static int ocfs2_get_inode_attr(struct inode *inode, unsigned *flags)
{
int status;
status = ocfs2_inode_lock(inode, NULL, 0);
if (status < 0) {
mlog_errno(status);
return status;
}
ocfs2_get_inode_flags(OCFS2_I(inode));
*flags = OCFS2_I(inode)->ip_attr;
ocfs2_inode_unlock(inode, 0);
mlog_exit(status);
return status;
}
static char *ocfs2_fast_symlink_getlink(struct inode *inode,
struct buffer_head **bh)
{
int status;
char *link = NULL;
struct ocfs2_dinode *fe;
mlog_entry_void();
status = ocfs2_read_block(OCFS2_SB(inode->i_sb),
OCFS2_I(inode)->ip_blkno,
bh,
OCFS2_BH_CACHED,
inode);
if (status < 0) {
mlog_errno(status);
link = ERR_PTR(status);
goto bail;
}
fe = (struct ocfs2_dinode *) (*bh)->b_data;
link = (char *) fe->id2.i_symlink;
bail:
mlog_exit(status);
return link;
}
static struct dentry *ocfs2_get_parent(struct dentry *child)
{
int status;
u64 blkno;
struct dentry *parent;
struct inode *dir = child->d_inode;
trace_ocfs2_get_parent(child, child->d_name.len, child->d_name.name,
(unsigned long long)OCFS2_I(dir)->ip_blkno);
status = ocfs2_inode_lock(dir, NULL, 0);
if (status < 0) {
if (status != -ENOENT)
mlog_errno(status);
parent = ERR_PTR(status);
goto bail;
}
status = ocfs2_lookup_ino_from_name(dir, "..", 2, &blkno);
if (status < 0) {
parent = ERR_PTR(-ENOENT);
goto bail_unlock;
}
parent = d_obtain_alias(ocfs2_iget(OCFS2_SB(dir->i_sb), blkno, 0, 0));
bail_unlock:
ocfs2_inode_unlock(dir, 0);
bail:
trace_ocfs2_get_parent_end(parent);
return parent;
}
/* Called from locking and called from ocfs2_clear_inode. Dump the
* cache for a given inode.
*
* This function is a few more lines longer than necessary due to some
* accounting done here, but I think it's worth tracking down those
* bugs sooner -- Mark */
void ocfs2_metadata_cache_purge(struct inode *inode)
{
struct ocfs2_inode_info *oi = OCFS2_I(inode);
unsigned int tree, to_purge, purged;
struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
struct rb_root root = RB_ROOT;
spin_lock(&oi->ip_lock);
tree = !(oi->ip_flags & OCFS2_INODE_CACHE_INLINE);
to_purge = ci->ci_num_cached;
mlog(0, "Purge %u %s items from Inode %llu\n", to_purge,
tree ? "array" : "tree", (unsigned long long)oi->ip_blkno);
/* If we're a tree, save off the root so that we can safely
* initialize the cache. We do the work to free tree members
* without the spinlock. */
if (tree)
root = ci->ci_cache.ci_tree;
ocfs2_metadata_cache_init(inode);
spin_unlock(&oi->ip_lock);
purged = ocfs2_purge_copied_metadata_tree(&root);
/* If possible, track the number wiped so that we can more
* easily detect counting errors. Unfortunately, this is only
* meaningful for trees. */
if (tree && purged != to_purge)
mlog(ML_ERROR, "Inode %llu, count = %u, purged = %u\n",
(unsigned long long)oi->ip_blkno, to_purge, purged);
}
static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
struct page *page = vmf->page;
struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
struct buffer_head *di_bh = NULL;
sigset_t blocked, oldset;
int ret, ret2;
ret = ocfs2_vm_op_block_sigs(&blocked, &oldset);
if (ret < 0) {
mlog_errno(ret);
return ret;
}
/*
* The cluster locks taken will block a truncate from another
* node. Taking the data lock will also ensure that we don't
* attempt page truncation as part of a downconvert.
*/
ret = ocfs2_inode_lock(inode, &di_bh, 1);
if (ret < 0) {
mlog_errno(ret);
goto out;
}
/*
* The alloc sem should be enough to serialize with
* ocfs2_truncate_file() changing i_size as well as any thread
* modifying the inode btree.
*/
down_write(&OCFS2_I(inode)->ip_alloc_sem);
ret = __ocfs2_page_mkwrite(inode, di_bh, page);
up_write(&OCFS2_I(inode)->ip_alloc_sem);
brelse(di_bh);
ocfs2_inode_unlock(inode, 1);
out:
ret2 = ocfs2_vm_op_unblock_sigs(&oldset);
if (ret2 < 0)
mlog_errno(ret2);
if (ret)
ret = VM_FAULT_SIGBUS;
return ret;
}
void ocfs2_metadata_cache_init(struct inode *inode)
{
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
oi->ip_flags |= OCFS2_INODE_CACHE_INLINE;
ci->ci_num_cached = 0;
}
static int ocfs2_writepage(struct page *page, struct writeback_control *wbc)
{
trace_ocfs2_writepage(
(unsigned long long)OCFS2_I(page->mapping->host)->ip_blkno,
page->index);
return block_write_full_page(page, ocfs2_get_block, wbc);
}
static struct dentry *ocfs2_get_parent(struct dentry *child)
{
int status;
u64 blkno;
struct dentry *parent;
struct inode *inode;
struct inode *dir = child->d_inode;
struct buffer_head *dirent_bh = NULL;
struct ocfs2_dir_entry *dirent;
mlog_entry("(0x%p, '%.*s')\n", child,
child->d_name.len, child->d_name.name);
mlog(0, "find parent of directory %llu\n",
(unsigned long long)OCFS2_I(dir)->ip_blkno);
status = ocfs2_meta_lock(dir, NULL, 0);
if (status < 0) {
if (status != -ENOENT)
mlog_errno(status);
parent = ERR_PTR(status);
goto bail;
}
status = ocfs2_find_files_on_disk("..", 2, &blkno, dir, &dirent_bh,
&dirent);
if (status < 0) {
parent = ERR_PTR(-ENOENT);
goto bail_unlock;
}
inode = ocfs2_iget(OCFS2_SB(dir->i_sb), blkno, 0);
if (IS_ERR(inode)) {
mlog(ML_ERROR, "Unable to create inode %llu\n",
(unsigned long long)blkno);
parent = ERR_PTR(-EACCES);
goto bail_unlock;
}
parent = d_alloc_anon(inode);
if (!parent) {
iput(inode);
parent = ERR_PTR(-ENOMEM);
}
parent->d_op = &ocfs2_dentry_ops;
bail_unlock:
ocfs2_meta_unlock(dir, 0);
if (dirent_bh)
brelse(dirent_bh);
bail:
mlog_exit_ptr(parent);
return parent;
}
static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
{
sector_t status;
u64 p_blkno = 0;
int err = 0;
struct inode *inode = mapping->host;
mlog_entry("(block = %llu)\n", (unsigned long long)block);
/* We don't need to lock journal system files, since they aren't
* accessed concurrently from multiple nodes.
*/
if (!INODE_JOURNAL(inode)) {
err = ocfs2_inode_lock(inode, NULL, 0);
if (err) {
if (err != -ENOENT)
mlog_errno(err);
goto bail;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
}
if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
if (!INODE_JOURNAL(inode)) {
up_read(&OCFS2_I(inode)->ip_alloc_sem);
ocfs2_inode_unlock(inode, 0);
}
if (err) {
mlog(ML_ERROR, "get_blocks() failed, block = %llu\n",
(unsigned long long)block);
mlog_errno(err);
goto bail;
}
bail:
status = err ? 0 : p_blkno;
mlog_exit((int)status);
return status;
}
int ocfs2_extent_map_init(struct inode *inode)
{
struct ocfs2_extent_map *em = &OCFS2_I(inode)->ip_map;
em->em_extents = RB_ROOT;
em->em_clusters = 0;
return 0;
}
static int ocfs2_block_group_fill(handle_t *handle,
struct inode *alloc_inode,
struct buffer_head *bg_bh,
u64 group_blkno,
u16 my_chain,
struct ocfs2_chain_list *cl)
{
int status = 0;
struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
struct super_block * sb = alloc_inode->i_sb;
mlog_entry_void();
if (((unsigned long long) bg_bh->b_blocknr) != group_blkno) {
ocfs2_error(alloc_inode->i_sb, "group block (%llu) != "
"b_blocknr (%llu)",
(unsigned long long)group_blkno,
(unsigned long long) bg_bh->b_blocknr);
status = -EIO;
goto bail;
}
status = ocfs2_journal_access(handle,
alloc_inode,
bg_bh,
OCFS2_JOURNAL_ACCESS_CREATE);
if (status < 0) {
mlog_errno(status);
goto bail;
}
memset(bg, 0, sb->s_blocksize);
strcpy(bg->bg_signature, OCFS2_GROUP_DESC_SIGNATURE);
bg->bg_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
bg->bg_size = cpu_to_le16(ocfs2_group_bitmap_size(sb));
bg->bg_bits = cpu_to_le16(ocfs2_bits_per_group(cl));
bg->bg_chain = cpu_to_le16(my_chain);
bg->bg_next_group = cl->cl_recs[my_chain].c_blkno;
bg->bg_parent_dinode = cpu_to_le64(OCFS2_I(alloc_inode)->ip_blkno);
bg->bg_blkno = cpu_to_le64(group_blkno);
/* set the 1st bit in the bitmap to account for the descriptor block */
ocfs2_set_bit(0, (unsigned long *)bg->bg_bitmap);
bg->bg_free_bits_count = cpu_to_le16(le16_to_cpu(bg->bg_bits) - 1);
status = ocfs2_journal_dirty(handle, bg_bh);
if (status < 0)
mlog_errno(status);
/* There is no need to zero out or otherwise initialize the
* other blocks in a group - All valid FS metadata in a block
* group stores the superblock fs_generation value at
* allocation time. */
bail:
mlog_exit(status);
return status;
}
/* Support function for ocfs2_delete_inode. Will help us keep the
* inode data in a consistent state for clear_inode. Always truncates
* pages, optionally sync's them first. */
static void ocfs2_cleanup_delete_inode(struct inode *inode,
int sync_data)
{
mlog(0, "Cleanup inode %llu, sync = %d\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
if (sync_data)
write_inode_now(inode, 1);
truncate_inode_pages(&inode->i_data, 0);
}
/* Support function for ocfs2_delete_inode. Will help us keep the
* inode data in a consistent state for clear_inode. Always truncates
* pages, optionally sync's them first. */
static void ocfs2_cleanup_delete_inode(struct inode *inode,
int sync_data)
{
trace_ocfs2_cleanup_delete_inode(
(unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
if (sync_data)
filemap_write_and_wait(inode->i_mapping);
truncate_inode_pages(&inode->i_data, 0);
}
/*
* TODO: this should probably be merged into ocfs2_get_block
*
* However, you now need to pay attention to the cont_prepare_write()
* stuff in ocfs2_get_block (that is, ocfs2_get_block pretty much
* expects never to extend).
*/
struct buffer_head *ocfs2_bread(struct inode *inode,
int block, int *err, int reada)
{
struct buffer_head *bh = NULL;
int tmperr;
u64 p_blkno;
int readflags = OCFS2_BH_CACHED;
if (reada)
readflags |= OCFS2_BH_READAHEAD;
if (((u64)block << inode->i_sb->s_blocksize_bits) >=
i_size_read(inode)) {
BUG_ON(!reada);
return NULL;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
tmperr = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
up_read(&OCFS2_I(inode)->ip_alloc_sem);
if (tmperr < 0) {
mlog_errno(tmperr);
goto fail;
}
tmperr = ocfs2_read_block(OCFS2_SB(inode->i_sb), p_blkno, &bh,
readflags, inode);
if (tmperr < 0)
goto fail;
tmperr = 0;
*err = 0;
return bh;
fail:
if (bh) {
brelse(bh);
bh = NULL;
}
*err = -EIO;
return NULL;
}
static struct inode * _ocfs2_get_system_file_inode(struct ocfs2_super *osb,
int type,
u32 slot)
{
char namebuf[40];
struct inode *inode = NULL;
u64 blkno;
int status = 0;
ocfs2_sprintf_system_inode_name(namebuf,
sizeof(namebuf),
type, slot);
status = ocfs2_lookup_ino_from_name(osb->sys_root_inode, namebuf,
strlen(namebuf), &blkno);
if (status < 0) {
goto bail;
}
inode = ocfs2_iget(osb, blkno, OCFS2_FI_FLAG_SYSFILE, type);
if (IS_ERR(inode)) {
mlog_errno(PTR_ERR(inode));
inode = NULL;
goto bail;
}
#ifdef CONFIG_DEBUG_LOCK_ALLOC
if (type == LOCAL_USER_QUOTA_SYSTEM_INODE ||
type == LOCAL_GROUP_QUOTA_SYSTEM_INODE ||
type == JOURNAL_SYSTEM_INODE) {
/* Ignore inode lock on these inodes as the lock does not
* really belong to any process and lockdep cannot handle
* that */
OCFS2_I(inode)->ip_inode_lockres.l_lockdep_map.key = NULL;
} else {
lockdep_init_map(&OCFS2_I(inode)->ip_inode_lockres.
l_lockdep_map,
ocfs2_system_inodes[type].si_name,
&ocfs2_sysfile_cluster_lock_key[type], 0);
}
#endif
bail:
return inode;
}
void ocfs2_evict_inode(struct inode *inode)
{
if (!inode->i_nlink ||
(OCFS2_I(inode)->ip_flags & OCFS2_INODE_MAYBE_ORPHANED)) {
ocfs2_delete_inode(inode);
} else {
truncate_inode_pages(&inode->i_data, 0);
}
ocfs2_clear_inode(inode);
}
static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
{
sector_t status;
u64 p_blkno = 0;
int err = 0;
struct inode *inode = mapping->host;
trace_ocfs2_bmap((unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)block);
if (!INODE_JOURNAL(inode)) {
err = ocfs2_inode_lock(inode, NULL, 0);
if (err) {
if (err != -ENOENT)
mlog_errno(err);
goto bail;
}
down_read(&OCFS2_I(inode)->ip_alloc_sem);
}
if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL))
err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
NULL);
if (!INODE_JOURNAL(inode)) {
up_read(&OCFS2_I(inode)->ip_alloc_sem);
ocfs2_inode_unlock(inode, 0);
}
if (err) {
mlog(ML_ERROR, "get_blocks() failed, block = %llu\n",
(unsigned long long)block);
mlog_errno(err);
goto bail;
}
bail:
status = err ? 0 : p_blkno;
return status;
}
/*
* initialize the new inode, but don't do anything that would cause
* us to sleep.
* return 0 on success, 1 on failure
*/
static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
{
struct ocfs2_find_inode_args *args = opaque;
mlog_entry("inode = %p, opaque = %p\n", inode, opaque);
inode->i_ino = args->fi_ino;
OCFS2_I(inode)->ip_blkno = args->fi_blkno;
mlog_exit(0);
return 0;
}
struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
int sysfile_type)
{
struct inode *inode = NULL;
struct super_block *sb = osb->sb;
struct ocfs2_find_inode_args args;
trace_ocfs2_iget_begin((unsigned long long)blkno, flags,
sysfile_type);
/* Ok. By now we've either got the offsets passed to us by the
* caller, or we just pulled them off the bh. Lets do some
* sanity checks to make sure they're OK. */
if (blkno == 0) {
inode = ERR_PTR(-EINVAL);
mlog_errno(PTR_ERR(inode));
goto bail;
}
args.fi_blkno = blkno;
args.fi_flags = flags;
args.fi_ino = ino_from_blkno(sb, blkno);
args.fi_sysfile_type = sysfile_type;
inode = iget5_locked(sb, args.fi_ino, ocfs2_find_actor,
ocfs2_init_locked_inode, &args);
/* inode was *not* in the inode cache. 2.6.x requires
* us to do our own read_inode call and unlock it
* afterwards. */
if (inode == NULL) {
inode = ERR_PTR(-ENOMEM);
mlog_errno(PTR_ERR(inode));
goto bail;
}
trace_ocfs2_iget5_locked(inode->i_state);
if (inode->i_state & I_NEW) {
ocfs2_read_locked_inode(inode, &args);
unlock_new_inode(inode);
}
if (is_bad_inode(inode)) {
iput(inode);
inode = ERR_PTR(-ESTALE);
goto bail;
}
bail:
if (!IS_ERR(inode)) {
trace_ocfs2_iget_end(inode,
(unsigned long long)OCFS2_I(inode)->ip_blkno);
}
return inode;
}
/*
* This is called from our getattr.
*/
int ocfs2_inode_revalidate(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
int status = 0;
trace_ocfs2_inode_revalidate(inode,
inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL,
inode ? (unsigned long long)OCFS2_I(inode)->ip_flags : 0);
if (!inode) {
status = -ENOENT;
goto bail;
}
spin_lock(&OCFS2_I(inode)->ip_lock);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_DELETED) {
spin_unlock(&OCFS2_I(inode)->ip_lock);
status = -ENOENT;
goto bail;
}
spin_unlock(&OCFS2_I(inode)->ip_lock);
/* Let ocfs2_inode_lock do the work of updating our struct
* inode for us. */
status = ocfs2_inode_lock(inode, NULL, 0);
if (status < 0) {
if (status != -ENOENT)
mlog_errno(status);
goto bail;
}
ocfs2_inode_unlock(inode, 0);
bail:
return status;
}
/*
* Remove all entries past new_clusters and also clip any extent
* straddling new_clusters, if there is one. This does not check
* or modify ip_clusters
*/
int ocfs2_extent_map_trunc(struct inode *inode, u32 new_clusters)
{
struct rb_node *free_head = NULL;
struct ocfs2_extent_map_entry *ent = NULL;
spin_lock(&OCFS2_I(inode)->ip_lock);
__ocfs2_extent_map_drop(inode, new_clusters, &free_head, &ent);
if (ent)
ent->e_rec.e_clusters = cpu_to_le32(new_clusters -
le32_to_cpu(ent->e_rec.e_cpos));
OCFS2_I(inode)->ip_map.em_clusters = new_clusters;
spin_unlock(&OCFS2_I(inode)->ip_lock);
if (free_head)
__ocfs2_extent_map_drop_cleanup(free_head);
return 0;
}
static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
{
sigset_t oldset;
int ret;
ocfs2_block_signals(&oldset);
ret = filemap_fault(area, vmf);
ocfs2_unblock_signals(&oldset);
trace_ocfs2_fault(OCFS2_I(area->vm_file->f_mapping->host)->ip_blkno,
area, vmf->page, vmf->pgoff);
return ret;
}
struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, int flags)
{
struct inode *inode = NULL;
struct super_block *sb = osb->sb;
struct ocfs2_find_inode_args args;
mlog_entry("(blkno = %llu)\n", (unsigned long long)blkno);
/* Ok. By now we've either got the offsets passed to us by the
* caller, or we just pulled them off the bh. Lets do some
* sanity checks to make sure they're OK. */
if (blkno == 0) {
inode = ERR_PTR(-EINVAL);
mlog_errno(PTR_ERR(inode));
goto bail;
}
args.fi_blkno = blkno;
args.fi_flags = flags;
args.fi_ino = ino_from_blkno(sb, blkno);
inode = iget5_locked(sb, args.fi_ino, ocfs2_find_actor,
ocfs2_init_locked_inode, &args);
/* inode was *not* in the inode cache. 2.6.x requires
* us to do our own read_inode call and unlock it
* afterwards. */
if (inode && inode->i_state & I_NEW) {
mlog(0, "Inode was not in inode cache, reading it.\n");
ocfs2_read_locked_inode(inode, &args);
unlock_new_inode(inode);
}
if (inode == NULL) {
inode = ERR_PTR(-ENOMEM);
mlog_errno(PTR_ERR(inode));
goto bail;
}
if (is_bad_inode(inode)) {
iput(inode);
inode = ERR_PTR(-ESTALE);
goto bail;
}
bail:
if (!IS_ERR(inode)) {
mlog(0, "returning inode with number %llu\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
mlog_exit_ptr(inode);
}
return inode;
}
/* There is a series of simple checks that should be done before a
* vote is even considered. Encapsulate those in this function. */
static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
{
int ret = 0;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
/* We shouldn't be getting here for the root directory
* inode.. */
if (inode == osb->root_inode) {
mlog(ML_ERROR, "Skipping delete of root inode.\n");
goto bail;
}
/* If we're coming from process_vote we can't go into our own
* voting [hello, deadlock city!], so unforuntately we just
* have to skip deleting this guy. That's OK though because
* the node who's doing the actual deleting should handle it
* anyway. */
if (current == osb->vote_task) {
mlog(0, "Skipping delete of %lu because we're currently "
"in process_vote\n", inode->i_ino);
goto bail;
}
spin_lock(&oi->ip_lock);
/* OCFS2 *never* deletes system files. This should technically
* never get here as system file inodes should always have a
* positive link count. */
if (oi->ip_flags & OCFS2_INODE_SYSTEM_FILE) {
mlog(ML_ERROR, "Skipping delete of system file %llu\n",
(unsigned long long)oi->ip_blkno);
goto bail_unlock;
}
/* If we have voted "yes" on the wipe of this inode for
* another node, it will be marked here so we can safely skip
* it. Recovery will cleanup any inodes we might inadvertantly
* skip here. */
if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE) {
mlog(0, "Skipping delete of %lu because another node "
"has done this for us.\n", inode->i_ino);
goto bail_unlock;
}
ret = 1;
bail_unlock:
spin_unlock(&oi->ip_lock);
bail:
return ret;
}
static int ocfs2_write_remove_suid(struct inode *inode)
{
int ret;
struct buffer_head *bh = NULL;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
handle_t *handle;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di;
mlog_entry("(Inode %llu, mode 0%o)\n",
(unsigned long long)oi->ip_blkno, inode->i_mode);
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (handle == NULL) {
ret = -ENOMEM;
mlog_errno(ret);
goto out;
}
ret = ocfs2_read_block(osb, oi->ip_blkno, &bh, OCFS2_BH_CACHED, inode);
if (ret < 0) {
mlog_errno(ret);
goto out_trans;
}
ret = ocfs2_journal_access(handle, inode, bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (ret < 0) {
mlog_errno(ret);
goto out_bh;
}
inode->i_mode &= ~S_ISUID;
if ((inode->i_mode & S_ISGID) && (inode->i_mode & S_IXGRP))
inode->i_mode &= ~S_ISGID;
di = (struct ocfs2_dinode *) bh->b_data;
di->i_mode = cpu_to_le16(inode->i_mode);
ret = ocfs2_journal_dirty(handle, bh);
if (ret < 0)
mlog_errno(ret);
out_bh:
brelse(bh);
out_trans:
ocfs2_commit_trans(osb, handle);
out:
mlog_exit(ret);
return ret;
}
/*
* initialize the new inode, but don't do anything that would cause
* us to sleep.
* return 0 on success, 1 on failure
*/
static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
{
struct ocfs2_find_inode_args *args = opaque;
static struct lock_class_key ocfs2_quota_ip_alloc_sem_key,
ocfs2_file_ip_alloc_sem_key;
inode->i_ino = args->fi_ino;
OCFS2_I(inode)->ip_blkno = args->fi_blkno;
if (args->fi_sysfile_type != 0)
lockdep_set_class(&inode->i_mutex,
&ocfs2_sysfile_lock_key[args->fi_sysfile_type]);
if (args->fi_sysfile_type == USER_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == GROUP_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == LOCAL_USER_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == LOCAL_GROUP_QUOTA_SYSTEM_INODE)
lockdep_set_class(&OCFS2_I(inode)->ip_alloc_sem,
&ocfs2_quota_ip_alloc_sem_key);
else
lockdep_set_class(&OCFS2_I(inode)->ip_alloc_sem,
&ocfs2_file_ip_alloc_sem_key);
return 0;
}
static int ocfs2_readpage_inline(struct inode *inode, struct page *page)
{
int ret;
struct buffer_head *di_bh = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
BUG_ON(!PageLocked(page));
BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL));
ret = ocfs2_read_block(osb, OCFS2_I(inode)->ip_blkno, &di_bh,
OCFS2_BH_CACHED, inode);
if (ret) {
mlog_errno(ret);
goto out;
}
ret = ocfs2_read_inline_data(inode, page, di_bh);
out:
unlock_page(page);
brelse(di_bh);
return ret;
}
/*
* Remove all entries past new_clusters, inclusive of an entry that
* contains new_clusters. This is effectively a cache forget.
*
* If you want to also clip the last extent by some number of clusters,
* you need to call ocfs2_extent_map_trunc().
* This code does not check or modify ip_clusters.
*/
int ocfs2_extent_map_drop(struct inode *inode, u32 new_clusters)
{
struct rb_node *free_head = NULL;
struct ocfs2_extent_map *em = &OCFS2_I(inode)->ip_map;
struct ocfs2_extent_map_entry *ent;
spin_lock(&OCFS2_I(inode)->ip_lock);
__ocfs2_extent_map_drop(inode, new_clusters, &free_head, &ent);
if (ent) {
rb_erase(&ent->e_node, &em->em_extents);
ent->e_node.rb_right = free_head;
free_head = &ent->e_node;
}
spin_unlock(&OCFS2_I(inode)->ip_lock);
if (free_head)
__ocfs2_extent_map_drop_cleanup(free_head);
return 0;
}
/*
* This is called from our getattr.
*/
int ocfs2_inode_revalidate(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
int status = 0;
mlog_entry("(inode = 0x%p, ino = %llu)\n", inode,
inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL);
if (!inode) {
mlog(0, "eep, no inode!\n");
status = -ENOENT;
goto bail;
}
spin_lock(&OCFS2_I(inode)->ip_lock);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_DELETED) {
spin_unlock(&OCFS2_I(inode)->ip_lock);
mlog(0, "inode deleted!\n");
status = -ENOENT;
goto bail;
}
spin_unlock(&OCFS2_I(inode)->ip_lock);
/* Let ocfs2_meta_lock do the work of updating our struct
* inode for us. */
status = ocfs2_meta_lock(inode, NULL, 0);
if (status < 0) {
if (status != -ENOENT)
mlog_errno(status);
goto bail;
}
ocfs2_meta_unlock(inode, 0);
bail:
mlog_exit(status);
return status;
}
static int __ocfs2_move_extents_range(struct buffer_head *di_bh,
struct ocfs2_move_extents_context *context)
{
int ret = 0, flags, do_defrag, skip = 0;
u32 cpos, phys_cpos, move_start, len_to_move, alloc_size;
u32 len_defraged = 0, defrag_thresh = 0, new_phys_cpos = 0;
struct inode *inode = context->inode;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
struct ocfs2_move_extents *range = context->range;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
if ((inode->i_size == 0) || (range->me_len == 0))
return 0;
if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
return 0;
context->refcount_loc = le64_to_cpu(di->i_refcount_loc);
ocfs2_init_dinode_extent_tree(&context->et, INODE_CACHE(inode), di_bh);
ocfs2_init_dealloc_ctxt(&context->dealloc);
/*
* TO-DO XXX:
*
* - xattr extents.
*/
do_defrag = context->auto_defrag;
/*
* extents moving happens in unit of clusters, for the sake
* of simplicity, we may ignore two clusters where 'byte_start'
* and 'byte_start + len' were within.
*/
move_start = ocfs2_clusters_for_bytes(osb->sb, range->me_start);
len_to_move = (range->me_start + range->me_len) >>
osb->s_clustersize_bits;
if (len_to_move >= move_start)
len_to_move -= move_start;
else
len_to_move = 0;
if (do_defrag) {
defrag_thresh = range->me_threshold >> osb->s_clustersize_bits;
if (defrag_thresh <= 1)
goto done;
} else