struct vnode *
vn_initialize(
struct inode *inode)
{
struct vnode *vp = LINVFS_GET_VP(inode);
XFS_STATS_INC(vn_active);
XFS_STATS_INC(vn_alloc);
vp->v_flag = VMODIFIED;
spinlock_init(&vp->v_lock, "v_lock");
spin_lock(&vnumber_lock);
if (!++vn_generation) /* v_number shouldn't be zero */
vn_generation++;
vp->v_number = vn_generation;
spin_unlock(&vnumber_lock);
ASSERT(VN_CACHED(vp) == 0);
/* Initialize the first behavior and the behavior chain head. */
vn_bhv_head_init(VN_BHV_HEAD(vp), "vnode");
#ifdef XFS_VNODE_TRACE
vp->v_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
#endif /* XFS_VNODE_TRACE */
vn_trace_exit(vp, "vn_initialize", (inst_t *)__return_address);
return vp;
}
/*
* Test whether it is appropriate to check an inode for and free post EOF
* blocks. The 'force' parameter determines whether we should also consider
* regular files that are marked preallocated or append-only.
*/
bool
xfs_can_free_eofblocks(struct xfs_inode *ip, bool force)
{
/* prealloc/delalloc exists only on regular files */
if (!S_ISREG(ip->i_d.di_mode))
return false;
/*
* Zero sized files with no cached pages and delalloc blocks will not
* have speculative prealloc/delalloc blocks to remove.
*/
if (VFS_I(ip)->i_size == 0 &&
VN_CACHED(VFS_I(ip)) == 0 &&
ip->i_delayed_blks == 0)
return false;
/* If we haven't read in the extent list, then don't do it now. */
if (!(ip->i_df.if_flags & XFS_IFEXTENTS))
return false;
/*
* Do not free real preallocated or append-only files unless the file
* has delalloc blocks and we are forced to remove them.
*/
if (ip->i_d.di_flags & (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND))
if (!force || ip->i_delayed_blks == 0)
return false;
return true;
}
/*
* vnode pcache layer for vnode_tosspages.
* 'last' parameter unused but left in for IRIX compatibility
*/
void
fs_tosspages(
bhv_desc_t *bdp,
xfs_off_t first,
xfs_off_t last,
int fiopt)
{
vnode_t *vp = BHV_TO_VNODE(bdp);
struct inode *ip = LINVFS_GET_IP(vp);
if (VN_CACHED(vp))
truncate_inode_pages(ip->i_mapping, first);
}
/*
* 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)
{
if (VN_CACHED(vp)) {
xfs_inval_cached_trace(io, offset, -1, ctooff(offtoct(offset)), -1);
VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(offset)), -1, FI_REMAPF_LOCKED);
}
}
/*
* vnode pcache layer for vnode_flushinval_pages.
* 'last' parameter unused but left in for IRIX compatibility
*/
void
fs_flushinval_pages(
bhv_desc_t *bdp,
xfs_off_t first,
xfs_off_t last,
int fiopt)
{
vnode_t *vp = BHV_TO_VNODE(bdp);
struct inode *ip = LINVFS_GET_IP(vp);
if (VN_CACHED(vp)) {
filemap_fdatasync(ip->i_mapping);
fsync_inode_data_buffers(ip);
filemap_fdatawait(ip->i_mapping);
truncate_inode_pages(ip->i_mapping, first);
}
}
/*
* 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);
}
}
/*
* vnode pcache layer for vnode_flush_pages.
* 'last' parameter unused but left in for IRIX compatibility
*/
int
fs_flush_pages(
bhv_desc_t *bdp,
xfs_off_t first,
xfs_off_t last,
uint64_t flags,
int fiopt)
{
vnode_t *vp = BHV_TO_VNODE(bdp);
struct inode *ip = LINVFS_GET_IP(vp);
if (VN_CACHED(vp)) {
filemap_fdatasync(ip->i_mapping);
fsync_inode_data_buffers(ip);
filemap_fdatawait(ip->i_mapping);
}
return 0;
}
int
xfs_swap_extents(
xfs_inode_t *ip, /* target inode */
xfs_inode_t *tip, /* tmp inode */
xfs_swapext_t *sxp)
{
xfs_mount_t *mp = ip->i_mount;
xfs_trans_t *tp;
xfs_bstat_t *sbp = &sxp->sx_stat;
xfs_ifork_t *tempifp, *ifp, *tifp;
int src_log_flags, target_log_flags;
int error = 0;
int aforkblks = 0;
int taforkblks = 0;
__uint64_t tmp;
tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL);
if (!tempifp) {
error = XFS_ERROR(ENOMEM);
goto out;
}
/*
* we have to do two separate lock calls here to keep lockdep
* happy. If we try to get all the locks in one call, lock will
* report false positives when we drop the ILOCK and regain them
* below.
*/
xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL);
xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
/* Verify that both files have the same format */
if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) {
error = XFS_ERROR(EINVAL);
goto out_unlock;
}
/* Verify both files are either real-time or non-realtime */
if (XFS_IS_REALTIME_INODE(ip) != XFS_IS_REALTIME_INODE(tip)) {
error = XFS_ERROR(EINVAL);
goto out_unlock;
}
error = -filemap_write_and_wait(VFS_I(tip)->i_mapping);
if (error)
goto out_unlock;
truncate_pagecache_range(VFS_I(tip), 0, -1);
/* Verify O_DIRECT for ftmp */
if (VN_CACHED(VFS_I(tip)) != 0) {
error = XFS_ERROR(EINVAL);
goto out_unlock;
}
/* Verify all data are being swapped */
if (sxp->sx_offset != 0 ||
sxp->sx_length != ip->i_d.di_size ||
sxp->sx_length != tip->i_d.di_size) {
error = XFS_ERROR(EFAULT);
goto out_unlock;
}
trace_xfs_swap_extent_before(ip, 0);
trace_xfs_swap_extent_before(tip, 1);
/* check inode formats now that data is flushed */
error = xfs_swap_extents_check_format(ip, tip);
if (error) {
xfs_notice(mp,
"%s: inode 0x%llx format is incompatible for exchanging.",
__func__, ip->i_ino);
goto out_unlock;
}
/*
* Compare the current change & modify times with that
* passed in. If they differ, we abort this swap.
* This is the mechanism used to ensure the calling
* process that the file was not changed out from
* under it.
*/
if ((sbp->bs_ctime.tv_sec != VFS_I(ip)->i_ctime.tv_sec) ||
(sbp->bs_ctime.tv_nsec != VFS_I(ip)->i_ctime.tv_nsec) ||
(sbp->bs_mtime.tv_sec != VFS_I(ip)->i_mtime.tv_sec) ||
(sbp->bs_mtime.tv_nsec != VFS_I(ip)->i_mtime.tv_nsec)) {
error = XFS_ERROR(EBUSY);
goto out_unlock;
}
/* We need to fail if the file is memory mapped. Once we have tossed
* all existing pages, the page fault will have no option
* but to go to the filesystem for pages. By making the page fault call
* vop_read (or write in the case of autogrow) they block on the iolock
* until we have switched the extents.
*/
if (VN_MAPPED(VFS_I(ip))) {
error = XFS_ERROR(EBUSY);
goto out_unlock;
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
xfs_iunlock(tip, XFS_ILOCK_EXCL);
/*
* There is a race condition here since we gave up the
* ilock. However, the data fork will not change since
* we have the iolock (locked for truncation too) so we
* are safe. We don't really care if non-io related
* fields change.
*/
truncate_pagecache_range(VFS_I(ip), 0, -1);
tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0);
if (error) {
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
xfs_iunlock(tip, XFS_IOLOCK_EXCL);
xfs_trans_cancel(tp, 0);
goto out;
}
xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
/*
* Count the number of extended attribute blocks
*/
if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) &&
(ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks);
if (error)
goto out_trans_cancel;
}
if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) &&
(tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK,
&taforkblks);
if (error)
goto out_trans_cancel;
}
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
/*
* Before we've swapped the forks, lets set the owners of the forks
* appropriately. We have to do this as we are demand paging the btree
* buffers, and so the validation done on read will expect the owner
* field to be correctly set. Once we change the owners, we can swap the
* inode forks.
*
* Note the trickiness in setting the log flags - we set the owner log
* flag on the opposite inode (i.e. the inode we are setting the new
* owner to be) because once we swap the forks and log that, log
* recovery is going to see the fork as owned by the swapped inode,
* not the pre-swapped inodes.
*/
src_log_flags = XFS_ILOG_CORE;
target_log_flags = XFS_ILOG_CORE;
if (ip->i_d.di_version == 3 &&
ip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
target_log_flags |= XFS_ILOG_DOWNER;
error = xfs_bmbt_change_owner(tp, ip, XFS_DATA_FORK,
tip->i_ino, NULL);
if (error)
goto out_trans_cancel;
}
if (tip->i_d.di_version == 3 &&
tip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
src_log_flags |= XFS_ILOG_DOWNER;
error = xfs_bmbt_change_owner(tp, tip, XFS_DATA_FORK,
ip->i_ino, NULL);
if (error)
goto out_trans_cancel;
}
/*
* Swap the data forks of the inodes
*/
ifp = &ip->i_df;
tifp = &tip->i_df;
*tempifp = *ifp; /* struct copy */
*ifp = *tifp; /* struct copy */
*tifp = *tempifp; /* struct copy */
/*
* Fix the on-disk inode values
*/
tmp = (__uint64_t)ip->i_d.di_nblocks;
ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks;
tip->i_d.di_nblocks = tmp + taforkblks - aforkblks;
tmp = (__uint64_t) ip->i_d.di_nextents;
ip->i_d.di_nextents = tip->i_d.di_nextents;
tip->i_d.di_nextents = tmp;
tmp = (__uint64_t) ip->i_d.di_format;
ip->i_d.di_format = tip->i_d.di_format;
tip->i_d.di_format = tmp;
/*
* The extents in the source inode could still contain speculative
* preallocation beyond EOF (e.g. the file is open but not modified
* while defrag is in progress). In that case, we need to copy over the
* number of delalloc blocks the data fork in the source inode is
* tracking beyond EOF so that when the fork is truncated away when the
* temporary inode is unlinked we don't underrun the i_delayed_blks
* counter on that inode.
*/
ASSERT(tip->i_delayed_blks == 0);
tip->i_delayed_blks = ip->i_delayed_blks;
ip->i_delayed_blks = 0;
switch (ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) {
ifp->if_u1.if_extents =
ifp->if_u2.if_inline_ext;
}
src_log_flags |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
ASSERT(ip->i_d.di_version < 3 ||
(src_log_flags & XFS_ILOG_DOWNER));
src_log_flags |= XFS_ILOG_DBROOT;
break;
}
switch (tip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) {
tifp->if_u1.if_extents =
tifp->if_u2.if_inline_ext;
}
target_log_flags |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
target_log_flags |= XFS_ILOG_DBROOT;
ASSERT(tip->i_d.di_version < 3 ||
(target_log_flags & XFS_ILOG_DOWNER));
break;
}
xfs_trans_log_inode(tp, ip, src_log_flags);
xfs_trans_log_inode(tp, tip, target_log_flags);
/*
* If this is a synchronous mount, make sure that the
* transaction goes to disk before returning to the user.
*/
if (mp->m_flags & XFS_MOUNT_WSYNC)
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
trace_xfs_swap_extent_after(ip, 0);
trace_xfs_swap_extent_after(tip, 1);
out:
kmem_free(tempifp);
return error;
out_unlock:
xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
xfs_iunlock(tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
goto out;
out_trans_cancel:
xfs_trans_cancel(tp, 0);
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;
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;
}
ssize_t /* bytes read, or (-) error */
xfs_read(
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;
struct inode *inode = file->f_mapping->host;
size_t size = 0;
ssize_t ret;
xfs_fsize_t n;
xfs_inode_t *ip;
xfs_mount_t *mp;
vnode_t *vp;
unsigned long seg;
ip = XFS_BHVTOI(bdp);
vp = BHV_TO_VNODE(bdp);
mp = ip->i_mount;
XFS_STATS_INC(xs_read_calls);
/* 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 (unlikely(ioflags & IO_ISDIRECT)) {
xfs_buftarg_t *target =
(ip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((*offset & target->bt_smask) ||
(size & target->bt_smask)) {
if (*offset == ip->i_d.di_size) {
return (0);
}
return -XFS_ERROR(EINVAL);
}
}
n = XFS_MAXIOFFSET(mp) - *offset;
if ((n <= 0) || (size == 0))
return 0;
if (n < size)
size = n;
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
if (unlikely(ioflags & IO_ISDIRECT))
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, XFS_IOLOCK_SHARED);
if (DM_EVENT_ENABLED(vp->v_vfsp, ip, DM_EVENT_READ) &&
!(ioflags & IO_INVIS)) {
vrwlock_t locktype = VRWLOCK_READ;
int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
ret = -XFS_SEND_DATA(mp, DM_EVENT_READ,
BHV_TO_VNODE(bdp), *offset, size,
dmflags, &locktype);
if (ret) {
xfs_iunlock(ip, XFS_IOLOCK_SHARED);
goto unlock_mutex;
}
}
if (unlikely((ioflags & IO_ISDIRECT) && VN_CACHED(vp)))
VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(*offset)),
-1, FI_REMAPF_LOCKED);
xfs_rw_enter_trace(XFS_READ_ENTER, &ip->i_iocore,
(void *)iovp, segs, *offset, ioflags);
ret = __generic_file_aio_read(iocb, iovp, segs, offset);
if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
ret = wait_on_sync_kiocb(iocb);
if (ret > 0)
XFS_STATS_ADD(xs_read_bytes, ret);
xfs_iunlock(ip, XFS_IOLOCK_SHARED);
unlock_mutex:
if (unlikely(ioflags & IO_ISDIRECT))
mutex_unlock(&inode->i_mutex);
return ret;
}
/*
* Syssgi interface for swapext
*/
int
xfs_swapext(
xfs_swapext_t __user *sxp)
{
xfs_swapext_t sx;
xfs_inode_t *ip=NULL, *tip=NULL, *ips[2];
xfs_trans_t *tp;
xfs_mount_t *mp;
xfs_bstat_t *sbp;
struct file *fp = NULL, *tfp = NULL;
vnode_t *vp, *tvp;
bhv_desc_t *bdp, *tbdp;
vn_bhv_head_t *bhp, *tbhp;
uint lock_flags=0;
int ilf_fields, tilf_fields;
int error = 0;
xfs_ifork_t tempif, *ifp, *tifp;
__uint64_t tmp;
int aforkblks = 0;
int taforkblks = 0;
int locked = 0;
if (copy_from_user(&sx, sxp, sizeof(sx)))
return XFS_ERROR(EFAULT);
/* Pull information for the target fd */
if (((fp = fget((int)sx.sx_fdtarget)) == NULL) ||
((vp = LINVFS_GET_VP(fp->f_dentry->d_inode)) == NULL)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
bhp = VN_BHV_HEAD(vp);
bdp = vn_bhv_lookup(bhp, &xfs_vnodeops);
if (bdp == NULL) {
error = XFS_ERROR(EBADF);
goto error0;
} else {
ip = XFS_BHVTOI(bdp);
}
if (((tfp = fget((int)sx.sx_fdtmp)) == NULL) ||
((tvp = LINVFS_GET_VP(tfp->f_dentry->d_inode)) == NULL)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
tbhp = VN_BHV_HEAD(tvp);
tbdp = vn_bhv_lookup(tbhp, &xfs_vnodeops);
if (tbdp == NULL) {
error = XFS_ERROR(EBADF);
goto error0;
} else {
tip = XFS_BHVTOI(tbdp);
}
if (ip->i_mount != tip->i_mount) {
error = XFS_ERROR(EINVAL);
goto error0;
}
if (ip->i_ino == tip->i_ino) {
error = XFS_ERROR(EINVAL);
goto error0;
}
mp = ip->i_mount;
sbp = &sx.sx_stat;
if (XFS_FORCED_SHUTDOWN(mp)) {
error = XFS_ERROR(EIO);
goto error0;
}
locked = 1;
/* Lock in i_ino order */
if (ip->i_ino < tip->i_ino) {
ips[0] = ip;
ips[1] = tip;
} else {
ips[0] = tip;
ips[1] = ip;
}
lock_flags = XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL;
xfs_lock_inodes(ips, 2, 0, lock_flags);
/* Check permissions */
error = xfs_iaccess(ip, S_IWUSR, NULL);
if (error)
goto error0;
error = xfs_iaccess(tip, S_IWUSR, NULL);
if (error)
goto error0;
/* Verify that both files have the same format */
if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/* Verify both files are either real-time or non-realtime */
if ((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) !=
(tip->i_d.di_flags & XFS_DIFLAG_REALTIME)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/* Should never get a local format */
if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||
tip->i_d.di_format == XFS_DINODE_FMT_LOCAL) {
error = XFS_ERROR(EINVAL);
goto error0;
}
if (VN_CACHED(tvp) != 0)
xfs_inval_cached_pages(XFS_ITOV(tip), &(tip->i_iocore),
(xfs_off_t)0, 0, 0);
/* Verify O_DIRECT for ftmp */
if (VN_CACHED(tvp) != 0) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/* Verify all data are being swapped */
if (sx.sx_offset != 0 ||
sx.sx_length != ip->i_d.di_size ||
sx.sx_length != tip->i_d.di_size) {
error = XFS_ERROR(EFAULT);
goto error0;
}
/*
* If the target has extended attributes, the tmp file
* must also in order to ensure the correct data fork
* format.
*/
if ( XFS_IFORK_Q(ip) != XFS_IFORK_Q(tip) ) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/*
* Compare the current change & modify times with that
* passed in. If they differ, we abort this swap.
* This is the mechanism used to ensure the calling
* process that the file was not changed out from
* under it.
*/
if ((sbp->bs_ctime.tv_sec != ip->i_d.di_ctime.t_sec) ||
(sbp->bs_ctime.tv_nsec != ip->i_d.di_ctime.t_nsec) ||
(sbp->bs_mtime.tv_sec != ip->i_d.di_mtime.t_sec) ||
(sbp->bs_mtime.tv_nsec != ip->i_d.di_mtime.t_nsec)) {
error = XFS_ERROR(EBUSY);
goto error0;
}
/* We need to fail if the file is memory mapped. Once we have tossed
* all existing pages, the page fault will have no option
* but to go to the filesystem for pages. By making the page fault call
* VOP_READ (or write in the case of autogrow) they block on the iolock
* until we have switched the extents.
*/
if (VN_MAPPED(vp)) {
error = XFS_ERROR(EBUSY);
goto error0;
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
xfs_iunlock(tip, XFS_ILOCK_EXCL);
/*
* There is a race condition here since we gave up the
* ilock. However, the data fork will not change since
* we have the iolock (locked for truncation too) so we
* are safe. We don't really care if non-io related
* fields change.
*/
VOP_TOSS_PAGES(vp, 0, -1, FI_REMAPF);
tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT);
if ((error = xfs_trans_reserve(tp, 0,
XFS_ICHANGE_LOG_RES(mp), 0,
0, 0))) {
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
xfs_iunlock(tip, XFS_IOLOCK_EXCL);
xfs_trans_cancel(tp, 0);
return error;
}
xfs_lock_inodes(ips, 2, 0, XFS_ILOCK_EXCL);
/*
* Count the number of extended attribute blocks
*/
if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) &&
(ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks);
if (error) {
xfs_iunlock(ip, lock_flags);
xfs_iunlock(tip, lock_flags);
xfs_trans_cancel(tp, 0);
return error;
}
}
if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) &&
(tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK,
&taforkblks);
if (error) {
xfs_iunlock(ip, lock_flags);
xfs_iunlock(tip, lock_flags);
xfs_trans_cancel(tp, 0);
return error;
}
}
/*
* Swap the data forks of the inodes
*/
ifp = &ip->i_df;
tifp = &tip->i_df;
tempif = *ifp; /* struct copy */
*ifp = *tifp; /* struct copy */
*tifp = tempif; /* struct copy */
/*
* Fix the on-disk inode values
*/
tmp = (__uint64_t)ip->i_d.di_nblocks;
ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks;
tip->i_d.di_nblocks = tmp + taforkblks - aforkblks;
tmp = (__uint64_t) ip->i_d.di_nextents;
ip->i_d.di_nextents = tip->i_d.di_nextents;
tip->i_d.di_nextents = tmp;
tmp = (__uint64_t) ip->i_d.di_format;
ip->i_d.di_format = tip->i_d.di_format;
tip->i_d.di_format = tmp;
ilf_fields = XFS_ILOG_CORE;
switch(ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) {
ifp->if_u1.if_extents =
ifp->if_u2.if_inline_ext;
}
ilf_fields |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
ilf_fields |= XFS_ILOG_DBROOT;
break;
}
tilf_fields = XFS_ILOG_CORE;
switch(tip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) {
tifp->if_u1.if_extents =
tifp->if_u2.if_inline_ext;
}
tilf_fields |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
tilf_fields |= XFS_ILOG_DBROOT;
break;
}
/*
* Increment vnode ref counts since xfs_trans_commit &
* xfs_trans_cancel will both unlock the inodes and
* decrement the associated ref counts.
*/
VN_HOLD(vp);
VN_HOLD(tvp);
xfs_trans_ijoin(tp, ip, lock_flags);
xfs_trans_ijoin(tp, tip, lock_flags);
xfs_trans_log_inode(tp, ip, ilf_fields);
xfs_trans_log_inode(tp, tip, tilf_fields);
/*
* If this is a synchronous mount, make sure that the
* transaction goes to disk before returning to the user.
*/
if (mp->m_flags & XFS_MOUNT_WSYNC) {
xfs_trans_set_sync(tp);
}
error = xfs_trans_commit(tp, XFS_TRANS_SWAPEXT, NULL);
fput(fp);
fput(tfp);
return error;
error0:
if (locked) {
xfs_iunlock(ip, lock_flags);
xfs_iunlock(tip, lock_flags);
}
if (fp != NULL) fput(fp);
if (tfp != NULL) fput(tfp);
return error;
}
int
xfs_swap_extents(
xfs_inode_t *ip, /* target inode */
xfs_inode_t *tip, /* tmp inode */
xfs_swapext_t *sxp)
{
xfs_mount_t *mp;
xfs_trans_t *tp;
xfs_bstat_t *sbp = &sxp->sx_stat;
xfs_ifork_t *tempifp, *ifp, *tifp;
int ilf_fields, tilf_fields;
static uint lock_flags = XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL;
int error = 0;
int aforkblks = 0;
int taforkblks = 0;
__uint64_t tmp;
char locked = 0;
mp = ip->i_mount;
tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL);
if (!tempifp) {
error = XFS_ERROR(ENOMEM);
goto error0;
}
sbp = &sxp->sx_stat;
/*
* we have to do two separate lock calls here to keep lockdep
* happy. If we try to get all the locks in one call, lock will
* report false positives when we drop the ILOCK and regain them
* below.
*/
xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL);
xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
locked = 1;
/* Verify that both files have the same format */
if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/* Verify both files are either real-time or non-realtime */
if (XFS_IS_REALTIME_INODE(ip) != XFS_IS_REALTIME_INODE(tip)) {
error = XFS_ERROR(EINVAL);
goto error0;
}
if (VN_CACHED(VFS_I(tip)) != 0) {
xfs_inval_cached_trace(tip, 0, -1, 0, -1);
error = xfs_flushinval_pages(tip, 0, -1,
FI_REMAPF_LOCKED);
if (error)
goto error0;
}
/* Verify O_DIRECT for ftmp */
if (VN_CACHED(VFS_I(tip)) != 0) {
error = XFS_ERROR(EINVAL);
goto error0;
}
/* Verify all data are being swapped */
if (sxp->sx_offset != 0 ||
sxp->sx_length != ip->i_d.di_size ||
sxp->sx_length != tip->i_d.di_size) {
error = XFS_ERROR(EFAULT);
goto error0;
}
/* check inode formats now that data is flushed */
error = xfs_swap_extents_check_format(ip, tip);
if (error) {
xfs_fs_cmn_err(CE_NOTE, mp,
"%s: inode 0x%llx format is incompatible for exchanging.",
__FILE__, ip->i_ino);
goto error0;
}
/*
* Compare the current change & modify times with that
* passed in. If they differ, we abort this swap.
* This is the mechanism used to ensure the calling
* process that the file was not changed out from
* under it.
*/
if ((sbp->bs_ctime.tv_sec != ip->i_d.di_ctime.t_sec) ||
(sbp->bs_ctime.tv_nsec != ip->i_d.di_ctime.t_nsec) ||
(sbp->bs_mtime.tv_sec != ip->i_d.di_mtime.t_sec) ||
(sbp->bs_mtime.tv_nsec != ip->i_d.di_mtime.t_nsec)) {
error = XFS_ERROR(EBUSY);
goto error0;
}
/* We need to fail if the file is memory mapped. Once we have tossed
* all existing pages, the page fault will have no option
* but to go to the filesystem for pages. By making the page fault call
* vop_read (or write in the case of autogrow) they block on the iolock
* until we have switched the extents.
*/
if (VN_MAPPED(VFS_I(ip))) {
error = XFS_ERROR(EBUSY);
goto error0;
}
xfs_iunlock(ip, XFS_ILOCK_EXCL);
xfs_iunlock(tip, XFS_ILOCK_EXCL);
/*
* There is a race condition here since we gave up the
* ilock. However, the data fork will not change since
* we have the iolock (locked for truncation too) so we
* are safe. We don't really care if non-io related
* fields change.
*/
xfs_tosspages(ip, 0, -1, FI_REMAPF);
tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT);
if ((error = xfs_trans_reserve(tp, 0,
XFS_ICHANGE_LOG_RES(mp), 0,
0, 0))) {
xfs_iunlock(ip, XFS_IOLOCK_EXCL);
xfs_iunlock(tip, XFS_IOLOCK_EXCL);
xfs_trans_cancel(tp, 0);
locked = 0;
goto error0;
}
xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL);
/*
* Count the number of extended attribute blocks
*/
if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) &&
(ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks);
if (error) {
xfs_trans_cancel(tp, 0);
goto error0;
}
}
if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) &&
(tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK,
&taforkblks);
if (error) {
xfs_trans_cancel(tp, 0);
goto error0;
}
}
/*
* Swap the data forks of the inodes
*/
ifp = &ip->i_df;
tifp = &tip->i_df;
*tempifp = *ifp; /* struct copy */
*ifp = *tifp; /* struct copy */
*tifp = *tempifp; /* struct copy */
/*
* Fix the in-memory data fork values that are dependent on the fork
* offset in the inode. We can't assume they remain the same as attr2
* has dynamic fork offsets.
*/
ifp->if_ext_max = XFS_IFORK_SIZE(ip, XFS_DATA_FORK) /
(uint)sizeof(xfs_bmbt_rec_t);
tifp->if_ext_max = XFS_IFORK_SIZE(tip, XFS_DATA_FORK) /
(uint)sizeof(xfs_bmbt_rec_t);
/*
* Fix the on-disk inode values
*/
tmp = (__uint64_t)ip->i_d.di_nblocks;
ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks;
tip->i_d.di_nblocks = tmp + taforkblks - aforkblks;
tmp = (__uint64_t) ip->i_d.di_nextents;
ip->i_d.di_nextents = tip->i_d.di_nextents;
tip->i_d.di_nextents = tmp;
tmp = (__uint64_t) ip->i_d.di_format;
ip->i_d.di_format = tip->i_d.di_format;
tip->i_d.di_format = tmp;
ilf_fields = XFS_ILOG_CORE;
switch(ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) {
ifp->if_u1.if_extents =
ifp->if_u2.if_inline_ext;
}
ilf_fields |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
ilf_fields |= XFS_ILOG_DBROOT;
break;
}
tilf_fields = XFS_ILOG_CORE;
switch(tip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
/* If the extents fit in the inode, fix the
* pointer. Otherwise it's already NULL or
* pointing to the extent.
*/
if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) {
tifp->if_u1.if_extents =
tifp->if_u2.if_inline_ext;
}
tilf_fields |= XFS_ILOG_DEXT;
break;
case XFS_DINODE_FMT_BTREE:
tilf_fields |= XFS_ILOG_DBROOT;
break;
}
IHOLD(ip);
xfs_trans_ijoin(tp, ip, lock_flags);
IHOLD(tip);
xfs_trans_ijoin(tp, tip, lock_flags);
xfs_trans_log_inode(tp, ip, ilf_fields);
xfs_trans_log_inode(tp, tip, tilf_fields);
/*
* If this is a synchronous mount, make sure that the
* transaction goes to disk before returning to the user.
*/
if (mp->m_flags & XFS_MOUNT_WSYNC) {
xfs_trans_set_sync(tp);
}
error = xfs_trans_commit(tp, XFS_TRANS_SWAPEXT);
locked = 0;
error0:
if (locked) {
xfs_iunlock(ip, lock_flags);
xfs_iunlock(tip, lock_flags);
}
if (tempifp != NULL)
kmem_free(tempifp);
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;
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(
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;
bhv_vnode_t *vp = XFS_ITOV(xip);
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;
bhv_vrwlock_t locktype;
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;
locktype = VRWLOCK_WRITE_DIRECT;
need_i_mutex = 0;
} else {
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
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, vp,
pos, count,
dmflags, &locktype);
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 && (VN_CACHED(vp) || pos > xip->i_size)) {
xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
iolock = XFS_IOLOCK_EXCL;
locktype = VRWLOCK_WRITE;
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))) {
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 > 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 = -remove_suid(file->f_path.dentry);
if (unlikely(error)) {
goto out_unlock_internal;
}
}
retry:
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
if ((ioflags & IO_ISDIRECT)) {
if (VN_CACHED(vp)) {
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;
locktype = VRWLOCK_WRITE_DIRECT;
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);
if (ret == -ENOSPC &&
DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
xfs_rwunlock(xip, 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 (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_rwlock(xip, locktype);
if (error)
goto out_unlock_internal;
pos = xip->i_size;
ret = 0;
goto retry;
}
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);
}
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_rwunlock(xip, locktype);
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_rwlock(xip, locktype);
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_rwunlock(xip, locktype);
out_unlock_mutex:
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
return -error;
}
ssize_t /* bytes read, or (-) error */
xfs_read(
xfs_inode_t *ip,
struct kiocb *iocb,
const struct iovec *iovp,
unsigned int segs,
loff_t *offset,
int ioflags)
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
bhv_vnode_t *vp = XFS_ITOV(ip);
xfs_mount_t *mp = ip->i_mount;
size_t size = 0;
ssize_t ret = 0;
xfs_fsize_t n;
unsigned long seg;
XFS_STATS_INC(xs_read_calls);
/* 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 (unlikely(ioflags & IO_ISDIRECT)) {
xfs_buftarg_t *target =
XFS_IS_REALTIME_INODE(ip) ?
mp->m_rtdev_targp : mp->m_ddev_targp;
if ((*offset & target->bt_smask) ||
(size & target->bt_smask)) {
if (*offset == ip->i_size) {
return (0);
}
return -XFS_ERROR(EINVAL);
}
}
n = XFS_MAXIOFFSET(mp) - *offset;
if ((n <= 0) || (size == 0))
return 0;
if (n < size)
size = n;
if (XFS_FORCED_SHUTDOWN(mp))
return -EIO;
if (unlikely(ioflags & IO_ISDIRECT))
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, XFS_IOLOCK_SHARED);
if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
bhv_vrwlock_t locktype = VRWLOCK_READ;
int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *offset, size,
dmflags, &locktype);
if (ret) {
xfs_iunlock(ip, XFS_IOLOCK_SHARED);
if (unlikely(ioflags & IO_ISDIRECT))
mutex_unlock(&inode->i_mutex);
return ret;
}
}
if (unlikely(ioflags & IO_ISDIRECT)) {
if (VN_CACHED(vp))
ret = xfs_flushinval_pages(ip, (*offset & PAGE_CACHE_MASK),
-1, FI_REMAPF_LOCKED);
mutex_unlock(&inode->i_mutex);
if (ret) {
xfs_iunlock(ip, XFS_IOLOCK_SHARED);
return ret;
}
}
xfs_rw_enter_trace(XFS_READ_ENTER, ip,
(void *)iovp, segs, *offset, ioflags);
iocb->ki_pos = *offset;
ret = generic_file_aio_read(iocb, iovp, segs, *offset);
if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
ret = wait_on_sync_kiocb(iocb);
if (ret > 0)
XFS_STATS_ADD(xs_read_bytes, ret);
xfs_iunlock(ip, XFS_IOLOCK_SHARED);
return ret;
}
