/*
* Common pre-write limit and setup checks.
*
* Called with the iolocked held either shared and exclusive according to
* @iolock, and returns with it held. Might upgrade the iolock to exclusive
* if called for a direct write beyond i_size.
*/
STATIC ssize_t
xfs_file_aio_write_checks(
struct file *file,
loff_t *pos,
size_t *count,
int *iolock)
{
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
int error = 0;
restart:
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error)
return error;
/*
* If the offset is beyond the size of the file, we need to zero any
* blocks that fall between the existing EOF and the start of this
* write. If zeroing is needed and we are currently holding the
* iolock shared, we need to update it to exclusive which implies
* having to redo all checks before.
*/
if (*pos > i_size_read(inode)) {
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_rw_iunlock(ip, *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_rw_ilock(ip, *iolock);
goto restart;
}
error = -xfs_zero_eof(ip, *pos, i_size_read(inode));
if (error)
return error;
}
/*
* Updating the timestamps will grab the ilock again from
* xfs_fs_dirty_inode, so we have to call it after dropping the
* lock above. Eventually we should look into a way to avoid
* the pointless lock roundtrip.
*/
if (likely(!(file->f_mode & FMODE_NOCMTIME))) {
error = file_update_time(file);
if (error)
return error;
}
/*
* If we're writing the file then make sure to clear the setuid and
* setgid bits if the process is not being run by root. This keeps
* people from modifying setuid and setgid binaries.
*/
return file_remove_suid(file);
}
/*
* Common pre-write limit and setup checks.
*
* Called with the iolocked held either shared and exclusive according to
* @iolock, and returns with it held. Might upgrade the iolock to exclusive
* if called for a direct write beyond i_size.
*/
STATIC ssize_t
xfs_file_aio_write_checks(
struct file *file,
loff_t *pos,
size_t *count,
int *iolock)
{
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
int error = 0;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
restart:
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
return error;
}
if (likely(!(file->f_mode & FMODE_NOCMTIME)))
file_update_time(file);
/*
* If the offset is beyond the size of the file, we need to zero any
* blocks that fall between the existing EOF and the start of this
* write. If zeroing is needed and we are currently holding the
* iolock shared, we need to update it to exclusive which involves
* dropping all locks and relocking to maintain correct locking order.
* If we do this, restart the function to ensure all checks and values
* are still valid.
*/
if (*pos > i_size_read(inode)) {
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
goto restart;
}
error = -xfs_zero_eof(ip, *pos, i_size_read(inode));
}
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
if (error)
return error;
/*
* If we're writing the file then make sure to clear the setuid and
* setgid bits if the process is not being run by root. This keeps
* people from modifying setuid and setgid binaries.
*/
return file_remove_suid(file);
}
/*
* Common pre-write limit and setup checks.
*
* Returns with iolock held according to @iolock.
*/
STATIC ssize_t
xfs_file_aio_write_checks(
struct file *file,
loff_t *pos,
size_t *count,
int *iolock)
{
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
xfs_fsize_t new_size;
int error = 0;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
*iolock = 0;
return error;
}
new_size = *pos + *count;
if (new_size > ip->i_size)
ip->i_new_size = new_size;
if (likely(!(file->f_mode & FMODE_NOCMTIME)))
file_update_time(file);
/*
* If the offset is beyond the size of the file, we need to zero any
* blocks that fall between the existing EOF and the start of this
* write.
*/
if (*pos > ip->i_size)
error = -xfs_zero_eof(ip, *pos, ip->i_size);
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
if (error)
return error;
/*
* If we're writing the file then make sure to clear the setuid and
* setgid bits if the process is not being run by root. This keeps
* people from modifying setuid and setgid binaries.
*/
return file_remove_suid(file);
}
/*
* xfs_inval_cached_pages
*
* This routine is responsible for keeping direct I/O and buffered I/O
* somewhat coherent. From here we make sure that we're at least
* temporarily holding the inode I/O lock exclusively and then call
* the page cache to flush and invalidate any cached pages. If there
* are no cached pages this routine will be very quick.
*/
void
xfs_inval_cached_pages(
vnode_t *vp,
xfs_iocore_t *io,
xfs_off_t offset,
int write,
int relock)
{
xfs_mount_t *mp;
if (!VN_CACHED(vp)) {
return;
}
mp = io->io_mount;
/*
* We need to get the I/O lock exclusively in order
* to safely invalidate pages and mappings.
*/
if (relock) {
XFS_IUNLOCK(mp, io, XFS_IOLOCK_SHARED);
XFS_ILOCK(mp, io, XFS_IOLOCK_EXCL);
}
/* Writing beyond EOF creates a hole that must be zeroed */
if (write && (offset > XFS_SIZE(mp, io))) {
xfs_fsize_t isize;
XFS_ILOCK(mp, io, XFS_ILOCK_EXCL|XFS_EXTSIZE_RD);
isize = XFS_SIZE(mp, io);
if (offset > isize) {
xfs_zero_eof(vp, io, offset, isize, offset);
}
XFS_IUNLOCK(mp, io, XFS_ILOCK_EXCL|XFS_EXTSIZE_RD);
}
xfs_inval_cached_trace(io, offset, -1, ctooff(offtoct(offset)), -1);
VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(offset)), -1, FI_REMAPF_LOCKED);
if (relock) {
XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
}
}
STATIC ssize_t
xfs_file_aio_write_checks(
struct file *file,
loff_t *pos,
size_t *count,
int *iolock)
{
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
int error = 0;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
restart:
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
return error;
}
if (*pos > i_size_read(inode)) {
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
goto restart;
}
error = -xfs_zero_eof(ip, *pos, i_size_read(inode));
}
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
if (error)
return error;
if (likely(!(file->f_mode & FMODE_NOCMTIME)))
file_update_time(file);
return file_remove_suid(file);
}
ssize_t /* bytes written, or (-) error */
xfs_write(
bhv_desc_t *bdp,
struct file *file,
const char *buf,
size_t size,
loff_t *offset,
int ioflags,
cred_t *credp)
{
xfs_inode_t *xip;
xfs_mount_t *mp;
ssize_t ret;
int error = 0;
xfs_fsize_t isize, new_size;
xfs_fsize_t n, limit;
xfs_iocore_t *io;
vnode_t *vp;
int iolock;
int eventsent = 0;
vrwlock_t locktype;
XFS_STATS_INC(xs_write_calls);
vp = BHV_TO_VNODE(bdp);
xip = XFS_BHVTOI(bdp);
if (size == 0)
return 0;
io = &xip->i_iocore;
mp = io->io_mount;
fs_check_frozen(vp->v_vfsp, SB_FREEZE_WRITE);
if (XFS_FORCED_SHUTDOWN(xip->i_mount)) {
return -EIO;
}
if (unlikely(ioflags & IO_ISDIRECT)) {
if (((__psint_t)buf & BBMASK) ||
(*offset & mp->m_blockmask) ||
(size & mp->m_blockmask)) {
return XFS_ERROR(-EINVAL);
}
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
} else {
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
}
if (ioflags & IO_ISLOCKED)
iolock = 0;
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
isize = xip->i_d.di_size;
limit = XFS_MAXIOFFSET(mp);
if (file->f_flags & O_APPEND)
*offset = isize;
start:
n = limit - *offset;
if (n <= 0) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return -EFBIG;
}
if (n < size)
size = n;
new_size = *offset + size;
if (new_size > isize) {
io->io_new_size = new_size;
}
if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
loff_t savedsize = *offset;
int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
*offset, size,
dmflags, &locktype);
if (error) {
if (iolock) xfs_iunlock(xip, iolock);
return -error;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reaquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) &&
savedsize != xip->i_d.di_size) {
*offset = isize = xip->i_d.di_size;
goto start;
}
}
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (!(ioflags & IO_ISDIRECT) && (*offset > isize && isize)) {
error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, *offset,
isize, *offset + size);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return(-error);
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (error) {
xfs_iunlock(xip, iolock);
return -error;
}
}
if ((ssize_t) size < 0) {
ret = -EINVAL;
goto error;
}
if (!access_ok(VERIFY_READ, buf, size)) {
ret = -EINVAL;
goto error;
}
retry:
if (unlikely(ioflags & IO_ISDIRECT)) {
xfs_inval_cached_pages(vp, io, *offset, 1, 1);
xfs_rw_enter_trace(XFS_DIOWR_ENTER,
io, buf, size, *offset, ioflags);
ret = do_generic_direct_write(file, buf, size, offset);
} else {
xfs_rw_enter_trace(XFS_WRITE_ENTER,
io, buf, size, *offset, ioflags);
ret = do_generic_file_write(file, buf, size, offset);
}
if (unlikely(ioflags & IO_INVIS)) {
/* generic_file_write updates the mtime/ctime but we need
* to undo that because this I/O was supposed to be
* invisible.
*/
struct inode *inode = LINVFS_GET_IP(vp);
inode->i_mtime = xip->i_d.di_mtime.t_sec;
inode->i_ctime = xip->i_d.di_ctime.t_sec;
} else {
xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
}
if ((ret == -ENOSPC) &&
DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
!(ioflags & IO_INVIS)) {
xfs_rwunlock(bdp, locktype);
error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (error)
return -error;
xfs_rwlock(bdp, locktype);
*offset = xip->i_d.di_size;
goto retry;
}
error:
if (ret <= 0) {
if (iolock)
xfs_rwunlock(bdp, locktype);
return ret;
}
XFS_STATS_ADD(xs_write_bytes, ret);
if (*offset > xip->i_d.di_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_d.di_size) {
struct inode *inode = LINVFS_GET_IP(vp);
xip->i_d.di_size = *offset;
i_size_write(inode, *offset);
xip->i_update_core = 1;
xip->i_update_size = 1;
mark_inode_dirty_sync(inode);
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(file->f_dentry->d_inode)) {
/*
* If we're treating this as O_DSYNC and we have not updated the
* size, force the log.
*/
if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC)
&& !(xip->i_update_size)) {
/*
* If an allocation transaction occurred
* without extending the size, then we have to force
* the log up the proper point to ensure that the
* allocation is permanent. We can't count on
* the fact that buffered writes lock out direct I/O
* writes - the direct I/O write could have extended
* the size nontransactionally, then finished before
* we started. xfs_write_file will think that the file
* didn't grow but the update isn't safe unless the
* size change is logged.
*
* Force the log if we've committed a transaction
* against the inode or if someone else has and
* the commit record hasn't gone to disk (e.g.
* the inode is pinned). This guarantees that
* all changes affecting the inode are permanent
* when we return.
*/
xfs_inode_log_item_t *iip;
xfs_lsn_t lsn;
iip = xip->i_itemp;
if (iip && iip->ili_last_lsn) {
lsn = iip->ili_last_lsn;
xfs_log_force(mp, lsn,
XFS_LOG_FORCE | XFS_LOG_SYNC);
} else if (xfs_ipincount(xip) > 0) {
xfs_log_force(mp, (xfs_lsn_t)0,
XFS_LOG_FORCE | XFS_LOG_SYNC);
}
} else {
xfs_trans_t *tp;
/*
* O_SYNC or O_DSYNC _with_ a size update are handled
* the same way.
*
* If the write was synchronous then we need to make
* sure that the inode modification time is permanent.
* We'll have updated the timestamp above, so here
* we use a synchronous transaction to log the inode.
* It's not fast, but it's necessary.
*
* If this a dsync write and the size got changed
* non-transactionally, then we need to ensure that
* the size change gets logged in a synchronous
* transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
if ((error = xfs_trans_reserve(tp, 0,
XFS_SWRITE_LOG_RES(mp),
0, 0, 0))) {
/* Transaction reserve failed */
xfs_trans_cancel(tp, 0);
} else {
/* Transaction reserve successful */
xfs_ilock(xip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, xip);
xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0, NULL);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
}
} /* (ioflags & O_SYNC) */
/*
* If we are coming from an nfsd thread then insert into the
* reference cache.
*/
if (!strcmp(current->comm, "nfsd"))
xfs_refcache_insert(xip);
/* Drop lock this way - the old refcache release is in here */
if (iolock)
xfs_rwunlock(bdp, locktype);
return(ret);
}
STATIC ssize_t
xfs_file_aio_write(
struct kiocb *iocb,
const struct iovec *iovp,
unsigned long nr_segs,
loff_t pos)
{
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
struct xfs_inode *ip = XFS_I(inode);
struct xfs_mount *mp = ip->i_mount;
ssize_t ret = 0, error = 0;
int ioflags = 0;
xfs_fsize_t isize, new_size;
int iolock;
int eventsent = 0;
size_t ocount = 0, count;
int need_i_mutex;
XFS_STATS_INC(xs_write_calls);
BUG_ON(iocb->ki_pos != pos);
if (unlikely(file->f_flags & O_DIRECT))
ioflags |= IO_ISDIRECT;
if (file->f_mode & FMODE_NOCMTIME)
ioflags |= IO_INVIS;
error = generic_segment_checks(iovp, &nr_segs, &ocount, VERIFY_READ);
if (error)
return error;
count = ocount;
if (count == 0)
return 0;
xfs_wait_for_freeze(mp, SB_FREEZE_WRITE);
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
relock:
if (ioflags & IO_ISDIRECT) {
iolock = XFS_IOLOCK_SHARED;
need_i_mutex = 0;
} else {
iolock = XFS_IOLOCK_EXCL;
need_i_mutex = 1;
mutex_lock(&inode->i_mutex);
}
xfs_ilock(ip, XFS_ILOCK_EXCL|iolock);
start:
error = -generic_write_checks(file, &pos, &count,
S_ISBLK(inode->i_mode));
if (error) {
xfs_iunlock(ip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_mutex;
}
if ((DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
int dmflags = FILP_DELAY_FLAG(file);
if (need_i_mutex)
dmflags |= DM_FLAGS_IMUX;
xfs_iunlock(ip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(ip->i_mount, DM_EVENT_WRITE, ip,
pos, count, dmflags, &iolock);
if (error) {
goto out_unlock_internal;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reacquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) && pos != ip->i_size)
goto start;
}
if (ioflags & IO_ISDIRECT) {
xfs_buftarg_t *target =
XFS_IS_REALTIME_INODE(ip) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((pos & target->bt_smask) || (count & target->bt_smask)) {
xfs_iunlock(ip, XFS_ILOCK_EXCL|iolock);
return XFS_ERROR(-EINVAL);
}
if (!need_i_mutex && (mapping->nrpages || pos > ip->i_size)) {
xfs_iunlock(ip, XFS_ILOCK_EXCL|iolock);
iolock = XFS_IOLOCK_EXCL;
need_i_mutex = 1;
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, XFS_ILOCK_EXCL|iolock);
goto start;
}
}
new_size = pos + count;
if (new_size > ip->i_size)
ip->i_new_size = new_size;
if (likely(!(ioflags & IO_INVIS)))
file_update_time(file);
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (pos > ip->i_size) {
error = xfs_zero_eof(ip, pos, ip->i_size);
if (error) {
xfs_iunlock(ip, XFS_ILOCK_EXCL);
goto out_unlock_internal;
}
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
error = -file_remove_suid(file);
if (unlikely(error))
goto out_unlock_internal;
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
if ((ioflags & IO_ISDIRECT)) {
if (mapping->nrpages) {
WARN_ON(need_i_mutex == 0);
error = xfs_flushinval_pages(ip,
(pos & PAGE_CACHE_MASK),
-1, FI_REMAPF_LOCKED);
if (error)
goto out_unlock_internal;
}
if (need_i_mutex) {
/* demote the lock now the cached pages are gone */
xfs_ilock_demote(ip, XFS_IOLOCK_EXCL);
mutex_unlock(&inode->i_mutex);
iolock = XFS_IOLOCK_SHARED;
need_i_mutex = 0;
}
trace_xfs_file_direct_write(ip, count, iocb->ki_pos, ioflags);
ret = generic_file_direct_write(iocb, iovp,
&nr_segs, pos, &iocb->ki_pos, count, ocount);
/*
* direct-io write to a hole: fall through to buffered I/O
* for completing the rest of the request.
*/
if (ret >= 0 && ret != count) {
XFS_STATS_ADD(xs_write_bytes, ret);
pos += ret;
count -= ret;
ioflags &= ~IO_ISDIRECT;
xfs_iunlock(ip, iolock);
goto relock;
}
} else {
int enospc = 0;
ssize_t ret2 = 0;
write_retry:
trace_xfs_file_buffered_write(ip, count, iocb->ki_pos, ioflags);
ret2 = generic_file_buffered_write(iocb, iovp, nr_segs,
pos, &iocb->ki_pos, count, ret);
/*
* if we just got an ENOSPC, flush the inode now we
* aren't holding any page locks and retry *once*
*/
if (ret2 == -ENOSPC && !enospc) {
error = xfs_flush_pages(ip, 0, -1, 0, FI_NONE);
if (error)
goto out_unlock_internal;
enospc = 1;
goto write_retry;
}
ret = ret2;
}
current->backing_dev_info = NULL;
isize = i_size_read(inode);
if (unlikely(ret < 0 && ret != -EFAULT && iocb->ki_pos > isize))
iocb->ki_pos = isize;
if (iocb->ki_pos > ip->i_size) {
xfs_ilock(ip, XFS_ILOCK_EXCL);
if (iocb->ki_pos > ip->i_size)
ip->i_size = iocb->ki_pos;
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
if (ret == -ENOSPC &&
DM_EVENT_ENABLED(ip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
xfs_iunlock(ip, iolock);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error = XFS_SEND_NAMESP(ip->i_mount, DM_EVENT_NOSPACE, ip,
DM_RIGHT_NULL, ip, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, iolock);
if (error)
goto out_unlock_internal;
goto start;
}
error = -ret;
if (ret <= 0)
goto out_unlock_internal;
XFS_STATS_ADD(xs_write_bytes, ret);
/* Handle various SYNC-type writes */
if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) {
loff_t end = pos + ret - 1;
int error2;
xfs_iunlock(ip, iolock);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error2 = filemap_write_and_wait_range(mapping, pos, end);
if (!error)
error = error2;
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, iolock);
error2 = -xfs_file_fsync(file, file->f_path.dentry,
(file->f_flags & __O_SYNC) ? 0 : 1);
if (!error)
error = error2;
}
out_unlock_internal:
if (ip->i_new_size) {
xfs_ilock(ip, XFS_ILOCK_EXCL);
ip->i_new_size = 0;
/*
* If this was a direct or synchronous I/O that failed (such
* as ENOSPC) then part of the I/O may have been written to
* disk before the error occured. In this case the on-disk
* file size may have been adjusted beyond the in-memory file
* size and now needs to be truncated back.
*/
if (ip->i_d.di_size > ip->i_size)
ip->i_d.di_size = ip->i_size;
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
xfs_iunlock(ip, iolock);
out_unlock_mutex:
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
return -error;
}
ssize_t /* bytes written, or (-) error */
xfs_write(
bhv_desc_t *bdp,
struct kiocb *iocb,
const struct iovec *iovp,
unsigned int nsegs,
loff_t *offset,
int ioflags,
cred_t *credp)
{
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
unsigned long segs = nsegs;
xfs_inode_t *xip;
xfs_mount_t *mp;
ssize_t ret = 0, error = 0;
xfs_fsize_t isize, new_size;
xfs_iocore_t *io;
vnode_t *vp;
unsigned long seg;
int iolock;
int eventsent = 0;
vrwlock_t locktype;
size_t ocount = 0, count;
loff_t pos;
int need_isem = 1, need_flush = 0;
XFS_STATS_INC(xs_write_calls);
vp = BHV_TO_VNODE(bdp);
xip = XFS_BHVTOI(bdp);
for (seg = 0; seg < segs; seg++) {
const struct iovec *iv = &iovp[seg];
/*
* If any segment has a negative length, or the cumulative
* length ever wraps negative then return -EINVAL.
*/
ocount += iv->iov_len;
if (unlikely((ssize_t)(ocount|iv->iov_len) < 0))
return -EINVAL;
if (access_ok(VERIFY_READ, iv->iov_base, iv->iov_len))
continue;
if (seg == 0)
return -EFAULT;
segs = seg;
ocount -= iv->iov_len; /* This segment is no good */
break;
}
count = ocount;
pos = *offset;
if (count == 0)
return 0;
io = &xip->i_iocore;
mp = io->io_mount;
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
if (ioflags & IO_ISDIRECT) {
xfs_buftarg_t *target =
(xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((pos & target->pbr_smask) || (count & target->pbr_smask))
return XFS_ERROR(-EINVAL);
if (!VN_CACHED(vp) && pos < i_size_read(inode))
need_isem = 0;
if (VN_CACHED(vp))
need_flush = 1;
}
relock:
if (need_isem) {
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
down(&inode->i_sem);
} else {
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
}
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
isize = i_size_read(inode);
if (file->f_flags & O_APPEND)
*offset = isize;
start:
error = -generic_write_checks(file, &pos, &count,
S_ISBLK(inode->i_mode));
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_isem;
}
new_size = pos + count;
if (new_size > isize)
io->io_new_size = new_size;
if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
loff_t savedsize = pos;
int dmflags = FILP_DELAY_FLAG(file);
if (need_isem)
dmflags |= DM_FLAGS_ISEM;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
pos, count,
dmflags, &locktype);
if (error) {
xfs_iunlock(xip, iolock);
goto out_unlock_isem;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reaquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) && savedsize != isize) {
pos = isize = xip->i_d.di_size;
goto start;
}
}
/*
* On Linux, generic_file_write updates the times even if
* no data is copied in so long as the write had a size.
*
* We must update xfs' times since revalidate will overcopy xfs.
*/
if (!(ioflags & IO_INVIS)) {
xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
inode_update_time(inode, 1);
}
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (pos > isize) {
error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos,
isize, pos + count);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_isem;
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (likely(!error))
error = -remove_suid(file->f_dentry);
if (unlikely(error)) {
xfs_iunlock(xip, iolock);
goto out_unlock_isem;
}
}
retry:
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
if ((ioflags & IO_ISDIRECT)) {
if (need_flush) {
xfs_inval_cached_trace(io, pos, -1,
ctooff(offtoct(pos)), -1);
VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(pos)),
-1, FI_REMAPF_LOCKED);
}
if (need_isem) {
/* demote the lock now the cached pages are gone */
XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
up(&inode->i_sem);
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
need_isem = 0;
}
xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_direct_write(iocb, iovp,
&segs, pos, offset, count, ocount);
/*
* direct-io write to a hole: fall through to buffered I/O
* for completing the rest of the request.
*/
if (ret >= 0 && ret != count) {
XFS_STATS_ADD(xs_write_bytes, ret);
pos += ret;
count -= ret;
need_isem = 1;
ioflags &= ~IO_ISDIRECT;
xfs_iunlock(xip, iolock);
goto relock;
}
} else {
xfs_rw_enter_trace(XFS_WRITE_ENTER, io, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_buffered_write(iocb, iovp, segs,
pos, offset, count, ret);
}
current->backing_dev_info = NULL;
if (ret == -EIOCBQUEUED)
ret = wait_on_sync_kiocb(iocb);
if ((ret == -ENOSPC) &&
DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
!(ioflags & IO_INVIS)) {
xfs_rwunlock(bdp, locktype);
error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (error)
goto out_unlock_isem;
xfs_rwlock(bdp, locktype);
pos = xip->i_d.di_size;
goto retry;
}
if (*offset > xip->i_d.di_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_d.di_size) {
xip->i_d.di_size = *offset;
i_size_write(inode, *offset);
xip->i_update_core = 1;
xip->i_update_size = 1;
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
error = -ret;
if (ret <= 0)
goto out_unlock_internal;
XFS_STATS_ADD(xs_write_bytes, ret);
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
/*
* If we're treating this as O_DSYNC and we have not updated the
* size, force the log.
*/
if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
!(xip->i_update_size)) {
xfs_inode_log_item_t *iip = xip->i_itemp;
/*
* If an allocation transaction occurred
* without extending the size, then we have to force
* the log up the proper point to ensure that the
* allocation is permanent. We can't count on
* the fact that buffered writes lock out direct I/O
* writes - the direct I/O write could have extended
* the size nontransactionally, then finished before
* we started. xfs_write_file will think that the file
* didn't grow but the update isn't safe unless the
* size change is logged.
*
* Force the log if we've committed a transaction
* against the inode or if someone else has and
* the commit record hasn't gone to disk (e.g.
* the inode is pinned). This guarantees that
* all changes affecting the inode are permanent
* when we return.
*/
if (iip && iip->ili_last_lsn) {
xfs_log_force(mp, iip->ili_last_lsn,
XFS_LOG_FORCE | XFS_LOG_SYNC);
} else if (xfs_ipincount(xip) > 0) {
xfs_log_force(mp, (xfs_lsn_t)0,
XFS_LOG_FORCE | XFS_LOG_SYNC);
}
} else {
xfs_trans_t *tp;
/*
* O_SYNC or O_DSYNC _with_ a size update are handled
* the same way.
*
* If the write was synchronous then we need to make
* sure that the inode modification time is permanent.
* We'll have updated the timestamp above, so here
* we use a synchronous transaction to log the inode.
* It's not fast, but it's necessary.
*
* If this a dsync write and the size got changed
* non-transactionally, then we need to ensure that
* the size change gets logged in a synchronous
* transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
if ((error = xfs_trans_reserve(tp, 0,
XFS_SWRITE_LOG_RES(mp),
0, 0, 0))) {
/* Transaction reserve failed */
xfs_trans_cancel(tp, 0);
} else {
/* Transaction reserve successful */
xfs_ilock(xip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, xip);
xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0, NULL);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
if (error)
goto out_unlock_internal;
}
}
xfs_rwunlock(bdp, locktype);
if (need_isem)
up(&inode->i_sem);
error = sync_page_range(inode, mapping, pos, ret);
if (!error)
error = ret;
return error;
}
out_unlock_internal:
xfs_rwunlock(bdp, locktype);
out_unlock_isem:
if (need_isem)
up(&inode->i_sem);
return -error;
}
/*
* Truncate file. Must have write permission and not be a directory.
*/
int
xfs_setattr_size(
struct xfs_inode *ip,
struct iattr *iattr)
{
struct xfs_mount *mp = ip->i_mount;
struct inode *inode = VFS_I(ip);
xfs_off_t oldsize, newsize;
struct xfs_trans *tp;
int error;
uint lock_flags = 0;
bool did_zeroing = false;
trace_xfs_setattr(ip);
if (mp->m_flags & XFS_MOUNT_RDONLY)
return -EROFS;
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
error = inode_change_ok(inode, iattr);
if (error)
return error;
ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL));
ASSERT(S_ISREG(inode->i_mode));
ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
oldsize = inode->i_size;
newsize = iattr->ia_size;
/*
* Short circuit the truncate case for zero length files.
*/
if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
if (!(iattr->ia_valid & (ATTR_CTIME|ATTR_MTIME)))
return 0;
/*
* Use the regular setattr path to update the timestamps.
*/
iattr->ia_valid &= ~ATTR_SIZE;
return xfs_setattr_nonsize(ip, iattr, 0);
}
/*
* Make sure that the dquots are attached to the inode.
*/
error = xfs_qm_dqattach(ip, 0);
if (error)
return error;
/*
* Wait for all direct I/O to complete.
*/
inode_dio_wait(inode);
/*
* File data changes must be complete before we start the transaction to
* modify the inode. 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.
*
* Start with zeroing any data beyond EOF that we may expose on file
* extension, or zeroing out the rest of the block on a downward
* truncate.
*/
if (newsize > oldsize) {
error = xfs_zero_eof(ip, newsize, oldsize, &did_zeroing);
} else {
error = iomap_truncate_page(inode, newsize, &did_zeroing,
&xfs_iomap_ops);
}
if (error)
return error;
/*
* We are going to log the inode size change in this transaction so
* any previous writes that are beyond the on disk EOF and the new
* EOF that have not been written out need to be written here. If we
* do not write the data out, we expose ourselves to the null files
* problem. Note that this includes any block zeroing we did above;
* otherwise those blocks may not be zeroed after a crash.
*/
if (did_zeroing ||
(newsize > ip->i_d.di_size && oldsize != ip->i_d.di_size)) {
error = filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
ip->i_d.di_size, newsize);
if (error)
return error;
}
/*
* We've already locked out new page faults, so now we can safely remove
* pages from the page cache knowing they won't get refaulted until we
* drop the XFS_MMAP_EXCL lock after the extent manipulations are
* complete. The truncate_setsize() call also cleans partial EOF page
* PTEs on extending truncates and hence ensures sub-page block size
* filesystems are correctly handled, too.
*
* We have to do all the page cache truncate work outside the
* transaction context as the "lock" order is page lock->log space
* reservation as defined by extent allocation in the writeback path.
* Hence a truncate can fail with ENOMEM from xfs_trans_alloc(), but
* having already truncated the in-memory version of the file (i.e. made
* user visible changes). There's not much we can do about this, except
* to hope that the caller sees ENOMEM and retries the truncate
* operation.
*/
truncate_setsize(inode, newsize);
error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
if (error)
return error;
lock_flags |= XFS_ILOCK_EXCL;
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, 0);
/*
* 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.
*
* The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
* special case where we need to update the times despite not having
* these flags set. For all other operations the VFS set these flags
* explicitly if it wants a timestamp update.
*/
if (newsize != oldsize &&
!(iattr->ia_valid & (ATTR_CTIME | ATTR_MTIME))) {
iattr->ia_ctime = iattr->ia_mtime =
current_fs_time(inode->i_sb);
iattr->ia_valid |= ATTR_CTIME | ATTR_MTIME;
}
/*
* The first thing we do is set the size to new_size permanently on
* disk. This way we don't have to worry about anyone ever being able
* to look at the data being freed even in the face of a crash.
* What we're getting around here is the case where we free a block, it
* is allocated to another file, it is written to, and then we crash.
* If the new data gets written to the file but the log buffers
* containing the free and reallocation don't, then we'd end up with
* garbage in the blocks being freed. As long as we make the new size
* permanent before actually freeing any blocks it doesn't matter if
* they get written to.
*/
ip->i_d.di_size = newsize;
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
if (newsize <= oldsize) {
error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, newsize);
if (error)
goto out_trans_cancel;
/*
* Truncated "down", so we're removing references to old data
* here - if we delay flushing for a long time, we expose
* ourselves unduly to the notorious NULL files problem. So,
* we mark this inode and flush it when the file is closed,
* and do not wait the usual (long) time for writeout.
*/
xfs_iflags_set(ip, XFS_ITRUNCATED);
/* A truncate down always removes post-EOF blocks. */
xfs_inode_clear_eofblocks_tag(ip);
}
if (iattr->ia_valid & ATTR_MODE)
xfs_setattr_mode(ip, iattr);
if (iattr->ia_valid & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
xfs_setattr_time(ip, iattr);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
XFS_STATS_INC(mp, xs_ig_attrchg);
if (mp->m_flags & XFS_MOUNT_WSYNC)
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp);
out_unlock:
if (lock_flags)
xfs_iunlock(ip, lock_flags);
return error;
out_trans_cancel:
xfs_trans_cancel(tp);
goto out_unlock;
}
/*
* Common pre-write limit and setup checks.
*
* Returns with iolock held according to @iolock.
*/
STATIC ssize_t
xfs_file_aio_write_checks(
struct file *file,
loff_t *pos,
size_t *count,
xfs_fsize_t *new_sizep,
int *iolock)
{
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
xfs_fsize_t new_size;
int error = 0;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
*new_sizep = 0;
restart:
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
*iolock = 0;
return error;
}
if (likely(!(file->f_mode & FMODE_NOCMTIME)))
file_update_time(file);
/*
* If the offset is beyond the size of the file, we need to zero any
* blocks that fall between the existing EOF and the start of this
* write. There is no need to issue zeroing if another in-flght IO ends
* at or before this one If zeronig is needed and we are currently
* holding the iolock shared, we need to update it to exclusive which
* involves dropping all locks and relocking to maintain correct locking
* order. If we do this, restart the function to ensure all checks and
* values are still valid.
*/
if ((ip->i_new_size && *pos > ip->i_new_size) ||
(!ip->i_new_size && *pos > ip->i_size)) {
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
goto restart;
}
error = -xfs_zero_eof(ip, *pos, ip->i_size);
}
/*
* If this IO extends beyond EOF, we may need to update ip->i_new_size.
* We have already zeroed space beyond EOF (if necessary). Only update
* ip->i_new_size if this IO ends beyond any other in-flight writes.
*/
new_size = *pos + *count;
if (new_size > ip->i_size) {
if (new_size > ip->i_new_size)
ip->i_new_size = new_size;
*new_sizep = new_size;
}
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL);
if (error)
return error;
/*
* If we're writing the file then make sure to clear the setuid and
* setgid bits if the process is not being run by root. This keeps
* people from modifying setuid and setgid binaries.
*/
return file_remove_suid(file);
}
/*
* Truncate file. Must have write permission and not be a directory.
*/
int
xfs_setattr_size(
struct xfs_inode *ip,
struct iattr *iattr,
int flags)
{
struct xfs_mount *mp = ip->i_mount;
struct inode *inode = VFS_I(ip);
int mask = iattr->ia_valid;
struct xfs_trans *tp;
int error;
uint lock_flags;
uint commit_flags = 0;
trace_xfs_setattr(ip);
if (mp->m_flags & XFS_MOUNT_RDONLY)
return XFS_ERROR(EROFS);
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
error = -inode_change_ok(inode, iattr);
if (error)
return XFS_ERROR(error);
ASSERT(S_ISREG(ip->i_d.di_mode));
ASSERT((mask & (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
ATTR_MTIME_SET|ATTR_KILL_SUID|ATTR_KILL_SGID|
ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
lock_flags = XFS_ILOCK_EXCL;
if (!(flags & XFS_ATTR_NOLOCK))
lock_flags |= XFS_IOLOCK_EXCL;
xfs_ilock(ip, lock_flags);
/*
* Short circuit the truncate case for zero length files.
*/
if (iattr->ia_size == 0 &&
ip->i_size == 0 && ip->i_d.di_nextents == 0) {
if (!(mask & (ATTR_CTIME|ATTR_MTIME)))
goto out_unlock;
/*
* Use the regular setattr path to update the timestamps.
*/
xfs_iunlock(ip, lock_flags);
iattr->ia_valid &= ~ATTR_SIZE;
return xfs_setattr_nonsize(ip, iattr, 0);
}
/*
* Make sure that the dquots are attached to the inode.
*/
error = xfs_qm_dqattach_locked(ip, 0);
if (error)
goto out_unlock;
/*
* Now we can make the changes. Before we join the inode to the
* transaction, 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 (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
* i_size.
*/
error = xfs_zero_eof(ip, iattr->ia_size, ip->i_size);
if (error)
goto out_unlock;
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
lock_flags &= ~XFS_ILOCK_EXCL;
/*
* We are going to log the inode size change in this transaction so
* any previous writes that are beyond the on disk EOF and the new
* EOF that have not been written out need to be written here. If we
* do not write the data out, we expose ourselves to the null files
* problem.
*
* Only flush from the on disk size to the smaller of the in memory
* file size or the new size as that's the range we really care about
* here and prevents waiting for other data not within the range we
* care about here.
*/
if (ip->i_size != ip->i_d.di_size && iattr->ia_size > ip->i_d.di_size) {
error = xfs_flush_pages(ip, ip->i_d.di_size, iattr->ia_size,
XBF_ASYNC, FI_NONE);
if (error)
goto out_unlock;
}
/*
* Wait for all I/O to complete.
*/
xfs_ioend_wait(ip);
error = -block_truncate_page(inode->i_mapping, iattr->ia_size,
xfs_get_blocks);
if (error)
goto out_unlock;
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0,
XFS_TRANS_PERM_LOG_RES,
XFS_ITRUNCATE_LOG_COUNT);
if (error)
goto out_trans_cancel;
truncate_setsize(inode, iattr->ia_size);
commit_flags = XFS_TRANS_RELEASE_LOG_RES;
lock_flags |= XFS_ILOCK_EXCL;
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip);
/*
* 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.
*
* The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
* special case where we need to update the times despite not having
* these flags set. For all other operations the VFS set these flags
* explicitly if it wants a timestamp update.
*/
if (iattr->ia_size != ip->i_size &&
(!(mask & (ATTR_CTIME | ATTR_MTIME)))) {
iattr->ia_ctime = iattr->ia_mtime =
current_fs_time(inode->i_sb);
mask |= ATTR_CTIME | ATTR_MTIME;
}
if (iattr->ia_size > ip->i_size) {
ip->i_d.di_size = iattr->ia_size;
ip->i_size = iattr->ia_size;
} else if (iattr->ia_size <= ip->i_size ||
(iattr->ia_size == 0 && ip->i_d.di_nextents)) {
error = xfs_itruncate_data(&tp, ip, iattr->ia_size);
if (error)
goto out_trans_abort;
/*
* Truncated "down", so we're removing references to old data
* here - if we delay flushing for a long time, we expose
* ourselves unduly to the notorious NULL files problem. So,
* we mark this inode and flush it when the file is closed,
* and do not wait the usual (long) time for writeout.
*/
xfs_iflags_set(ip, XFS_ITRUNCATED);
}
if (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;
}
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;
ip->i_update_core = 1;
}
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
XFS_STATS_INC(xs_ig_attrchg);
if (mp->m_flags & XFS_MOUNT_WSYNC)
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
out_unlock:
if (lock_flags)
xfs_iunlock(ip, lock_flags);
return error;
out_trans_abort:
commit_flags |= XFS_TRANS_ABORT;
out_trans_cancel:
xfs_trans_cancel(tp, commit_flags);
goto out_unlock;
}
int
xfs_setattr_size(
struct xfs_inode *ip,
struct iattr *iattr,
int flags)
{
struct xfs_mount *mp = ip->i_mount;
struct inode *inode = VFS_I(ip);
int mask = iattr->ia_valid;
xfs_off_t oldsize, newsize;
struct xfs_trans *tp;
int error;
uint lock_flags;
uint commit_flags = 0;
trace_xfs_setattr(ip);
if (mp->m_flags & XFS_MOUNT_RDONLY)
return XFS_ERROR(EROFS);
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
error = -inode_change_ok(inode, iattr);
if (error)
return XFS_ERROR(error);
ASSERT(S_ISREG(ip->i_d.di_mode));
ASSERT((mask & (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
ATTR_MTIME_SET|ATTR_KILL_SUID|ATTR_KILL_SGID|
ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
lock_flags = XFS_ILOCK_EXCL;
if (!(flags & XFS_ATTR_NOLOCK))
lock_flags |= XFS_IOLOCK_EXCL;
xfs_ilock(ip, lock_flags);
oldsize = inode->i_size;
newsize = iattr->ia_size;
if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
if (!(mask & (ATTR_CTIME|ATTR_MTIME)))
goto out_unlock;
xfs_iunlock(ip, lock_flags);
iattr->ia_valid &= ~ATTR_SIZE;
return xfs_setattr_nonsize(ip, iattr, 0);
}
error = xfs_qm_dqattach_locked(ip, 0);
if (error)
goto out_unlock;
if (newsize > oldsize) {
error = xfs_zero_eof(ip, newsize, oldsize);
if (error)
goto out_unlock;
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
lock_flags &= ~XFS_ILOCK_EXCL;
/*
* We are going to log the inode size change in this transaction so
* any previous writes that are beyond the on disk EOF and the new
* EOF that have not been written out need to be written here. If we
* do not write the data out, we expose ourselves to the null files
* problem.
*
* Only flush from the on disk size to the smaller of the in memory
* file size or the new size as that's the range we really care about
* here and prevents waiting for other data not within the range we
* care about here.
*/
if (oldsize != ip->i_d.di_size && newsize > ip->i_d.di_size) {
error = xfs_flush_pages(ip, ip->i_d.di_size, newsize, 0,
FI_NONE);
if (error)
goto out_unlock;
}
inode_dio_wait(inode);
error = -block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
if (error)
goto out_unlock;
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0,
XFS_TRANS_PERM_LOG_RES,
XFS_ITRUNCATE_LOG_COUNT);
if (error)
goto out_trans_cancel;
truncate_setsize(inode, newsize);
commit_flags = XFS_TRANS_RELEASE_LOG_RES;
lock_flags |= XFS_ILOCK_EXCL;
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, 0);
if (newsize != oldsize && (!(mask & (ATTR_CTIME | ATTR_MTIME)))) {
iattr->ia_ctime = iattr->ia_mtime =
current_fs_time(inode->i_sb);
mask |= ATTR_CTIME | ATTR_MTIME;
}
/*
* The first thing we do is set the size to new_size permanently on
* disk. This way we don't have to worry about anyone ever being able
* to look at the data being freed even in the face of a crash.
* What we're getting around here is the case where we free a block, it
* is allocated to another file, it is written to, and then we crash.
* If the new data gets written to the file but the log buffers
* containing the free and reallocation don't, then we'd end up with
* garbage in the blocks being freed. As long as we make the new size
* permanent before actually freeing any blocks it doesn't matter if
* they get written to.
*/
ip->i_d.di_size = newsize;
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
if (newsize <= oldsize) {
error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, newsize);
if (error)
goto out_trans_abort;
xfs_iflags_set(ip, XFS_ITRUNCATED);
}
if (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;
}
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;
}
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
XFS_STATS_INC(xs_ig_attrchg);
if (mp->m_flags & XFS_MOUNT_WSYNC)
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
out_unlock:
if (lock_flags)
xfs_iunlock(ip, lock_flags);
return error;
out_trans_abort:
commit_flags |= XFS_TRANS_ABORT;
out_trans_cancel:
xfs_trans_cancel(tp, commit_flags);
goto out_unlock;
}
ssize_t /* bytes written, or (-) error */
xfs_write(
bhv_desc_t *bdp,
struct kiocb *iocb,
const struct iovec *iovp,
unsigned int nsegs,
loff_t *offset,
int ioflags,
cred_t *credp)
{
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
unsigned long segs = nsegs;
xfs_inode_t *xip;
xfs_mount_t *mp;
ssize_t ret = 0, error = 0;
xfs_fsize_t isize, new_size;
xfs_iocore_t *io;
bhv_vnode_t *vp;
unsigned long seg;
int iolock;
int eventsent = 0;
bhv_vrwlock_t locktype;
size_t ocount = 0, count;
loff_t pos;
int need_i_mutex = 1, need_flush = 0;
XFS_STATS_INC(xs_write_calls);
vp = BHV_TO_VNODE(bdp);
xip = XFS_BHVTOI(bdp);
for (seg = 0; seg < segs; seg++) {
const struct iovec *iv = &iovp[seg];
/*
* If any segment has a negative length, or the cumulative
* length ever wraps negative then return -EINVAL.
*/
ocount += iv->iov_len;
if (unlikely((ssize_t)(ocount|iv->iov_len) < 0))
return -EINVAL;
if (access_ok(VERIFY_READ, iv->iov_base, iv->iov_len))
continue;
if (seg == 0)
return -EFAULT;
segs = seg;
ocount -= iv->iov_len; /* This segment is no good */
break;
}
count = ocount;
pos = *offset;
if (count == 0)
return 0;
io = &xip->i_iocore;
mp = io->io_mount;
vfs_wait_for_freeze(vp->v_vfsp, SB_FREEZE_WRITE);
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
if (ioflags & IO_ISDIRECT) {
xfs_buftarg_t *target =
(xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((pos & target->bt_smask) || (count & target->bt_smask))
return XFS_ERROR(-EINVAL);
if (!VN_CACHED(vp) && pos < i_size_read(inode))
need_i_mutex = 0;
if (VN_CACHED(vp))
need_flush = 1;
}
relock:
if (need_i_mutex) {
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
mutex_lock(&inode->i_mutex);
} else {
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
}
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
isize = i_size_read(inode);
if (file->f_flags & O_APPEND)
*offset = isize;
start:
error = -generic_write_checks(file, &pos, &count,
S_ISBLK(inode->i_mode));
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_mutex;
}
new_size = pos + count;
if (new_size > isize)
io->io_new_size = new_size;
if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
loff_t savedsize = pos;
int dmflags = FILP_DELAY_FLAG(file);
if (need_i_mutex)
dmflags |= DM_FLAGS_IMUX;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
pos, count,
dmflags, &locktype);
if (error) {
xfs_iunlock(xip, iolock);
goto out_unlock_mutex;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reacquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) && savedsize != isize) {
pos = isize = xip->i_d.di_size;
goto start;
}
}
if (likely(!(ioflags & IO_INVIS))) {
file_update_time(file);
xfs_ichgtime_fast(xip, inode,
XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
}
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (pos > isize) {
error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos, isize);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_mutex;
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (likely(!error))
error = -remove_suid(file->f_path.dentry);
if (unlikely(error)) {
xfs_iunlock(xip, iolock);
goto out_unlock_mutex;
}
}
retry:
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
if ((ioflags & IO_ISDIRECT)) {
if (need_flush) {
xfs_inval_cached_trace(io, pos, -1,
ctooff(offtoct(pos)), -1);
bhv_vop_flushinval_pages(vp, ctooff(offtoct(pos)),
-1, FI_REMAPF_LOCKED);
}
if (need_i_mutex) {
/* demote the lock now the cached pages are gone */
XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
mutex_unlock(&inode->i_mutex);
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
need_i_mutex = 0;
}
xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_direct_write(iocb, iovp,
&segs, pos, offset, count, ocount);
/*
* direct-io write to a hole: fall through to buffered I/O
* for completing the rest of the request.
*/
if (ret >= 0 && ret != count) {
XFS_STATS_ADD(xs_write_bytes, ret);
pos += ret;
count -= ret;
need_i_mutex = 1;
ioflags &= ~IO_ISDIRECT;
xfs_iunlock(xip, iolock);
goto relock;
}
} else {
xfs_rw_enter_trace(XFS_WRITE_ENTER, io, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_buffered_write(iocb, iovp, segs,
pos, offset, count, ret);
}
current->backing_dev_info = NULL;
if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
ret = wait_on_sync_kiocb(iocb);
if ((ret == -ENOSPC) &&
DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
!(ioflags & IO_INVIS)) {
xfs_rwunlock(bdp, locktype);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (error)
goto out_nounlocks;
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_rwlock(bdp, locktype);
pos = xip->i_d.di_size;
ret = 0;
goto retry;
}
isize = i_size_read(inode);
if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
*offset = isize;
if (*offset > xip->i_d.di_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_d.di_size) {
xip->i_d.di_size = *offset;
i_size_write(inode, *offset);
xip->i_update_core = 1;
xip->i_update_size = 1;
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
error = -ret;
if (ret <= 0)
goto out_unlock_internal;
XFS_STATS_ADD(xs_write_bytes, ret);
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
error = xfs_write_sync_logforce(mp, xip);
if (error)
goto out_unlock_internal;
xfs_rwunlock(bdp, locktype);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error = sync_page_range(inode, mapping, pos, ret);
if (!error)
error = ret;
return error;
}
out_unlock_internal:
xfs_rwunlock(bdp, locktype);
out_unlock_mutex:
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
out_nounlocks:
return -error;
}
ssize_t /* bytes written, or (-) error */
xfs_write(
bhv_desc_t *bdp,
uio_t *uio,
int ioflag,
cred_t *credp)
{
xfs_inode_t *xip;
xfs_mount_t *mp;
ssize_t ret = 0;
int error = 0;
xfs_fsize_t isize, new_size;
xfs_fsize_t n, limit;
xfs_fsize_t size;
xfs_iocore_t *io;
xfs_vnode_t *vp;
int iolock;
//int eventsent = 0;
vrwlock_t locktype;
xfs_off_t offset_c;
xfs_off_t *offset;
xfs_off_t pos;
XFS_STATS_INC(xs_write_calls);
vp = BHV_TO_VNODE(bdp);
xip = XFS_BHVTOI(bdp);
io = &xip->i_iocore;
mp = io->io_mount;
if (XFS_FORCED_SHUTDOWN(xip->i_mount)) {
return EIO;
}
size = uio->uio_resid;
pos = offset_c = uio->uio_offset;
offset = &offset_c;
if (unlikely(ioflag & IO_ISDIRECT)) {
if (((__psint_t)buf & BBMASK) ||
(*offset & mp->m_blockmask) ||
(size & mp->m_blockmask)) {
return EINVAL;
}
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
} else {
if (io->io_flags & XFS_IOCORE_RT)
return EINVAL;
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
}
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
isize = xip->i_d.di_size;
limit = XFS_MAXIOFFSET(mp);
if (ioflag & O_APPEND)
*offset = isize;
//start:
n = limit - *offset;
if (n <= 0) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return EFBIG;
}
if (n < size)
size = n;
new_size = *offset + size;
if (new_size > isize) {
io->io_new_size = new_size;
}
#ifdef RMC
/* probably be a long time before if ever that we do dmapi */
if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
loff_t savedsize = *offset;
int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
*offset, size,
dmflags, &locktype);
if (error) {
if (iolock) xfs_iunlock(xip, iolock);
return -error;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reaquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) &&
savedsize != xip->i_d.di_size) {
*offset = isize = xip->i_d.di_size;
goto start;
}
}
#endif
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (!(ioflag & IO_ISDIRECT) && (*offset > isize && isize)) {
error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, *offset,
isize, *offset + size);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return(-error);
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
#if 0
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (likely(!error))
error = -remove_suid(file->f_dentry);
if (unlikely(error)) {
xfs_iunlock(xip, iolock);
goto out_unlock_mutex;
}
}
#endif
//retry:
if (unlikely(ioflag & IO_ISDIRECT)) {
#ifdef RMC
xfs_off_t pos = *offset;
struct address_space *mapping = file->f_dentry->d_inode->i_mapping;
struct inode *inode = mapping->host;
ret = precheck_file_write(file, inode, &size, &pos);
if (ret || size == 0)
goto error;
xfs_inval_cached_pages(vp, io, pos, 1, 1);
inode->i_ctime = inode->i_mtime = CURRENT_TIME;
/* mark_inode_dirty_sync(inode); - we do this later */
xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, buf, size, pos, ioflags);
ret = generic_file_direct_IO(WRITE, file, (char *)buf, size, pos);
xfs_inval_cached_pages(vp, io, pos, 1, 1);
if (ret > 0)
*offset += ret;
#endif
} else {
xfs_rw_enter_trace(XFS_WRITE_ENTER, io, buf, size, *offset, ioflags);
ret = xfs_write_file(xip,uio,ioflag);
}
xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
//error:
if (ret <= 0) {
if (iolock)
xfs_rwunlock(bdp, locktype);
return ret;
}
XFS_STATS_ADD(xs_write_bytes, ret);
if (*offset > xip->i_d.di_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_d.di_size) {
printf("xfs_write look at doing more here %s:%d\n",__FILE__,__LINE__);
#ifdef RMC
struct inode *inode = LINVFS_GET_IP(vp);
i_size_write(inode, *offset);
mark_inode_dirty_sync(inode);
#endif
xip->i_d.di_size = *offset;
xip->i_update_core = 1;
xip->i_update_size = 1;
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
/* Handle various SYNC-type writes */
#if 0
// if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
#endif
if (ioflag & IO_SYNC) {
/*
* If we're treating this as O_DSYNC and we have not updated the
* size, force the log.
*/
if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
!(xip->i_update_size)) {
xfs_inode_log_item_t *iip = xip->i_itemp;
/*
* If an allocation transaction occurred
* without extending the size, then we have to force
* the log up the proper point to ensure that the
* allocation is permanent. We can't count on
* the fact that buffered writes lock out direct I/O
* writes - the direct I/O write could have extended
* the size nontransactionally, then finished before
* we started. xfs_write_file will think that the file
* didn't grow but the update isn't safe unless the
* size change is logged.
*
* Force the log if we've committed a transaction
* against the inode or if someone else has and
* the commit record hasn't gone to disk (e.g.
* the inode is pinned). This guarantees that
* all changes affecting the inode are permanent
* when we return.
*/
if (iip && iip->ili_last_lsn) {
xfs_log_force(mp, iip->ili_last_lsn,
XFS_LOG_FORCE | XFS_LOG_SYNC);
} else if (xfs_ipincount(xip) > 0) {
xfs_log_force(mp, (xfs_lsn_t)0,
XFS_LOG_FORCE | XFS_LOG_SYNC);
}
} else {
xfs_trans_t *tp;
/*
* O_SYNC or O_DSYNC _with_ a size update are handled
* the same way.
*
* If the write was synchronous then we need to make
* sure that the inode modification time is permanent.
* We'll have updated the timestamp above, so here
* we use a synchronous transaction to log the inode.
* It's not fast, but it's necessary.
*
* If this a dsync write and the size got changed
* non-transactionally, then we need to ensure that
* the size change gets logged in a synchronous
* transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
if ((error = xfs_trans_reserve(tp, 0,
XFS_SWRITE_LOG_RES(mp),
0, 0, 0))) {
/* Transaction reserve failed */
xfs_trans_cancel(tp, 0);
} else {
/* Transaction reserve successful */
xfs_ilock(xip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, xip);
xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0, NULL);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
if (error)
goto out_unlock_internal;
}
xfs_rwunlock(bdp, locktype);
return ret;
} /* (ioflags & O_SYNC) */
out_unlock_internal:
xfs_rwunlock(bdp, locktype);
#if 0
out_unlock_mutex:
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
#endif
//out_nounlocks:
return -error;
}
ssize_t /* bytes written, or (-) error */
xfs_write(
struct xfs_inode *xip,
struct kiocb *iocb,
const struct iovec *iovp,
unsigned int nsegs,
loff_t *offset,
int ioflags)
{
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
unsigned long segs = nsegs;
xfs_mount_t *mp;
ssize_t ret = 0, error = 0;
xfs_fsize_t isize, new_size;
int iolock;
int eventsent = 0;
size_t ocount = 0, count;
loff_t pos;
int need_i_mutex;
XFS_STATS_INC(xs_write_calls);
error = generic_segment_checks(iovp, &segs, &ocount, VERIFY_READ);
if (error)
return error;
count = ocount;
pos = *offset;
if (count == 0)
return 0;
mp = xip->i_mount;
xfs_wait_for_freeze(mp, SB_FREEZE_WRITE);
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
relock:
if (ioflags & IO_ISDIRECT) {
iolock = XFS_IOLOCK_SHARED;
need_i_mutex = 0;
} else {
iolock = XFS_IOLOCK_EXCL;
need_i_mutex = 1;
mutex_lock(&inode->i_mutex);
}
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
start:
error = -generic_write_checks(file, &pos, &count,
S_ISBLK(inode->i_mode));
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
goto out_unlock_mutex;
}
if ((DM_EVENT_ENABLED(xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
int dmflags = FILP_DELAY_FLAG(file);
if (need_i_mutex)
dmflags |= DM_FLAGS_IMUX;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, xip,
pos, count, dmflags, &iolock);
if (error) {
goto out_unlock_internal;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reacquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) && pos != xip->i_size)
goto start;
}
if (ioflags & IO_ISDIRECT) {
xfs_buftarg_t *target =
XFS_IS_REALTIME_INODE(xip) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((pos & target->bt_smask) || (count & target->bt_smask)) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return XFS_ERROR(-EINVAL);
}
if (!need_i_mutex && (mapping->nrpages || pos > xip->i_size)) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
iolock = XFS_IOLOCK_EXCL;
need_i_mutex = 1;
mutex_lock(&inode->i_mutex);
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
goto start;
}
}
new_size = pos + count;
if (new_size > xip->i_size)
xip->i_new_size = new_size;
if (likely(!(ioflags & IO_INVIS)))
xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (pos > xip->i_size) {
error = xfs_zero_eof(xip, pos, xip->i_size);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL);
goto out_unlock_internal;
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (likely(!error))
error = -file_remove_suid(file);
if (unlikely(error)) {
goto out_unlock_internal;
}
}
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
if ((ioflags & IO_ISDIRECT)) {
if (mapping->nrpages) {
WARN_ON(need_i_mutex == 0);
xfs_inval_cached_trace(xip, pos, -1,
(pos & PAGE_CACHE_MASK), -1);
error = xfs_flushinval_pages(xip,
(pos & PAGE_CACHE_MASK),
-1, FI_REMAPF_LOCKED);
if (error)
goto out_unlock_internal;
}
if (need_i_mutex) {
/* demote the lock now the cached pages are gone */
xfs_ilock_demote(xip, XFS_IOLOCK_EXCL);
mutex_unlock(&inode->i_mutex);
iolock = XFS_IOLOCK_SHARED;
need_i_mutex = 0;
}
xfs_rw_enter_trace(XFS_DIOWR_ENTER, xip, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_direct_write(iocb, iovp,
&segs, pos, offset, count, ocount);
/*
* direct-io write to a hole: fall through to buffered I/O
* for completing the rest of the request.
*/
if (ret >= 0 && ret != count) {
XFS_STATS_ADD(xs_write_bytes, ret);
pos += ret;
count -= ret;
ioflags &= ~IO_ISDIRECT;
xfs_iunlock(xip, iolock);
goto relock;
}
} else {
xfs_rw_enter_trace(XFS_WRITE_ENTER, xip, (void *)iovp, segs,
*offset, ioflags);
ret = generic_file_buffered_write(iocb, iovp, segs,
pos, offset, count, ret);
}
current->backing_dev_info = NULL;
if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
ret = wait_on_sync_kiocb(iocb);
isize = i_size_read(inode);
if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
*offset = isize;
if (*offset > xip->i_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_size)
xip->i_size = *offset;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
if (ret == -ENOSPC &&
DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
xfs_iunlock(xip, iolock);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, xip,
DM_RIGHT_NULL, xip, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_ilock(xip, iolock);
if (error)
goto out_unlock_internal;
goto start;
}
error = -ret;
if (ret <= 0)
goto out_unlock_internal;
XFS_STATS_ADD(xs_write_bytes, ret);
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
int error2;
xfs_iunlock(xip, iolock);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error2 = sync_page_range(inode, mapping, pos, ret);
if (!error)
error = error2;
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_ilock(xip, iolock);
error2 = xfs_write_sync_logforce(mp, xip);
if (!error)
error = error2;
}
out_unlock_internal:
if (xip->i_new_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
xip->i_new_size = 0;
/*
* If this was a direct or synchronous I/O that failed (such
* as ENOSPC) then part of the I/O may have been written to
* disk before the error occured. In this case the on-disk
* file size may have been adjusted beyond the in-memory file
* size and now needs to be truncated back.
*/
if (xip->i_d.di_size > xip->i_size)
xip->i_d.di_size = xip->i_size;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
xfs_iunlock(xip, iolock);
out_unlock_mutex:
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
return -error;
}
/*
* Truncate file. Must have write permission and not be a directory.
*/
int
xfs_setattr_size(
struct xfs_inode *ip,
struct iattr *iattr)
{
struct xfs_mount *mp = ip->i_mount;
struct inode *inode = VFS_I(ip);
xfs_off_t oldsize, newsize;
struct xfs_trans *tp;
int error;
uint lock_flags = 0;
uint commit_flags = 0;
trace_xfs_setattr(ip);
if (mp->m_flags & XFS_MOUNT_RDONLY)
return XFS_ERROR(EROFS);
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
error = -inode_change_ok(inode, iattr);
if (error)
return XFS_ERROR(error);
ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
ASSERT(S_ISREG(ip->i_d.di_mode));
ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET|
ATTR_MTIME_SET|ATTR_KILL_PRIV|ATTR_TIMES_SET)) == 0);
oldsize = inode->i_size;
newsize = iattr->ia_size;
/*
* Short circuit the truncate case for zero length files.
*/
if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
if (!(iattr->ia_valid & (ATTR_CTIME|ATTR_MTIME)))
return 0;
/*
* Use the regular setattr path to update the timestamps.
*/
iattr->ia_valid &= ~ATTR_SIZE;
return xfs_setattr_nonsize(ip, iattr, 0);
}
/*
* Make sure that the dquots are attached to the inode.
*/
error = xfs_qm_dqattach(ip, 0);
if (error)
return error;
/*
* Now we can make the changes. Before we join the inode to the
* transaction, 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 (newsize > oldsize) {
/*
* 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
* i_size.
*/
error = xfs_zero_eof(ip, newsize, oldsize);
if (error)
return error;
}
/*
* We are going to log the inode size change in this transaction so
* any previous writes that are beyond the on disk EOF and the new
* EOF that have not been written out need to be written here. If we
* do not write the data out, we expose ourselves to the null files
* problem.
*
* Only flush from the on disk size to the smaller of the in memory
* file size or the new size as that's the range we really care about
* here and prevents waiting for other data not within the range we
* care about here.
*/
if (oldsize != ip->i_d.di_size && newsize > ip->i_d.di_size) {
error = -filemap_write_and_wait_range(VFS_I(ip)->i_mapping,
ip->i_d.di_size, newsize);
if (error)
return error;
}
/*
* Wait for all direct I/O to complete.
*/
inode_dio_wait(inode);
error = -block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
if (error)
return error;
tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
if (error)
goto out_trans_cancel;
truncate_setsize(inode, newsize);
commit_flags = XFS_TRANS_RELEASE_LOG_RES;
lock_flags |= XFS_ILOCK_EXCL;
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, 0);
/*
* 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.
*
* The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
* special case where we need to update the times despite not having
* these flags set. For all other operations the VFS set these flags
* explicitly if it wants a timestamp update.
*/
if (newsize != oldsize &&
!(iattr->ia_valid & (ATTR_CTIME | ATTR_MTIME))) {
iattr->ia_ctime = iattr->ia_mtime =
current_fs_time(inode->i_sb);
iattr->ia_valid |= ATTR_CTIME | ATTR_MTIME;
}
/*
* The first thing we do is set the size to new_size permanently on
* disk. This way we don't have to worry about anyone ever being able
* to look at the data being freed even in the face of a crash.
* What we're getting around here is the case where we free a block, it
* is allocated to another file, it is written to, and then we crash.
* If the new data gets written to the file but the log buffers
* containing the free and reallocation don't, then we'd end up with
* garbage in the blocks being freed. As long as we make the new size
* permanent before actually freeing any blocks it doesn't matter if
* they get written to.
*/
ip->i_d.di_size = newsize;
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
if (newsize <= oldsize) {
error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, newsize);
if (error)
goto out_trans_abort;
/*
* Truncated "down", so we're removing references to old data
* here - if we delay flushing for a long time, we expose
* ourselves unduly to the notorious NULL files problem. So,
* we mark this inode and flush it when the file is closed,
* and do not wait the usual (long) time for writeout.
*/
xfs_iflags_set(ip, XFS_ITRUNCATED);
/* A truncate down always removes post-EOF blocks. */
xfs_inode_clear_eofblocks_tag(ip);
}
if (iattr->ia_valid & ATTR_MODE)
xfs_setattr_mode(ip, iattr);
if (iattr->ia_valid & (ATTR_ATIME|ATTR_CTIME|ATTR_MTIME))
xfs_setattr_time(ip, iattr);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
XFS_STATS_INC(xs_ig_attrchg);
if (mp->m_flags & XFS_MOUNT_WSYNC)
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
out_unlock:
if (lock_flags)
xfs_iunlock(ip, lock_flags);
return error;
out_trans_abort:
commit_flags |= XFS_TRANS_ABORT;
out_trans_cancel:
xfs_trans_cancel(tp, commit_flags);
goto out_unlock;
}
/*
* Common pre-write limit and setup checks.
*
* Called with the iolocked held either shared and exclusive according to
* @iolock, and returns with it held. Might upgrade the iolock to exclusive
* if called for a direct write beyond i_size.
*/
STATIC ssize_t
xfs_file_aio_write_checks(
struct kiocb *iocb,
struct iov_iter *from,
int *iolock)
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
ssize_t error = 0;
size_t count = iov_iter_count(from);
bool drained_dio = false;
restart:
error = generic_write_checks(iocb, from);
if (error <= 0)
return error;
error = xfs_break_layouts(inode, iolock);
if (error)
return error;
/*
* For changing security info in file_remove_privs() we need i_rwsem
* exclusively.
*/
if (*iolock == XFS_IOLOCK_SHARED && !IS_NOSEC(inode)) {
xfs_iunlock(ip, *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_ilock(ip, *iolock);
goto restart;
}
/*
* If the offset is beyond the size of the file, we need to zero any
* blocks that fall between the existing EOF and the start of this
* write. If zeroing is needed and we are currently holding the
* iolock shared, we need to update it to exclusive which implies
* having to redo all checks before.
*
* We need to serialise against EOF updates that occur in IO
* completions here. We want to make sure that nobody is changing the
* size while we do this check until we have placed an IO barrier (i.e.
* hold the XFS_IOLOCK_EXCL) that prevents new IO from being dispatched.
* The spinlock effectively forms a memory barrier once we have the
* XFS_IOLOCK_EXCL so we are guaranteed to see the latest EOF value
* and hence be able to correctly determine if we need to run zeroing.
*/
spin_lock(&ip->i_flags_lock);
if (iocb->ki_pos > i_size_read(inode)) {
spin_unlock(&ip->i_flags_lock);
if (!drained_dio) {
if (*iolock == XFS_IOLOCK_SHARED) {
xfs_iunlock(ip, *iolock);
*iolock = XFS_IOLOCK_EXCL;
xfs_ilock(ip, *iolock);
iov_iter_reexpand(from, count);
}
/*
* We now have an IO submission barrier in place, but
* AIO can do EOF updates during IO completion and hence
* we now need to wait for all of them to drain. Non-AIO
* DIO will have drained before we are given the
* XFS_IOLOCK_EXCL, and so for most cases this wait is a
* no-op.
*/
inode_dio_wait(inode);
drained_dio = true;
goto restart;
}
error = xfs_zero_eof(ip, iocb->ki_pos, i_size_read(inode), NULL);
if (error)
return error;
} else
spin_unlock(&ip->i_flags_lock);
/*
* Updating the timestamps will grab the ilock again from
* xfs_fs_dirty_inode, so we have to call it after dropping the
* lock above. Eventually we should look into a way to avoid
* the pointless lock roundtrip.
*/
if (likely(!(file->f_mode & FMODE_NOCMTIME))) {
error = file_update_time(file);
if (error)
return error;
}
/*
* If we're writing the file then make sure to clear the setuid and
* setgid bits if the process is not being run by root. This keeps
* people from modifying setuid and setgid binaries.
*/
if (!IS_NOSEC(inode))
return file_remove_privs(file);
return 0;
}
ssize_t /* bytes written, or (-) error */
xfs_write(
bhv_desc_t *bdp,
struct kiocb *iocb,
const struct iovec *iovp,
unsigned int segs,
loff_t *offset,
int ioflags,
cred_t *credp)
{
struct file *file = iocb->ki_filp;
size_t size = 0;
xfs_inode_t *xip;
xfs_mount_t *mp;
ssize_t ret;
int error = 0;
xfs_fsize_t isize, new_size;
xfs_fsize_t n, limit;
xfs_iocore_t *io;
vnode_t *vp;
unsigned long seg;
int iolock;
int eventsent = 0;
vrwlock_t locktype;
XFS_STATS_INC(xs_write_calls);
vp = BHV_TO_VNODE(bdp);
vn_trace_entry(vp, "xfs_write", (inst_t *)__return_address);
xip = XFS_BHVTOI(bdp);
/* START copy & waste from filemap.c */
for (seg = 0; seg < segs; seg++) {
const struct iovec *iv = &iovp[seg];
/*
* If any segment has a negative length, or the cumulative
* length ever wraps negative then return -EINVAL.
*/
size += iv->iov_len;
if (unlikely((ssize_t)(size|iv->iov_len) < 0))
return XFS_ERROR(-EINVAL);
}
/* END copy & waste from filemap.c */
if (size == 0)
return 0;
io = &(xip->i_iocore);
mp = io->io_mount;
xfs_check_frozen(mp, bdp, XFS_FREEZE_WRITE);
if (XFS_FORCED_SHUTDOWN(mp)) {
return -EIO;
}
if (ioflags & IO_ISDIRECT) {
pb_target_t *target =
(xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((*offset & target->pbr_smask) ||
(size & target->pbr_smask)) {
return XFS_ERROR(-EINVAL);
}
iolock = XFS_IOLOCK_SHARED;
locktype = VRWLOCK_WRITE_DIRECT;
} else {
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
}
xfs_ilock(xip, XFS_ILOCK_EXCL|iolock);
isize = xip->i_d.di_size;
limit = XFS_MAXIOFFSET(mp);
if (file->f_flags & O_APPEND)
*offset = isize;
start:
n = limit - *offset;
if (n <= 0) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return -EFBIG;
}
if (n < size)
size = n;
new_size = *offset + size;
if (new_size > isize) {
io->io_new_size = new_size;
}
if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) &&
!(ioflags & IO_INVIS) && !eventsent)) {
loff_t savedsize = *offset;
xfs_iunlock(xip, XFS_ILOCK_EXCL);
error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp,
*offset, size,
FILP_DELAY_FLAG(file), &locktype);
if (error) {
xfs_iunlock(xip, iolock);
return -error;
}
xfs_ilock(xip, XFS_ILOCK_EXCL);
eventsent = 1;
/*
* The iolock was dropped and reaquired in XFS_SEND_DATA
* so we have to recheck the size when appending.
* We will only "goto start;" once, since having sent the
* event prevents another call to XFS_SEND_DATA, which is
* what allows the size to change in the first place.
*/
if ((file->f_flags & O_APPEND) &&
savedsize != xip->i_d.di_size) {
*offset = isize = xip->i_d.di_size;
goto start;
}
}
/*
* On Linux, generic_file_write updates the times even if
* no data is copied in so long as the write had a size.
*
* We must update xfs' times since revalidate will overcopy xfs.
*/
if (size && !(ioflags & IO_INVIS))
xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
/*
* If the offset is beyond the size of the file, we have a couple
* of things to do. First, if there is already space allocated
* we need to either create holes or zero the disk or ...
*
* If there is a page where the previous size lands, we need
* to zero it out up to the new size.
*/
if (!(ioflags & IO_ISDIRECT) && (*offset > isize && isize)) {
error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, *offset,
isize, *offset + size);
if (error) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
return(-error);
}
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
/*
* If we're writing the file then make sure to clear the
* setuid and setgid bits if the process is not being run
* by root. This keeps people from modifying setuid and
* setgid binaries.
*/
if (((xip->i_d.di_mode & S_ISUID) ||
((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) ==
(S_ISGID | S_IXGRP))) &&
!capable(CAP_FSETID)) {
error = xfs_write_clear_setuid(xip);
if (error) {
xfs_iunlock(xip, iolock);
return -error;
}
}
retry:
if (ioflags & IO_ISDIRECT) {
xfs_inval_cached_pages(vp, &xip->i_iocore, *offset, 1, 1);
}
ret = generic_file_aio_write_nolock(iocb, iovp, segs, offset);
if ((ret == -ENOSPC) &&
DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) &&
!(ioflags & IO_INVIS)) {
xfs_rwunlock(bdp, locktype);
error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
0, 0, 0); /* Delay flag intentionally unused */
if (error)
return -error;
xfs_rwlock(bdp, locktype);
*offset = xip->i_d.di_size;
goto retry;
}
if (*offset > xip->i_d.di_size) {
xfs_ilock(xip, XFS_ILOCK_EXCL);
if (*offset > xip->i_d.di_size) {
struct inode *inode = LINVFS_GET_IP(vp);
xip->i_d.di_size = *offset;
i_size_write(inode, *offset);
xip->i_update_core = 1;
xip->i_update_size = 1;
}
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
if (ret <= 0) {
xfs_rwunlock(bdp, locktype);
return ret;
}
XFS_STATS_ADD(xs_write_bytes, ret);
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(file->f_dentry->d_inode)) {
/*
* If we're treating this as O_DSYNC and we have not updated the
* size, force the log.
*/
if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC)
&& !(xip->i_update_size)) {
/*
* If an allocation transaction occurred
* without extending the size, then we have to force
* the log up the proper point to ensure that the
* allocation is permanent. We can't count on
* the fact that buffered writes lock out direct I/O
* writes - the direct I/O write could have extended
* the size nontransactionally, then finished before
* we started. xfs_write_file will think that the file
* didn't grow but the update isn't safe unless the
* size change is logged.
*
* Force the log if we've committed a transaction
* against the inode or if someone else has and
* the commit record hasn't gone to disk (e.g.
* the inode is pinned). This guarantees that
* all changes affecting the inode are permanent
* when we return.
*/
xfs_inode_log_item_t *iip;
xfs_lsn_t lsn;
iip = xip->i_itemp;
if (iip && iip->ili_last_lsn) {
lsn = iip->ili_last_lsn;
xfs_log_force(mp, lsn,
XFS_LOG_FORCE | XFS_LOG_SYNC);
} else if (xfs_ipincount(xip) > 0) {
xfs_log_force(mp, (xfs_lsn_t)0,
XFS_LOG_FORCE | XFS_LOG_SYNC);
}
} else {
xfs_trans_t *tp;
/*
* O_SYNC or O_DSYNC _with_ a size update are handled
* the same way.
*
* If the write was synchronous then we need to make
* sure that the inode modification time is permanent.
* We'll have updated the timestamp above, so here
* we use a synchronous transaction to log the inode.
* It's not fast, but it's necessary.
*
* If this a dsync write and the size got changed
* non-transactionally, then we need to ensure that
* the size change gets logged in a synchronous
* transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
if ((error = xfs_trans_reserve(tp, 0,
XFS_SWRITE_LOG_RES(mp),
0, 0, 0))) {
/* Transaction reserve failed */
xfs_trans_cancel(tp, 0);
} else {
/* Transaction reserve successful */
xfs_ilock(xip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, xip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, xip);
xfs_trans_log_inode(tp, xip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0, (xfs_lsn_t)0);
xfs_iunlock(xip, XFS_ILOCK_EXCL);
}
}
} /* (ioflags & O_SYNC) */
xfs_rwunlock(bdp, locktype);
return(ret);
}
