (XFS_IS_UQUOTA_ON(mp)) {
- ASSERT(mask & XFS_AT_UID);
+ if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_UQUOTA_ON(mp)) {
+ ASSERT(mask & ATTR_UID);
ASSERT(udqp);
- olddquot1 = XFS_QM_DQVOPCHOWN(mp, tp, ip,
+ olddquot1 = xfs_qm_vop_chown(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)) {
+ if (XFS_IS_QUOTA_RUNNING(mp) && 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,
+ olddquot2 = xfs_qm_vop_chown(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);
timeflags |= XFS_ICHGTIME_CHG;
}
+ /*
+ * Change file access modes.
+ */
+ if (mask & ATTR_MODE) {
+ umode_t mode = iattr->ia_mode;
+
+ if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
+ mode &= ~S_ISGID;
+
+ ip->i_d.di_mode &= S_IFMT;
+ ip->i_d.di_mode |= mode & ~S_IFMT;
+
+ inode->i_mode &= S_IFMT;
+ inode->i_mode |= mode & ~S_IFMT;
+
+ xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+ timeflags |= XFS_ICHGTIME_CHG;
+ }
/*
* 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);
/*
* Release any dquot(s) the inode had kept before chown.
*/
- XFS_QM_DQRELE(mp, olddquot1);
- XFS_QM_DQRELE(mp, olddquot2);
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ xfs_qm_dqrele(olddquot1);
+ xfs_qm_dqrele(olddquot2);
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
- if (code) {
+ if (code)
return code;
+
+ /*
+ * XXX(hch): Updating the ACL entries is not atomic vs the i_mode
+ * update. We could avoid this with linked transactions
+ * and passing down the transaction pointer all the way
+ * to attr_set. No previous user of the generic
+ * Posix ACL code seems to care about this issue either.
+ */
+ if ((mask & ATTR_MODE) && !(flags & XFS_ATTR_NOACL)) {
+ code = -xfs_acl_chmod(inode);
+ if (code)
+ return XFS_ERROR(code);
}
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));
commit_flags |= XFS_TRANS_ABORT;
/* FALLTHROUGH */
error_return:
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
if (tp) {
xfs_trans_cancel(tp, commit_flags);
}
d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock);
byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount);
- bp = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0);
+ bp = xfs_buf_read_flags(mp->m_ddev_targp, d, BTOBB(byte_cnt),
+ XBF_LOCK | XBF_MAPPED |
+ XBF_DONT_BLOCK);
error = XFS_BUF_GETERROR(bp);
if (error) {
xfs_ioerror_alert("xfs_readlink",
/*
* 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 = xfs_wait_on_pages(ip, 0, -1);
+ 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);
/*
* Attach the dquots to the inode up front.
*/
- if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
return error;
/*
* Find the block(s) so we can inval and unmap them.
*/
done = 0;
- XFS_BMAP_INIT(&free_list, &first_block);
+ xfs_bmap_init(&free_list, &first_block);
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,
goto error0;
}
/*
+ * 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);
+
+ /*
* Remove the memory for extent descriptions (just bookkeeping).
*/
if (ip->i_df.if_bytes)
int error;
xfs_mount_t *mp;
- ASSERT(ismrlocked(&ip->i_iolock, MR_UPDATE));
+ ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
tp = *tpp;
mp = ip->i_mount;
ASSERT(ip->i_d.di_forkoff != 0);
xfs_release(
xfs_inode_t *ip)
{
- bhv_vnode_t *vp = XFS_ITOV(ip);
xfs_mount_t *mp = ip->i_mount;
int error;
* be exposed to that problem.
*/
truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED);
- if (truncated && VN_DIRTY(vp) && ip->i_delayed_blks > 0)
+ if (truncated && VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0)
xfs_flush_pages(ip, 0, -1, XFS_B_ASYNC, FI_NONE);
}
if (ip->i_d.di_nlink != 0) {
if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
- ((ip->i_size > 0) || (VN_CACHED(vp) > 0 ||
+ ((ip->i_size > 0) || (VN_CACHED(VFS_I(ip)) > 0 ||
ip->i_delayed_blks > 0)) &&
(ip->i_df.if_flags & XFS_IFEXTENTS)) &&
(!(ip->i_d.di_flags &
xfs_inactive(
xfs_inode_t *ip)
{
- bhv_vnode_t *vp = XFS_ITOV(ip);
xfs_bmap_free_t free_list;
xfs_fsblock_t first_block;
int committed;
* If the inode is already free, then there can be nothing
* to clean up here.
*/
- if (ip->i_d.di_mode == 0 || VN_BAD(vp)) {
+ if (ip->i_d.di_mode == 0 || is_bad_inode(VFS_I(ip))) {
ASSERT(ip->i_df.if_real_bytes == 0);
ASSERT(ip->i_df.if_broot_bytes == 0);
return VN_INACTIVE_CACHE;
if (ip->i_d.di_nlink != 0) {
if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) &&
- ((ip->i_size > 0) || (VN_CACHED(vp) > 0 ||
+ ((ip->i_size > 0) || (VN_CACHED(VFS_I(ip)) > 0 ||
ip->i_delayed_blks > 0)) &&
(ip->i_df.if_flags & XFS_IFEXTENTS) &&
(!(ip->i_d.di_flags &
ASSERT(ip->i_d.di_nlink == 0);
- if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
return VN_INACTIVE_CACHE;
tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
/*
* Free the inode.
*/
- XFS_BMAP_INIT(&free_list, &first_block);
+ xfs_bmap_init(&free_list, &first_block);
error = xfs_ifree(tp, ip, &free_list);
if (error) {
/*
/*
* Credit the quota account(s). The inode is gone.
*/
- XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, XFS_TRANS_DQ_ICOUNT, -1);
+ xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, -1);
/*
* Just ignore errors at this point. There is nothing we can
xfs_fs_cmn_err(CE_NOTE, mp, "xfs_inactive: "
"xfs_trans_commit() returned error %d", error);
}
+
/*
* Release the dquots held by inode, if any.
*/
- XFS_QM_DQDETACH(mp, ip);
-
+ xfs_qm_dqdetach(ip);
xfs_iunlock(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
out:
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;
}
xfs_inode_t **ipp,
cred_t *credp)
{
- xfs_mount_t *mp = dp->i_mount;
- xfs_inode_t *ip;
- xfs_trans_t *tp;
+ int is_dir = S_ISDIR(mode);
+ struct xfs_mount *mp = dp->i_mount;
+ struct xfs_inode *ip = NULL;
+ struct xfs_trans *tp = NULL;
int error;
xfs_bmap_free_t free_list;
xfs_fsblock_t first_block;
boolean_t unlock_dp_on_error = B_FALSE;
- int dm_event_sent = 0;
uint cancel_flags;
int committed;
xfs_prid_t prid;
- struct xfs_dquot *udqp, *gdqp;
+ struct xfs_dquot *udqp = NULL;
+ struct xfs_dquot *gdqp = NULL;
uint resblks;
+ uint log_res;
+ uint log_count;
- ASSERT(!*ipp);
xfs_itrace_entry(dp);
+ if (XFS_FORCED_SHUTDOWN(mp))
+ return XFS_ERROR(EIO);
+
if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) {
error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE,
dp, DM_RIGHT_NULL, NULL,
if (error)
return error;
- dm_event_sent = 1;
}
- if (XFS_FORCED_SHUTDOWN(mp))
- return XFS_ERROR(EIO);
-
- /* Return through std_return after this point. */
-
- udqp = gdqp = NULL;
if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
prid = dp->i_d.di_projid;
else
- prid = (xfs_prid_t)dfltprid;
+ prid = dfltprid;
/*
* Make sure that we have allocated dquot(s) on disk.
*/
- error = XFS_QM_DQVOPALLOC(mp, dp,
- current_fsuid(credp), current_fsgid(credp), prid,
- XFS_QMOPT_QUOTALL|XFS_QMOPT_INHERIT, &udqp, &gdqp);
+ error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
+ XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, &udqp, &gdqp);
if (error)
goto std_return;
- ip = NULL;
+ if (is_dir) {
+ rdev = 0;
+ resblks = XFS_MKDIR_SPACE_RES(mp, name->len);
+ log_res = XFS_MKDIR_LOG_RES(mp);
+ log_count = XFS_MKDIR_LOG_COUNT;
+ tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR);
+ } else {
+ resblks = XFS_CREATE_SPACE_RES(mp, name->len);
+ log_res = XFS_CREATE_LOG_RES(mp);
+ log_count = XFS_CREATE_LOG_COUNT;
+ tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE);
+ }
- tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE);
cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
- resblks = XFS_CREATE_SPACE_RES(mp, name->len);
+
/*
* Initially assume that the file does not exist and
* reserve the resources for that case. If that is not
* the case we'll drop the one we have and get a more
* appropriate transaction later.
*/
- error = xfs_trans_reserve(tp, resblks, XFS_CREATE_LOG_RES(mp), 0,
+ error = xfs_trans_reserve(tp, resblks, log_res, 0,
+ XFS_TRANS_PERM_LOG_RES, log_count);
+ if (error == ENOSPC) {
+ /* flush outstanding delalloc blocks and retry */
+ xfs_flush_inodes(dp);
+ error = xfs_trans_reserve(tp, resblks, XFS_CREATE_LOG_RES(mp), 0,
XFS_TRANS_PERM_LOG_RES, XFS_CREATE_LOG_COUNT);
+ }
if (error == ENOSPC) {
+ /* No space at all so try a "no-allocation" reservation */
resblks = 0;
- error = xfs_trans_reserve(tp, 0, XFS_CREATE_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_CREATE_LOG_COUNT);
+ error = xfs_trans_reserve(tp, 0, log_res, 0,
+ XFS_TRANS_PERM_LOG_RES, log_count);
}
if (error) {
cancel_flags = 0;
- goto error_return;
+ goto out_trans_cancel;
}
xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
unlock_dp_on_error = B_TRUE;
- XFS_BMAP_INIT(&free_list, &first_block);
+ /*
+ * Check for directory link count overflow.
+ */
+ if (is_dir && dp->i_d.di_nlink >= XFS_MAXLINK) {
+ error = XFS_ERROR(EMLINK);
+ goto out_trans_cancel;
+ }
- ASSERT(ip == NULL);
+ xfs_bmap_init(&free_list, &first_block);
/*
* Reserve disk quota and the inode.
*/
- error = XFS_TRANS_RESERVE_QUOTA(mp, tp, udqp, gdqp, resblks, 1, 0);
+ error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp, resblks, 1, 0);
if (error)
- goto error_return;
+ goto out_trans_cancel;
error = xfs_dir_canenter(tp, dp, name, resblks);
if (error)
- goto error_return;
- error = xfs_dir_ialloc(&tp, dp, mode, 1,
- rdev, credp, prid, resblks > 0,
- &ip, &committed);
+ goto out_trans_cancel;
+
+ /*
+ * A newly created regular or special file just has one directory
+ * entry pointing to them, but a directory also the "." entry
+ * pointing to itself.
+ */
+ error = xfs_dir_ialloc(&tp, dp, mode, is_dir ? 2 : 1, rdev, credp,
+ prid, resblks > 0, &ip, &committed);
if (error) {
if (error == ENOSPC)
- goto error_return;
- goto abort_return;
+ goto out_trans_cancel;
+ goto out_trans_abort;
}
- xfs_itrace_ref(ip);
/*
* At this point, we've gotten a newly allocated inode.
* It is locked (and joined to the transaction).
*/
-
- ASSERT(ismrlocked (&ip->i_lock, MR_UPDATE));
+ xfs_itrace_ref(ip);
+ ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
/*
* Now we join the directory inode to the transaction. We do not do it
resblks - XFS_IALLOC_SPACE_RES(mp) : 0);
if (error) {
ASSERT(error != ENOSPC);
- goto abort_return;
+ goto out_trans_abort;
}
xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
+ if (is_dir) {
+ error = xfs_dir_init(tp, ip, dp);
+ if (error)
+ goto out_bmap_cancel;
+
+ error = xfs_bumplink(tp, dp);
+ if (error)
+ goto out_bmap_cancel;
+ }
+
/*
* If this is a synchronous mount, make sure that the
* create 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);
- }
-
- dp->i_gen++;
/*
* Attach the dquot(s) to the inodes and modify them incore.
* These ids of the inode couldn't have changed since the new
* inode has been locked ever since it was created.
*/
- XFS_QM_DQVOPCREATE(mp, tp, ip, udqp, gdqp);
+ xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp);
/*
* xfs_trans_commit normally decrements the vnode ref count
IHOLD(ip);
error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error) {
- xfs_bmap_cancel(&free_list);
- goto abort_rele;
- }
+ if (error)
+ goto out_abort_rele;
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
if (error) {
IRELE(ip);
- tp = NULL;
- goto error_return;
+ goto out_dqrele;
}
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
*ipp = ip;
/* Fallthrough to std_return with error = 0 */
-
-std_return:
- if ((*ipp || (error != 0 && dm_event_sent != 0)) &&
- DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE,
- dp, DM_RIGHT_NULL,
- *ipp ? ip : NULL,
- DM_RIGHT_NULL, name->name, NULL,
- mode, error, 0);
+ std_return:
+ if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) {
+ XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, dp, DM_RIGHT_NULL,
+ ip, DM_RIGHT_NULL, name->name, NULL, mode,
+ error, 0);
}
+
return error;
- abort_return:
+ out_bmap_cancel:
+ xfs_bmap_cancel(&free_list);
+ out_trans_abort:
cancel_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
-
- error_return:
- if (tp != NULL)
- xfs_trans_cancel(tp, cancel_flags);
-
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ out_trans_cancel:
+ xfs_trans_cancel(tp, cancel_flags);
+ out_dqrele:
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
if (unlock_dp_on_error)
xfs_iunlock(dp, XFS_ILOCK_EXCL);
goto std_return;
- abort_rele:
+ out_abort_rele:
/*
* Wait until after the current transaction is aborted to
* release the inode. This prevents recursive transactions
* and deadlocks from xfs_inactive.
*/
+ xfs_bmap_cancel(&free_list);
cancel_flags |= XFS_TRANS_ABORT;
xfs_trans_cancel(tp, cancel_flags);
IRELE(ip);
-
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
-
- goto std_return;
+ unlock_dp_on_error = B_FALSE;
+ goto out_dqrele;
}
#ifdef DEBUG
-/*
- * Some counters to see if (and how often) we are hitting some deadlock
- * prevention code paths.
- */
-
-int xfs_rm_locks;
-int xfs_rm_lock_delays;
-int xfs_rm_attempts;
+int xfs_locked_n;
+int xfs_small_retries;
+int xfs_middle_retries;
+int xfs_lots_retries;
+int xfs_lock_delays;
#endif
/*
- * The following routine will lock the inodes associated with the
- * directory and the named entry in the directory. The locks are
- * acquired in increasing inode number.
- *
- * If the entry is "..", then only the directory is locked. The
- * vnode ref count will still include that from the .. entry in
- * this case.
- *
- * There is a deadlock we need to worry about. If the locked directory is
- * in the AIL, it might be blocking up the log. The next inode we lock
- * could be already locked by another thread waiting for log space (e.g
- * a permanent log reservation with a long running transaction (see
- * xfs_itruncate_finish)). To solve this, we must check if the directory
- * is in the ail and use lock_nowait. If we can't lock, we need to
- * drop the inode lock on the directory and try again. xfs_iunlock will
- * potentially push the tail if we were holding up the log.
+ * Bump the subclass so xfs_lock_inodes() acquires each lock with
+ * a different value
*/
-STATIC int
-xfs_lock_dir_and_entry(
- xfs_inode_t *dp,
- xfs_inode_t *ip) /* inode of entry 'name' */
+static inline int
+xfs_lock_inumorder(int lock_mode, int subclass)
{
- int attempts;
- xfs_ino_t e_inum;
- xfs_inode_t *ips[2];
- xfs_log_item_t *lp;
-
-#ifdef DEBUG
- xfs_rm_locks++;
-#endif
- attempts = 0;
-
-again:
- xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
-
- e_inum = ip->i_ino;
-
- xfs_itrace_ref(ip);
-
- /*
- * We want to lock in increasing inum. Since we've already
- * acquired the lock on the directory, we may need to release
- * if if the inum of the entry turns out to be less.
- */
- if (e_inum > dp->i_ino) {
- /*
- * We are already in the right order, so just
- * lock on the inode of the entry.
- * We need to use nowait if dp is in the AIL.
- */
-
- lp = (xfs_log_item_t *)dp->i_itemp;
- if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
- if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
- attempts++;
-#ifdef DEBUG
- xfs_rm_attempts++;
-#endif
-
- /*
- * Unlock dp and try again.
- * xfs_iunlock will try to push the tail
- * if the inode is in the AIL.
- */
-
- xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
- if ((attempts % 5) == 0) {
- delay(1); /* Don't just spin the CPU */
-#ifdef DEBUG
- xfs_rm_lock_delays++;
-#endif
- }
- goto again;
- }
- } else {
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- }
- } else if (e_inum < dp->i_ino) {
- xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
- ips[0] = ip;
- ips[1] = dp;
- xfs_lock_inodes(ips, 2, 0, XFS_ILOCK_EXCL);
- }
- /* else e_inum == dp->i_ino */
- /* This can happen if we're asked to lock /x/..
- * the entry is "..", which is also the parent directory.
- */
-
- return 0;
-}
-
-#ifdef DEBUG
-int xfs_locked_n;
-int xfs_small_retries;
-int xfs_middle_retries;
-int xfs_lots_retries;
-int xfs_lock_delays;
-#endif
-
-/*
- * Bump the subclass so xfs_lock_inodes() acquires each lock with
- * a different value
- */
-static inline int
-xfs_lock_inumorder(int lock_mode, int subclass)
-{
- if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
- lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_IOLOCK_SHIFT;
- if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))
- lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_ILOCK_SHIFT;
+ if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
+ lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_IOLOCK_SHIFT;
+ if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))
+ lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_ILOCK_SHIFT;
return lock_mode;
}
xfs_lock_inodes(
xfs_inode_t **ips,
int inodes,
- int first_locked,
uint lock_mode)
{
int attempts = 0, i, j, try_lock;
ASSERT(ips && (inodes >= 2)); /* we need at least two */
- if (first_locked) {
- try_lock = 1;
- i = 1;
- } else {
- try_lock = 0;
- i = 0;
- }
+ try_lock = 0;
+ i = 0;
again:
for (; i < inodes; i++) {
#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 */
+/*
+ * xfs_lock_two_inodes() can only be used to lock one type of lock
+ * at a time - the iolock or the ilock, but not both at once. If
+ * we lock both at once, lockdep will report false positives saying
+ * we have violated locking orders.
+ */
+void
+xfs_lock_two_inodes(
+ xfs_inode_t *ip0,
+ xfs_inode_t *ip1,
+ uint lock_mode)
+{
+ xfs_inode_t *temp;
+ int attempts = 0;
+ xfs_log_item_t *lp;
+
+ if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
+ ASSERT((lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) == 0);
+ ASSERT(ip0->i_ino != ip1->i_ino);
+
+ if (ip0->i_ino > ip1->i_ino) {
+ temp = ip0;
+ ip0 = ip1;
+ ip1 = temp;
+ }
+
+ again:
+ xfs_ilock(ip0, xfs_lock_inumorder(lock_mode, 0));
+
+ /*
+ * If the first lock we have locked is in the AIL, we must TRY to get
+ * the second lock. If we can't get it, we must release the first one
+ * and try again.
+ */
+ lp = (xfs_log_item_t *)ip0->i_itemp;
+ if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
+ if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(lock_mode, 1))) {
+ xfs_iunlock(ip0, lock_mode);
+ if ((++attempts % 5) == 0)
+ delay(1); /* Don't just spin the CPU */
+ goto again;
+ }
+ } else {
+ xfs_ilock(ip1, xfs_lock_inumorder(lock_mode, 1));
+ }
+}
int
xfs_remove(
{
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;
}
- /*
- * We need to get a reference to ip before we get our log
- * reservation. The reason for this is that we cannot call
- * xfs_iget for an inode for which we do not have a reference
- * once we've acquired a log reservation. This is because the
- * inode we are trying to get might be in xfs_inactive going
- * 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 reference
- * to the inode before getting our log reservation.
- */
- IHOLD(ip);
-
- xfs_itrace_entry(ip);
- xfs_itrace_ref(ip);
+ error = xfs_qm_dqattach(dp, 0);
+ if (error)
+ goto std_return;
- error = XFS_QM_DQATTACH(mp, dp, 0);
- if (!error && dp != ip)
- error = XFS_QM_DQATTACH(mp, ip, 0);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- IRELE(ip);
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
goto std_return;
- }
- tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE);
+ 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);
- IRELE(ip);
- 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);
- IRELE(ip);
- goto std_return;
- }
+ xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL);
/*
* At this point, we've gotten both the directory and the entry
* inodes locked.
*/
+ IHOLD(ip);
xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
- if (dp != ip) {
- /*
- * Increment vnode ref count only in this case since
- * there's an extra vnode reference in the case where
- * dp == ip.
- */
- IHOLD(dp);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- }
+
+ IHOLD(dp);
+ xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
/*
- * Entry must exist since we did a lookup in xfs_lock_dir_and_entry.
+ * If we're removing a directory perform some additional validation.
*/
- XFS_BMAP_INIT(&free_list, &first_block);
+ 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;
+ }
+ }
+
+ xfs_bmap_init(&free_list, &first_block);
error = xfs_dir_removename(tp, dp, name, ip->i_ino,
&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);
- 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;
- }
-
- /* Determine if this is the last link while
- * we are in the transaction.
- */
- link_zero = (ip)->i_d.di_nlink==0;
-
- /*
- * Take an extra ref on the inode so that it doesn't
- * go to xfs_inactive() from within the commit.
- */
- IHOLD(ip);
-
- /*
- * 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)) {
- xfs_trans_set_sync(tp);
- }
-
- error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error) {
- REMOVE_DEBUG_TRACE(__LINE__);
- goto error_rele;
- }
-
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error) {
- IRELE(ip);
- goto std_return;
- }
-
- /*
- * If we are using filestreams, kill the stream association.
- * If the file is still open it may get a new one but that
- * 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))
- xfs_filestream_deassociate(ip);
-
- xfs_itrace_exit(ip);
- IRELE(ip);
-
-/* 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);
- }
- return error;
-
- error1:
- xfs_bmap_cancel(&free_list);
- cancel_flags |= XFS_TRANS_ABORT;
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
-
- 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.
- */
- xfs_bmap_cancel(&free_list);
- cancel_flags |= XFS_TRANS_ABORT;
- xfs_trans_cancel(tp, cancel_flags);
-
- IRELE(ip);
-
- goto std_return;
-}
-
-int
-xfs_link(
- xfs_inode_t *tdp,
- xfs_inode_t *sip,
- struct xfs_name *target_name)
-{
- xfs_mount_t *mp = tdp->i_mount;
- xfs_trans_t *tp;
- xfs_inode_t *ips[2];
- int error;
- xfs_bmap_free_t free_list;
- xfs_fsblock_t first_block;
- int cancel_flags;
- int committed;
- int resblks;
-
- xfs_itrace_entry(tdp);
- xfs_itrace_entry(sip);
-
- ASSERT(!S_ISDIR(sip->i_d.di_mode));
-
- if (XFS_FORCED_SHUTDOWN(mp))
- return XFS_ERROR(EIO);
-
- if (DM_EVENT_ENABLED(tdp, DM_EVENT_LINK)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_LINK,
- tdp, DM_RIGHT_NULL,
- sip, DM_RIGHT_NULL,
- target_name->name, NULL, 0, 0, 0);
+ if (is_dir) {
+ /*
+ * Drop the link from ip's "..".
+ */
+ error = xfs_droplink(tp, dp);
if (error)
- return error;
- }
-
- /* Return through std_return after this point. */
+ goto out_bmap_cancel;
- error = XFS_QM_DQATTACH(mp, sip, 0);
- if (!error && sip != tdp)
- error = XFS_QM_DQATTACH(mp, tdp, 0);
- if (error)
- goto std_return;
-
- tp = xfs_trans_alloc(mp, XFS_TRANS_LINK);
- cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
- resblks = XFS_LINK_SPACE_RES(mp, target_name->len);
- error = xfs_trans_reserve(tp, resblks, XFS_LINK_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_LINK_LOG_COUNT);
- if (error == ENOSPC) {
- resblks = 0;
- error = xfs_trans_reserve(tp, 0, XFS_LINK_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_LINK_LOG_COUNT);
- }
- if (error) {
- cancel_flags = 0;
- goto error_return;
- }
-
- if (sip->i_ino < tdp->i_ino) {
- ips[0] = sip;
- ips[1] = tdp;
- } else {
- ips[0] = tdp;
- ips[1] = sip;
- }
-
- xfs_lock_inodes(ips, 2, 0, XFS_ILOCK_EXCL);
-
- /*
- * Increment vnode ref counts since xfs_trans_commit &
- * xfs_trans_cancel will both unlock the inodes and
- * decrement the associated ref counts.
- */
- IHOLD(sip);
- IHOLD(tdp);
- xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);
-
- /*
- * If the source has too many links, we can't make any more to it.
- */
- if (sip->i_d.di_nlink >= XFS_MAXLINK) {
- error = XFS_ERROR(EMLINK);
- 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;
- }
-
- error = xfs_dir_canenter(tp, tdp, target_name, resblks);
- if (error)
- goto error_return;
-
- XFS_BMAP_INIT(&free_list, &first_block);
-
- error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
- &first_block, &free_list, resblks);
- if (error)
- goto abort_return;
- xfs_ichgtime(tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
- tdp->i_gen++;
- xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);
-
- error = xfs_bumplink(tp, sip);
- if (error)
- goto abort_return;
-
- /*
- * If this is a synchronous mount, make sure that the
- * link 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);
- goto abort_return;
- }
-
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error)
- goto std_return;
-
- /* Fall through to std_return with error = 0. */
-std_return:
- if (DM_EVENT_ENABLED(sip, DM_EVENT_POSTLINK)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTLINK,
- tdp, DM_RIGHT_NULL,
- sip, DM_RIGHT_NULL,
- target_name->name, NULL, 0, error, 0);
- }
- return error;
-
- abort_return:
- cancel_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
-
- error_return:
- xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
-}
-
-
-int
-xfs_mkdir(
- xfs_inode_t *dp,
- struct xfs_name *dir_name,
- mode_t mode,
- xfs_inode_t **ipp,
- cred_t *credp)
-{
- xfs_mount_t *mp = dp->i_mount;
- xfs_inode_t *cdp; /* inode of created dir */
- xfs_trans_t *tp;
- int cancel_flags;
- int error;
- int committed;
- xfs_bmap_free_t free_list;
- xfs_fsblock_t first_block;
- boolean_t unlock_dp_on_error = B_FALSE;
- boolean_t created = B_FALSE;
- int dm_event_sent = 0;
- xfs_prid_t prid;
- struct xfs_dquot *udqp, *gdqp;
- uint resblks;
-
- if (XFS_FORCED_SHUTDOWN(mp))
- return XFS_ERROR(EIO);
-
- tp = NULL;
-
- if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE,
- dp, DM_RIGHT_NULL, NULL,
- DM_RIGHT_NULL, dir_name->name, NULL,
- mode, 0, 0);
+ /*
+ * Drop the "." link from ip to self.
+ */
+ error = xfs_droplink(tp, ip);
if (error)
- return error;
- dm_event_sent = 1;
- }
-
- /* Return through std_return after this point. */
-
- xfs_itrace_entry(dp);
-
- mp = dp->i_mount;
- udqp = gdqp = NULL;
- if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
- prid = dp->i_d.di_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), prid,
- XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, &udqp, &gdqp);
- if (error)
- goto std_return;
-
- tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR);
- cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
- resblks = XFS_MKDIR_SPACE_RES(mp, dir_name->len);
- error = xfs_trans_reserve(tp, resblks, XFS_MKDIR_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES, XFS_MKDIR_LOG_COUNT);
- if (error == ENOSPC) {
- resblks = 0;
- error = xfs_trans_reserve(tp, 0, XFS_MKDIR_LOG_RES(mp), 0,
- XFS_TRANS_PERM_LOG_RES,
- XFS_MKDIR_LOG_COUNT);
- }
- if (error) {
- cancel_flags = 0;
- goto error_return;
- }
-
- xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
- unlock_dp_on_error = B_TRUE;
-
- /*
- * Check for directory link count overflow.
- */
- if (dp->i_d.di_nlink >= XFS_MAXLINK) {
- error = XFS_ERROR(EMLINK);
- goto error_return;
- }
-
- /*
- * Reserve disk quota and the inode.
- */
- error = XFS_TRANS_RESERVE_QUOTA(mp, tp, udqp, gdqp, resblks, 1, 0);
- if (error)
- goto error_return;
-
- error = xfs_dir_canenter(tp, dp, dir_name, resblks);
- if (error)
- goto error_return;
- /*
- * create the directory inode.
- */
- error = xfs_dir_ialloc(&tp, dp, mode, 2,
- 0, credp, prid, resblks > 0,
- &cdp, NULL);
- if (error) {
- if (error == ENOSPC)
- goto error_return;
- goto abort_return;
- }
- xfs_itrace_ref(cdp);
-
- /*
- * Now we add the directory inode to the transaction.
- * We waited until now since xfs_dir_ialloc might start
- * a new transaction. Had we joined the transaction
- * earlier, the locks might have gotten released. An error
- * from here on will result in the transaction cancel
- * unlocking dp so don't do it explicitly in the error path.
- */
- IHOLD(dp);
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
- unlock_dp_on_error = B_FALSE;
-
- XFS_BMAP_INIT(&free_list, &first_block);
-
- error = xfs_dir_createname(tp, dp, dir_name, cdp->i_ino,
- &first_block, &free_list, resblks ?
- resblks - XFS_IALLOC_SPACE_RES(mp) : 0);
- if (error) {
- ASSERT(error != ENOSPC);
- goto error1;
+ goto out_bmap_cancel;
+ } else {
+ /*
+ * When removing a non-directory we need to log the parent
+ * inode here. For a directory this is done implicitly
+ * by the xfs_droplink call for the ".." entry.
+ */
+ xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
}
- xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
/*
- * Bump the in memory version number of the parent directory
- * so that other processes accessing it will recognize that
- * the directory has changed.
+ * Drop the link from dp to ip.
*/
- dp->i_gen++;
-
- error = xfs_dir_init(tp, cdp, dp);
- if (error)
- goto error2;
-
- cdp->i_gen = 1;
- error = xfs_bumplink(tp, dp);
- if (error)
- goto error2;
-
- created = B_TRUE;
-
- *ipp = cdp;
- IHOLD(cdp);
+ error = xfs_droplink(tp, ip);
+ if (error)
+ goto out_bmap_cancel;
/*
- * Attach the dquots to the new inode and modify the icount incore.
+ * Determine if this is the last link while
+ * we are in the transaction.
*/
- XFS_QM_DQVOPCREATE(mp, tp, cdp, udqp, gdqp);
+ link_zero = (ip->i_d.di_nlink == 0);
/*
* If this is a synchronous mount, make sure that the
- * mkdir transaction goes to disk before returning to
+ * 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) {
- IRELE(cdp);
- goto error2;
- }
+ if (error)
+ goto out_bmap_cancel;
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
- if (error) {
- IRELE(cdp);
- }
+ if (error)
+ goto std_return;
- /* Fall through to std_return with error = 0 or errno from
- * xfs_trans_commit. */
+ /*
+ * If we are using filestreams, kill the stream association.
+ * If the file is still open it may get a new one but that
+ * will get killed on last close in xfs_close() so we don't
+ * have to worry about that.
+ */
+ if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
+ xfs_filestream_deassociate(ip);
-std_return:
- if ((created || (error != 0 && dm_event_sent != 0)) &&
- DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE,
- dp, DM_RIGHT_NULL,
- created ? cdp : NULL,
- DM_RIGHT_NULL,
- dir_name->name, NULL,
- mode, error, 0);
+ xfs_itrace_exit(ip);
+ xfs_itrace_exit(dp);
+
+ std_return:
+ if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) {
+ 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;
- error2:
- error1:
+ out_bmap_cancel:
xfs_bmap_cancel(&free_list);
- abort_return:
cancel_flags |= XFS_TRANS_ABORT;
- error_return:
+ out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
-
- if (unlock_dp_on_error)
- xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
goto std_return;
}
int
-xfs_rmdir(
- xfs_inode_t *dp,
- struct xfs_name *name,
- xfs_inode_t *cdp)
+xfs_link(
+ xfs_inode_t *tdp,
+ xfs_inode_t *sip,
+ struct xfs_name *target_name)
{
- bhv_vnode_t *dir_vp = XFS_ITOV(dp);
- xfs_mount_t *mp = dp->i_mount;
- xfs_trans_t *tp;
- int error;
+ xfs_mount_t *mp = tdp->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;
+ int resblks;
- xfs_itrace_entry(dp);
+ xfs_itrace_entry(tdp);
+ xfs_itrace_entry(sip);
+
+ ASSERT(!S_ISDIR(sip->i_d.di_mode));
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 (DM_EVENT_ENABLED(tdp, DM_EVENT_LINK)) {
+ error = XFS_SEND_NAMESP(mp, DM_EVENT_LINK,
+ tdp, DM_RIGHT_NULL,
+ sip, DM_RIGHT_NULL,
+ target_name->name, NULL, 0, 0, 0);
if (error)
- return XFS_ERROR(error);
+ return error;
}
- /*
- * We need to get a reference to cdp before we get our log
- * reservation. The reason for this is that we cannot call
- * xfs_iget for an inode for which we do not have a reference
- * once we've acquired a log reservation. This is because the
- * inode we are trying to get might be in xfs_inactive going
- * 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 reference
- * to the inode before getting our log reservation.
- */
- IHOLD(cdp);
+ /* Return through std_return after this point. */
- /*
- * Get the dquots for the inodes.
- */
- error = XFS_QM_DQATTACH(mp, dp, 0);
- if (!error && dp != cdp)
- error = XFS_QM_DQATTACH(mp, cdp, 0);
- if (error) {
- IRELE(cdp);
- REMOVE_DEBUG_TRACE(__LINE__);
+ error = xfs_qm_dqattach(sip, 0);
+ if (error)
+ goto std_return;
+
+ error = xfs_qm_dqattach(tdp, 0);
+ if (error)
goto std_return;
- }
- tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR);
+ tp = xfs_trans_alloc(mp, XFS_TRANS_LINK);
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);
+ resblks = XFS_LINK_SPACE_RES(mp, target_name->len);
+ error = xfs_trans_reserve(tp, resblks, XFS_LINK_LOG_RES(mp), 0,
+ XFS_TRANS_PERM_LOG_RES, XFS_LINK_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);
+ error = xfs_trans_reserve(tp, 0, XFS_LINK_LOG_RES(mp), 0,
+ XFS_TRANS_PERM_LOG_RES, XFS_LINK_LOG_COUNT);
}
if (error) {
- ASSERT(error != ENOSPC);
cancel_flags = 0;
- IRELE(cdp);
- 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);
- IRELE(cdp);
- goto std_return;
- }
-
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
- if (dp != cdp) {
- /*
- * Only increment the parent directory vnode count if
- * we didn't bump it in looking up cdp. The only time
- * we don't bump it is when we're looking up ".".
- */
- VN_HOLD(dir_vp);
- }
-
- xfs_itrace_ref(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);
+ xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL);
/*
- * Bump the in memory generation count on the parent
- * directory so that other can know that it has changed.
+ * Increment vnode ref counts since xfs_trans_commit &
+ * xfs_trans_cancel will both unlock the inodes and
+ * decrement the associated ref counts.
*/
- dp->i_gen++;
+ IHOLD(sip);
+ IHOLD(tdp);
+ xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);
/*
- * Drop the link from cdp's "..".
+ * If the source has too many links, we can't make any more to it.
*/
- error = xfs_droplink(tp, dp);
- if (error) {
- goto error1;
+ if (sip->i_d.di_nlink >= XFS_MAXLINK) {
+ error = XFS_ERROR(EMLINK);
+ goto error_return;
}
/*
- * Drop the link from dp to cdp.
+ * 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.
*/
- error = xfs_droplink(tp, cdp);
- if (error) {
- goto error1;
+ 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;
}
- /*
- * Drop the "." link from cdp to self.
- */
- error = xfs_droplink(tp, cdp);
- if (error) {
- goto error1;
- }
+ error = xfs_dir_canenter(tp, tdp, target_name, resblks);
+ if (error)
+ goto error_return;
- /* Determine these before committing transaction */
- last_cdp_link = (cdp)->i_d.di_nlink==0;
+ xfs_bmap_init(&free_list, &first_block);
- /*
- * Take an extra ref on the child vnode so that it
- * does not go to xfs_inactive() from within the commit.
- */
- IHOLD(cdp);
+ error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
+ &first_block, &free_list, resblks);
+ if (error)
+ goto abort_return;
+ xfs_ichgtime(tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
+ xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);
+
+ error = xfs_bumplink(tp, sip);
+ if (error)
+ goto abort_return;
/*
* If this is a synchronous mount, make sure that the
- * rmdir transaction goes to disk before returning to
+ * link transaction goes to disk before returning to
* the user.
*/
if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
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));
- IRELE(cdp);
- goto std_return;
+ goto abort_return;
}
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error) {
- IRELE(cdp);
+ if (error)
goto std_return;
- }
-
- IRELE(cdp);
-
- /* 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);
+ /* Fall through to std_return with error = 0. */
+std_return:
+ if (DM_EVENT_ENABLED(sip, DM_EVENT_POSTLINK)) {
+ (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTLINK,
+ tdp, DM_RIGHT_NULL,
+ sip, DM_RIGHT_NULL,
+ target_name->name, NULL, 0, error, 0);
}
return error;
- error1:
- xfs_bmap_cancel(&free_list);
+ abort_return:
cancel_flags |= XFS_TRANS_ABORT;
/* FALLTHROUGH */
/*
* Make sure that we have allocated dquot(s) on disk.
*/
- error = XFS_QM_DQVOPALLOC(mp, dp,
- current_fsuid(credp), current_fsgid(credp), prid,
+ error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT, &udqp, &gdqp);
if (error)
goto std_return;
/*
* Reserve disk quota : blocks and inode.
*/
- error = XFS_TRANS_RESERVE_QUOTA(mp, tp, udqp, gdqp, resblks, 1, 0);
+ error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp, resblks, 1, 0);
if (error)
goto error_return;
* Initialize the bmap freelist prior to calling either
* bmapi or the directory create code.
*/
- XFS_BMAP_INIT(&free_list, &first_block);
+ xfs_bmap_init(&free_list, &first_block);
/*
* Allocate an inode for the symlink.
/*
* Also attach the dquot(s) to it, if applicable.
*/
- XFS_QM_DQVOPCREATE(mp, tp, ip, udqp, gdqp);
+ xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp);
if (resblks)
resblks -= XFS_IALLOC_SPACE_RES(mp);
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
/*
- * Bump the in memory version number of the parent directory
- * so that other processes accessing it will recognize that
- * the directory has changed.
- */
- dp->i_gen++;
-
- /*
* If this is a synchronous mount, make sure that the
* symlink transaction goes to disk before returning to
* the user.
goto error2;
}
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
/* Fall through to std_return with error = 0 or errno from
* xfs_trans_commit */
cancel_flags |= XFS_TRANS_ABORT;
error_return:
xfs_trans_cancel(tp, cancel_flags);
- XFS_QM_DQRELE(mp, udqp);
- XFS_QM_DQRELE(mp, gdqp);
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
if (unlock_dp_on_error)
xfs_iunlock(dp, XFS_ILOCK_EXCL);
}
int
-xfs_inode_flush(
- xfs_inode_t *ip,
- int flags)
-{
- xfs_mount_t *mp = ip->i_mount;
- int error = 0;
-
- if (XFS_FORCED_SHUTDOWN(mp))
- return XFS_ERROR(EIO);
-
- /*
- * Bypass inodes which have already been cleaned by
- * the inode flush clustering code inside xfs_iflush
- */
- if (xfs_inode_clean(ip))
- return 0;
-
- /*
- * We make this non-blocking if the inode is contended,
- * return EAGAIN to indicate to the caller that they
- * did not succeed. This prevents the flush path from
- * blocking on inodes inside another operation right
- * now, they get caught later by xfs_sync.
- */
- if (flags & FLUSH_SYNC) {
- xfs_ilock(ip, XFS_ILOCK_SHARED);
- xfs_iflock(ip);
- } else if (xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
- if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) {
- xfs_iunlock(ip, XFS_ILOCK_SHARED);
- return EAGAIN;
- }
- } else {
- return EAGAIN;
- }
-
- error = xfs_iflush(ip, (flags & FLUSH_SYNC) ? XFS_IFLUSH_SYNC
- : XFS_IFLUSH_ASYNC_NOBLOCK);
- xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
- return error;
-}
-
-
-int
xfs_set_dmattrs(
xfs_inode_t *ip,
u_int evmask,
xfs_reclaim(
xfs_inode_t *ip)
{
- bhv_vnode_t *vp = XFS_ITOV(ip);
xfs_itrace_entry(ip);
- ASSERT(!VN_MAPPED(vp));
+ ASSERT(!VN_MAPPED(VFS_I(ip)));
/* bad inode, get out here ASAP */
- if (VN_BAD(vp)) {
+ if (is_bad_inode(VFS_I(ip))) {
xfs_ireclaim(ip);
return 0;
}
- vn_iowait(ip);
+ xfs_ioend_wait(ip);
ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
if (!ip->i_update_core && (ip->i_itemp == NULL)) {
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_iflock(ip);
- return xfs_finish_reclaim(ip, 1, XFS_IFLUSH_DELWRI_ELSE_SYNC);
- } else {
- xfs_mount_t *mp = ip->i_mount;
-
- /* Protect sync and unpin from us */
- XFS_MOUNT_ILOCK(mp);
- spin_lock(&ip->i_flags_lock);
- __xfs_iflags_set(ip, XFS_IRECLAIMABLE);
- vn_to_inode(vp)->i_private = NULL;
- ip->i_vnode = NULL;
- spin_unlock(&ip->i_flags_lock);
- list_add_tail(&ip->i_reclaim, &mp->m_del_inodes);
- XFS_MOUNT_IUNLOCK(mp);
- }
- return 0;
-}
-
-int
-xfs_finish_reclaim(
- xfs_inode_t *ip,
- int locked,
- int sync_mode)
-{
- 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;
-
- /* The hash lock here protects a thread in xfs_iget_core from
- * racing with us on linking the inode back with a vnode.
- * Once we have the XFS_IRECLAIM flag set it will not touch
- * us.
- */
- write_lock(&pag->pag_ici_lock);
- spin_lock(&ip->i_flags_lock);
- if (__xfs_iflags_test(ip, XFS_IRECLAIM) ||
- (!__xfs_iflags_test(ip, XFS_IRECLAIMABLE) && vp == NULL)) {
- spin_unlock(&ip->i_flags_lock);
- write_unlock(&pag->pag_ici_lock);
- if (locked) {
- xfs_ifunlock(ip);
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- }
- return 1;
- }
- __xfs_iflags_set(ip, XFS_IRECLAIM);
- spin_unlock(&ip->i_flags_lock);
- write_unlock(&pag->pag_ici_lock);
- xfs_put_perag(ip->i_mount, pag);
-
- /*
- * If the inode is still dirty, then flush it out. If the inode
- * is not in the AIL, then it will be OK to flush it delwri as
- * long as xfs_iflush() does not keep any references to the inode.
- * We leave that decision up to xfs_iflush() since it has the
- * knowledge of whether it's OK to simply do a delwri flush of
- * the inode or whether we need to wait until the inode is
- * pulled from the AIL.
- * We get the flush lock regardless, though, just to make sure
- * we don't free it while it is being flushed.
- */
- if (!locked) {
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- 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);
- }
-
- xfs_ifunlock(ip);
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
- reclaim:
- xfs_ireclaim(ip);
- return 0;
-}
-
-int
-xfs_finish_reclaim_all(xfs_mount_t *mp, int noblock)
-{
- int purged;
- xfs_inode_t *ip, *n;
- int done = 0;
-
- while (!done) {
- purged = 0;
- XFS_MOUNT_ILOCK(mp);
- list_for_each_entry_safe(ip, n, &mp->m_del_inodes, i_reclaim) {
- if (noblock) {
- if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0)
- continue;
- if (xfs_ipincount(ip) ||
- !xfs_iflock_nowait(ip)) {
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- continue;
- }
- }
- XFS_MOUNT_IUNLOCK(mp);
- if (xfs_finish_reclaim(ip, noblock,
- XFS_IFLUSH_DELWRI_ELSE_ASYNC))
- delay(1);
- purged = 1;
- break;
- }
-
- done = !purged;
+ xfs_iflags_set(ip, XFS_IRECLAIMABLE);
+ return xfs_reclaim_inode(ip, 1, XFS_IFLUSH_DELWRI_ELSE_SYNC);
}
-
- XFS_MOUNT_IUNLOCK(mp);
+ xfs_inode_set_reclaim_tag(ip);
return 0;
}
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
return error;
if (len <= 0)
/* 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;
break;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, tp, ip,
- qblocks, 0, quota_flag);
+ error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks,
+ 0, quota_flag);
if (error)
goto error1;
/*
* Issue the xfs_bmapi() call to allocate the blocks
*/
- XFS_BMAP_INIT(&free_list, &firstfsb);
+ xfs_bmap_init(&free_list, &firstfsb);
error = xfs_bmapi(tp, ip, startoffset_fsb,
allocatesize_fsb, bmapi_flag,
&firstfsb, 0, imapp, &nimaps,
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,
error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */
xfs_bmap_cancel(&free_list);
- XFS_TRANS_UNRESERVE_QUOTA_NBLKS(mp, tp, ip, qblocks, 0, quota_flag);
+ xfs_trans_unreserve_quota_nblks(tp, ip, qblocks, 0, quota_flag);
error1: /* Just cancel transaction */
xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
/*
* Zero file bytes between startoff and endoff inclusive.
* The iolock is held exclusive and no blocks are buffered.
+ *
+ * This function is used by xfs_free_file_space() to zero
+ * partial blocks when the range to free is not block aligned.
+ * When unreserving space with boundaries that are not block
+ * aligned we round up the start and round down the end
+ * boundaries and then use this function to zero the parts of
+ * the blocks that got dropped during the rounding.
*/
STATIC int
xfs_zero_remaining_bytes(
int nimap;
int error = 0;
+ /*
+ * Avoid doing I/O beyond eof - it's not necessary
+ * since nothing can read beyond eof. The space will
+ * be zeroed when the file is extended anyway.
+ */
+ if (startoff >= ip->i_size)
+ return 0;
+
+ if (endoff > ip->i_size)
+ endoff = ip->i_size;
+
bp = xfs_buf_get_noaddr(mp->m_sb.sb_blocksize,
XFS_IS_REALTIME_INODE(ip) ?
mp->m_rtdev_targp : mp->m_ddev_targp);
+ if (!bp)
+ return XFS_ERROR(ENOMEM);
for (offset = startoff; offset <= endoff; offset = lastoffset + 1) {
offset_fsb = XFS_B_TO_FSBT(mp, offset);
XFS_BUF_UNDONE(bp);
XFS_BUF_UNWRITE(bp);
XFS_BUF_READ(bp);
- XFS_BUF_SET_ADDR(bp, XFS_FSB_TO_DB(ip, imap.br_startblock));
+ XFS_BUF_SET_ADDR(bp, xfs_fsb_to_db(ip, imap.br_startblock));
xfsbdstrat(mp, bp);
error = xfs_iowait(bp);
if (error) {
xfs_off_t len,
int attr_flags)
{
- bhv_vnode_t *vp;
int committed;
int done;
xfs_off_t end_dmi_offset;
xfs_trans_t *tp;
int need_iolock = 1;
- vp = XFS_ITOV(ip);
mp = ip->i_mount;
xfs_itrace_entry(ip);
- if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
return error;
error = 0;
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);
- vn_iowait(ip); /* wait for the completion of any pending DIOs */
+ /* wait for the completion of any pending DIOs */
+ xfs_ioend_wait(ip);
}
rounding = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE);
ioffset = offset & ~(rounding - 1);
- if (VN_CACHED(vp) != 0) {
+ if (VN_CACHED(VFS_I(ip)) != 0) {
xfs_inval_cached_trace(ip, ioffset, -1, ioffset, -1);
error = xfs_flushinval_pages(ip, ioffset, -1, FI_REMAPF_LOCKED);
if (error)
/*
* Need to zero the stuff we're not freeing, on disk.
- * If its a realtime file & can't use unwritten extents then we
+ * If it's a realtime file & can't use unwritten extents then we
* actually need to zero the extent edges. Otherwise xfs_bunmapi
* will take care of it for us.
*/
break;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- error = XFS_TRANS_RESERVE_QUOTA(mp, tp,
- ip->i_udquot, ip->i_gdquot, resblks, 0,
- XFS_QMOPT_RES_REGBLKS);
+ error = xfs_trans_reserve_quota(tp, mp,
+ ip->i_udquot, ip->i_gdquot,
+ resblks, 0, XFS_QMOPT_RES_REGBLKS);
if (error)
goto error1;
/*
* issue the bunmapi() call to free the blocks
*/
- XFS_BMAP_INIT(&free_list, &firstfsb);
+ xfs_bmap_init(&free_list, &firstfsb);
error = xfs_bunmapi(tp, ip, startoffset_fsb,
endoffset_fsb - startoffset_fsb,
0, 2, &firstfsb, &free_list, NULL, &done);
int cmd,
xfs_flock64_t *bf,
xfs_off_t offset,
- cred_t *credp,
int attr_flags)
{
xfs_mount_t *mp = ip->i_mount;
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);
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;
/*
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff