xref: /kernel/linux/linux-5.10/fs/ocfs2/namei.c (revision 8c2ecf20)
1// SPDX-License-Identifier: GPL-2.0-or-later
2/* -*- mode: c; c-basic-offset: 8; -*-
3 * vim: noexpandtab sw=8 ts=8 sts=0:
4 *
5 * namei.c
6 *
7 * Create and rename file, directory, symlinks
8 *
9 * Copyright (C) 2002, 2004 Oracle.  All rights reserved.
10 *
11 *  Portions of this code from linux/fs/ext3/dir.c
12 *
13 *  Copyright (C) 1992, 1993, 1994, 1995
14 *  Remy Card (card@masi.ibp.fr)
15 *  Laboratoire MASI - Institut Blaise pascal
16 *  Universite Pierre et Marie Curie (Paris VI)
17 *
18 *   from
19 *
20 *   linux/fs/minix/dir.c
21 *
22 *   Copyright (C) 1991, 1992 Linux Torvalds
23 */
24
25#include <linux/fs.h>
26#include <linux/types.h>
27#include <linux/slab.h>
28#include <linux/highmem.h>
29#include <linux/quotaops.h>
30#include <linux/iversion.h>
31
32#include <cluster/masklog.h>
33
34#include "ocfs2.h"
35
36#include "alloc.h"
37#include "dcache.h"
38#include "dir.h"
39#include "dlmglue.h"
40#include "extent_map.h"
41#include "file.h"
42#include "inode.h"
43#include "journal.h"
44#include "namei.h"
45#include "suballoc.h"
46#include "super.h"
47#include "symlink.h"
48#include "sysfile.h"
49#include "uptodate.h"
50#include "xattr.h"
51#include "acl.h"
52#include "ocfs2_trace.h"
53
54#include "buffer_head_io.h"
55
56static int ocfs2_mknod_locked(struct ocfs2_super *osb,
57			      struct inode *dir,
58			      struct inode *inode,
59			      dev_t dev,
60			      struct buffer_head **new_fe_bh,
61			      struct buffer_head *parent_fe_bh,
62			      handle_t *handle,
63			      struct ocfs2_alloc_context *inode_ac);
64
65static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
66				    struct inode **ret_orphan_dir,
67				    u64 blkno,
68				    char *name,
69				    struct ocfs2_dir_lookup_result *lookup,
70				    bool dio);
71
72static int ocfs2_orphan_add(struct ocfs2_super *osb,
73			    handle_t *handle,
74			    struct inode *inode,
75			    struct buffer_head *fe_bh,
76			    char *name,
77			    struct ocfs2_dir_lookup_result *lookup,
78			    struct inode *orphan_dir_inode,
79			    bool dio);
80
81static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
82				     handle_t *handle,
83				     struct inode *inode,
84				     const char *symname);
85
86static int ocfs2_double_lock(struct ocfs2_super *osb,
87			     struct buffer_head **bh1,
88			     struct inode *inode1,
89			     struct buffer_head **bh2,
90			     struct inode *inode2,
91			     int rename);
92
93static void ocfs2_double_unlock(struct inode *inode1, struct inode *inode2);
94/* An orphan dir name is an 8 byte value, printed as a hex string */
95#define OCFS2_ORPHAN_NAMELEN ((int)(2 * sizeof(u64)))
96
97static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
98				   unsigned int flags)
99{
100	int status;
101	u64 blkno;
102	struct inode *inode = NULL;
103	struct dentry *ret;
104	struct ocfs2_inode_info *oi;
105
106	trace_ocfs2_lookup(dir, dentry, dentry->d_name.len,
107			   dentry->d_name.name,
108			   (unsigned long long)OCFS2_I(dir)->ip_blkno, 0);
109
110	if (dentry->d_name.len > OCFS2_MAX_FILENAME_LEN) {
111		ret = ERR_PTR(-ENAMETOOLONG);
112		goto bail;
113	}
114
115	status = ocfs2_inode_lock_nested(dir, NULL, 0, OI_LS_PARENT);
116	if (status < 0) {
117		if (status != -ENOENT)
118			mlog_errno(status);
119		ret = ERR_PTR(status);
120		goto bail;
121	}
122
123	status = ocfs2_lookup_ino_from_name(dir, dentry->d_name.name,
124					    dentry->d_name.len, &blkno);
125	if (status < 0)
126		goto bail_add;
127
128	inode = ocfs2_iget(OCFS2_SB(dir->i_sb), blkno, 0, 0);
129	if (IS_ERR(inode)) {
130		ret = ERR_PTR(-EACCES);
131		goto bail_unlock;
132	}
133
134	oi = OCFS2_I(inode);
135	/* Clear any orphaned state... If we were able to look up the
136	 * inode from a directory, it certainly can't be orphaned. We
137	 * might have the bad state from a node which intended to
138	 * orphan this inode but crashed before it could commit the
139	 * unlink. */
140	spin_lock(&oi->ip_lock);
141	oi->ip_flags &= ~OCFS2_INODE_MAYBE_ORPHANED;
142	spin_unlock(&oi->ip_lock);
143
144bail_add:
145	ret = d_splice_alias(inode, dentry);
146
147	if (inode) {
148		/*
149		 * If d_splice_alias() finds a DCACHE_DISCONNECTED
150		 * dentry, it will d_move() it on top of ourse. The
151		 * return value will indicate this however, so in
152		 * those cases, we switch them around for the locking
153		 * code.
154		 *
155		 * NOTE: This dentry already has ->d_op set from
156		 * ocfs2_get_parent() and ocfs2_get_dentry()
157		 */
158		if (!IS_ERR_OR_NULL(ret))
159			dentry = ret;
160
161		status = ocfs2_dentry_attach_lock(dentry, inode,
162						  OCFS2_I(dir)->ip_blkno);
163		if (status) {
164			mlog_errno(status);
165			ret = ERR_PTR(status);
166			goto bail_unlock;
167		}
168	} else
169		ocfs2_dentry_attach_gen(dentry);
170
171bail_unlock:
172	/* Don't drop the cluster lock until *after* the d_add --
173	 * unlink on another node will message us to remove that
174	 * dentry under this lock so otherwise we can race this with
175	 * the downconvert thread and have a stale dentry. */
176	ocfs2_inode_unlock(dir, 0);
177
178bail:
179
180	trace_ocfs2_lookup_ret(ret);
181
182	return ret;
183}
184
185static struct inode *ocfs2_get_init_inode(struct inode *dir, umode_t mode)
186{
187	struct inode *inode;
188	int status;
189
190	inode = new_inode(dir->i_sb);
191	if (!inode) {
192		mlog(ML_ERROR, "new_inode failed!\n");
193		return ERR_PTR(-ENOMEM);
194	}
195
196	/* populate as many fields early on as possible - many of
197	 * these are used by the support functions here and in
198	 * callers. */
199	if (S_ISDIR(mode))
200		set_nlink(inode, 2);
201	mode = mode_strip_sgid(dir, mode);
202	inode_init_owner(inode, dir, mode);
203	status = dquot_initialize(inode);
204	if (status)
205		return ERR_PTR(status);
206
207	return inode;
208}
209
210static void ocfs2_cleanup_add_entry_failure(struct ocfs2_super *osb,
211		struct dentry *dentry, struct inode *inode)
212{
213	struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
214
215	ocfs2_simple_drop_lockres(osb, &dl->dl_lockres);
216	ocfs2_lock_res_free(&dl->dl_lockres);
217	BUG_ON(dl->dl_count != 1);
218	spin_lock(&dentry_attach_lock);
219	dentry->d_fsdata = NULL;
220	spin_unlock(&dentry_attach_lock);
221	kfree(dl);
222	iput(inode);
223}
224
225static int ocfs2_mknod(struct inode *dir,
226		       struct dentry *dentry,
227		       umode_t mode,
228		       dev_t dev)
229{
230	int status = 0;
231	struct buffer_head *parent_fe_bh = NULL;
232	handle_t *handle = NULL;
233	struct ocfs2_super *osb;
234	struct ocfs2_dinode *dirfe;
235	struct ocfs2_dinode *fe = NULL;
236	struct buffer_head *new_fe_bh = NULL;
237	struct inode *inode = NULL;
238	struct ocfs2_alloc_context *inode_ac = NULL;
239	struct ocfs2_alloc_context *data_ac = NULL;
240	struct ocfs2_alloc_context *meta_ac = NULL;
241	int want_clusters = 0;
242	int want_meta = 0;
243	int xattr_credits = 0;
244	struct ocfs2_security_xattr_info si = {
245		.name = NULL,
246		.enable = 1,
247	};
248	int did_quota_inode = 0;
249	struct ocfs2_dir_lookup_result lookup = { NULL, };
250	sigset_t oldset;
251	int did_block_signals = 0;
252	struct ocfs2_dentry_lock *dl = NULL;
253
254	trace_ocfs2_mknod(dir, dentry, dentry->d_name.len, dentry->d_name.name,
255			  (unsigned long long)OCFS2_I(dir)->ip_blkno,
256			  (unsigned long)dev, mode);
257
258	status = dquot_initialize(dir);
259	if (status) {
260		mlog_errno(status);
261		return status;
262	}
263
264	/* get our super block */
265	osb = OCFS2_SB(dir->i_sb);
266
267	status = ocfs2_inode_lock(dir, &parent_fe_bh, 1);
268	if (status < 0) {
269		if (status != -ENOENT)
270			mlog_errno(status);
271		return status;
272	}
273
274	if (S_ISDIR(mode) && (dir->i_nlink >= ocfs2_link_max(osb))) {
275		status = -EMLINK;
276		goto leave;
277	}
278
279	dirfe = (struct ocfs2_dinode *) parent_fe_bh->b_data;
280	if (!ocfs2_read_links_count(dirfe)) {
281		/* can't make a file in a deleted directory. */
282		status = -ENOENT;
283		goto leave;
284	}
285
286	status = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
287					   dentry->d_name.len);
288	if (status)
289		goto leave;
290
291	/* get a spot inside the dir. */
292	status = ocfs2_prepare_dir_for_insert(osb, dir, parent_fe_bh,
293					      dentry->d_name.name,
294					      dentry->d_name.len, &lookup);
295	if (status < 0) {
296		mlog_errno(status);
297		goto leave;
298	}
299
300	/* reserve an inode spot */
301	status = ocfs2_reserve_new_inode(osb, &inode_ac);
302	if (status < 0) {
303		if (status != -ENOSPC)
304			mlog_errno(status);
305		goto leave;
306	}
307
308	inode = ocfs2_get_init_inode(dir, mode);
309	if (IS_ERR(inode)) {
310		status = PTR_ERR(inode);
311		inode = NULL;
312		mlog_errno(status);
313		goto leave;
314	}
315
316	/* get security xattr */
317	status = ocfs2_init_security_get(inode, dir, &dentry->d_name, &si);
318	if (status) {
319		if (status == -EOPNOTSUPP)
320			si.enable = 0;
321		else {
322			mlog_errno(status);
323			goto leave;
324		}
325	}
326
327	/* calculate meta data/clusters for setting security and acl xattr */
328	status = ocfs2_calc_xattr_init(dir, parent_fe_bh, mode,
329				       &si, &want_clusters,
330				       &xattr_credits, &want_meta);
331	if (status < 0) {
332		mlog_errno(status);
333		goto leave;
334	}
335
336	/* Reserve a cluster if creating an extent based directory. */
337	if (S_ISDIR(mode) && !ocfs2_supports_inline_data(osb)) {
338		want_clusters += 1;
339
340		/* Dir indexing requires extra space as well */
341		if (ocfs2_supports_indexed_dirs(osb))
342			want_meta++;
343	}
344
345	status = ocfs2_reserve_new_metadata_blocks(osb, want_meta, &meta_ac);
346	if (status < 0) {
347		if (status != -ENOSPC)
348			mlog_errno(status);
349		goto leave;
350	}
351
352	status = ocfs2_reserve_clusters(osb, want_clusters, &data_ac);
353	if (status < 0) {
354		if (status != -ENOSPC)
355			mlog_errno(status);
356		goto leave;
357	}
358
359	handle = ocfs2_start_trans(osb, ocfs2_mknod_credits(osb->sb,
360							    S_ISDIR(mode),
361							    xattr_credits));
362	if (IS_ERR(handle)) {
363		status = PTR_ERR(handle);
364		handle = NULL;
365		mlog_errno(status);
366		goto leave;
367	}
368
369	/* Starting to change things, restart is no longer possible. */
370	ocfs2_block_signals(&oldset);
371	did_block_signals = 1;
372
373	status = dquot_alloc_inode(inode);
374	if (status)
375		goto leave;
376	did_quota_inode = 1;
377
378	/* do the real work now. */
379	status = ocfs2_mknod_locked(osb, dir, inode, dev,
380				    &new_fe_bh, parent_fe_bh, handle,
381				    inode_ac);
382	if (status < 0) {
383		mlog_errno(status);
384		goto leave;
385	}
386
387	fe = (struct ocfs2_dinode *) new_fe_bh->b_data;
388	if (S_ISDIR(mode)) {
389		status = ocfs2_fill_new_dir(osb, handle, dir, inode,
390					    new_fe_bh, data_ac, meta_ac);
391		if (status < 0) {
392			mlog_errno(status);
393			goto leave;
394		}
395
396		status = ocfs2_journal_access_di(handle, INODE_CACHE(dir),
397						 parent_fe_bh,
398						 OCFS2_JOURNAL_ACCESS_WRITE);
399		if (status < 0) {
400			mlog_errno(status);
401			goto leave;
402		}
403		ocfs2_add_links_count(dirfe, 1);
404		ocfs2_journal_dirty(handle, parent_fe_bh);
405		inc_nlink(dir);
406	}
407
408	status = ocfs2_init_acl(handle, inode, dir, new_fe_bh, parent_fe_bh,
409			 meta_ac, data_ac);
410
411	if (status < 0) {
412		mlog_errno(status);
413		goto roll_back;
414	}
415
416	if (si.enable) {
417		status = ocfs2_init_security_set(handle, inode, new_fe_bh, &si,
418						 meta_ac, data_ac);
419		if (status < 0) {
420			mlog_errno(status);
421			goto roll_back;
422		}
423	}
424
425	/*
426	 * Do this before adding the entry to the directory. We add
427	 * also set d_op after success so that ->d_iput() will cleanup
428	 * the dentry lock even if ocfs2_add_entry() fails below.
429	 */
430	status = ocfs2_dentry_attach_lock(dentry, inode,
431					  OCFS2_I(dir)->ip_blkno);
432	if (status) {
433		mlog_errno(status);
434		goto roll_back;
435	}
436
437	dl = dentry->d_fsdata;
438
439	status = ocfs2_add_entry(handle, dentry, inode,
440				 OCFS2_I(inode)->ip_blkno, parent_fe_bh,
441				 &lookup);
442	if (status < 0) {
443		mlog_errno(status);
444		goto roll_back;
445	}
446
447	insert_inode_hash(inode);
448	d_instantiate(dentry, inode);
449	status = 0;
450
451roll_back:
452	if (status < 0 && S_ISDIR(mode)) {
453		ocfs2_add_links_count(dirfe, -1);
454		drop_nlink(dir);
455	}
456
457leave:
458	if (status < 0 && did_quota_inode)
459		dquot_free_inode(inode);
460	if (handle) {
461		if (status < 0 && fe)
462			ocfs2_set_links_count(fe, 0);
463		ocfs2_commit_trans(osb, handle);
464	}
465
466	ocfs2_inode_unlock(dir, 1);
467	if (did_block_signals)
468		ocfs2_unblock_signals(&oldset);
469
470	brelse(new_fe_bh);
471	brelse(parent_fe_bh);
472	kfree(si.value);
473
474	ocfs2_free_dir_lookup_result(&lookup);
475
476	if (inode_ac)
477		ocfs2_free_alloc_context(inode_ac);
478
479	if (data_ac)
480		ocfs2_free_alloc_context(data_ac);
481
482	if (meta_ac)
483		ocfs2_free_alloc_context(meta_ac);
484
485	/*
486	 * We should call iput after the i_mutex of the bitmap been
487	 * unlocked in ocfs2_free_alloc_context, or the
488	 * ocfs2_delete_inode will mutex_lock again.
489	 */
490	if ((status < 0) && inode) {
491		if (dl)
492			ocfs2_cleanup_add_entry_failure(osb, dentry, inode);
493
494		OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR;
495		clear_nlink(inode);
496		iput(inode);
497	}
498
499	if (status)
500		mlog_errno(status);
501
502	return status;
503}
504
505static int __ocfs2_mknod_locked(struct inode *dir,
506				struct inode *inode,
507				dev_t dev,
508				struct buffer_head **new_fe_bh,
509				struct buffer_head *parent_fe_bh,
510				handle_t *handle,
511				struct ocfs2_alloc_context *inode_ac,
512				u64 fe_blkno, u64 suballoc_loc, u16 suballoc_bit)
513{
514	int status = 0;
515	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
516	struct ocfs2_dinode *fe = NULL;
517	struct ocfs2_extent_list *fel;
518	u16 feat;
519	struct ocfs2_inode_info *oi = OCFS2_I(inode);
520	struct timespec64 ts;
521
522	*new_fe_bh = NULL;
523
524	/* populate as many fields early on as possible - many of
525	 * these are used by the support functions here and in
526	 * callers. */
527	inode->i_ino = ino_from_blkno(osb->sb, fe_blkno);
528	oi->ip_blkno = fe_blkno;
529	spin_lock(&osb->osb_lock);
530	inode->i_generation = osb->s_next_generation++;
531	spin_unlock(&osb->osb_lock);
532
533	*new_fe_bh = sb_getblk(osb->sb, fe_blkno);
534	if (!*new_fe_bh) {
535		status = -ENOMEM;
536		mlog_errno(status);
537		goto leave;
538	}
539	ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), *new_fe_bh);
540
541	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
542					 *new_fe_bh,
543					 OCFS2_JOURNAL_ACCESS_CREATE);
544	if (status < 0) {
545		mlog_errno(status);
546		goto leave;
547	}
548
549	fe = (struct ocfs2_dinode *) (*new_fe_bh)->b_data;
550	memset(fe, 0, osb->sb->s_blocksize);
551
552	fe->i_generation = cpu_to_le32(inode->i_generation);
553	fe->i_fs_generation = cpu_to_le32(osb->fs_generation);
554	fe->i_blkno = cpu_to_le64(fe_blkno);
555	fe->i_suballoc_loc = cpu_to_le64(suballoc_loc);
556	fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
557	fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot);
558	fe->i_uid = cpu_to_le32(i_uid_read(inode));
559	fe->i_gid = cpu_to_le32(i_gid_read(inode));
560	fe->i_mode = cpu_to_le16(inode->i_mode);
561	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
562		fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
563
564	ocfs2_set_links_count(fe, inode->i_nlink);
565
566	fe->i_last_eb_blk = 0;
567	strcpy(fe->i_signature, OCFS2_INODE_SIGNATURE);
568	fe->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
569	ktime_get_real_ts64(&ts);
570	fe->i_atime = fe->i_ctime = fe->i_mtime =
571		cpu_to_le64(ts.tv_sec);
572	fe->i_mtime_nsec = fe->i_ctime_nsec = fe->i_atime_nsec =
573		cpu_to_le32(ts.tv_nsec);
574	fe->i_dtime = 0;
575
576	/*
577	 * If supported, directories start with inline data. If inline
578	 * isn't supported, but indexing is, we start them as indexed.
579	 */
580	feat = le16_to_cpu(fe->i_dyn_features);
581	if (S_ISDIR(inode->i_mode) && ocfs2_supports_inline_data(osb)) {
582		fe->i_dyn_features = cpu_to_le16(feat | OCFS2_INLINE_DATA_FL);
583
584		fe->id2.i_data.id_count = cpu_to_le16(
585				ocfs2_max_inline_data_with_xattr(osb->sb, fe));
586	} else {
587		fel = &fe->id2.i_list;
588		fel->l_tree_depth = 0;
589		fel->l_next_free_rec = 0;
590		fel->l_count = cpu_to_le16(ocfs2_extent_recs_per_inode(osb->sb));
591	}
592
593	ocfs2_journal_dirty(handle, *new_fe_bh);
594
595	ocfs2_populate_inode(inode, fe, 1);
596	ocfs2_ci_set_new(osb, INODE_CACHE(inode));
597	if (!ocfs2_mount_local(osb)) {
598		status = ocfs2_create_new_inode_locks(inode);
599		if (status < 0)
600			mlog_errno(status);
601	}
602
603	ocfs2_update_inode_fsync_trans(handle, inode, 1);
604
605leave:
606	if (status < 0) {
607		if (*new_fe_bh) {
608			brelse(*new_fe_bh);
609			*new_fe_bh = NULL;
610		}
611	}
612
613	if (status)
614		mlog_errno(status);
615	return status;
616}
617
618static int ocfs2_mknod_locked(struct ocfs2_super *osb,
619			      struct inode *dir,
620			      struct inode *inode,
621			      dev_t dev,
622			      struct buffer_head **new_fe_bh,
623			      struct buffer_head *parent_fe_bh,
624			      handle_t *handle,
625			      struct ocfs2_alloc_context *inode_ac)
626{
627	int status = 0;
628	u64 suballoc_loc, fe_blkno = 0;
629	u16 suballoc_bit;
630
631	*new_fe_bh = NULL;
632
633	status = ocfs2_claim_new_inode(handle, dir, parent_fe_bh,
634				       inode_ac, &suballoc_loc,
635				       &suballoc_bit, &fe_blkno);
636	if (status < 0) {
637		mlog_errno(status);
638		return status;
639	}
640
641	return __ocfs2_mknod_locked(dir, inode, dev, new_fe_bh,
642				    parent_fe_bh, handle, inode_ac,
643				    fe_blkno, suballoc_loc, suballoc_bit);
644}
645
646static int ocfs2_mkdir(struct inode *dir,
647		       struct dentry *dentry,
648		       umode_t mode)
649{
650	int ret;
651
652	trace_ocfs2_mkdir(dir, dentry, dentry->d_name.len, dentry->d_name.name,
653			  OCFS2_I(dir)->ip_blkno, mode);
654	ret = ocfs2_mknod(dir, dentry, mode | S_IFDIR, 0);
655	if (ret)
656		mlog_errno(ret);
657
658	return ret;
659}
660
661static int ocfs2_create(struct inode *dir,
662			struct dentry *dentry,
663			umode_t mode,
664			bool excl)
665{
666	int ret;
667
668	trace_ocfs2_create(dir, dentry, dentry->d_name.len, dentry->d_name.name,
669			   (unsigned long long)OCFS2_I(dir)->ip_blkno, mode);
670	ret = ocfs2_mknod(dir, dentry, mode | S_IFREG, 0);
671	if (ret)
672		mlog_errno(ret);
673
674	return ret;
675}
676
677static int ocfs2_link(struct dentry *old_dentry,
678		      struct inode *dir,
679		      struct dentry *dentry)
680{
681	handle_t *handle;
682	struct inode *inode = d_inode(old_dentry);
683	struct inode *old_dir = d_inode(old_dentry->d_parent);
684	int err;
685	struct buffer_head *fe_bh = NULL;
686	struct buffer_head *old_dir_bh = NULL;
687	struct buffer_head *parent_fe_bh = NULL;
688	struct ocfs2_dinode *fe = NULL;
689	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
690	struct ocfs2_dir_lookup_result lookup = { NULL, };
691	sigset_t oldset;
692	u64 old_de_ino;
693
694	trace_ocfs2_link((unsigned long long)OCFS2_I(inode)->ip_blkno,
695			 old_dentry->d_name.len, old_dentry->d_name.name,
696			 dentry->d_name.len, dentry->d_name.name);
697
698	if (S_ISDIR(inode->i_mode))
699		return -EPERM;
700
701	err = dquot_initialize(dir);
702	if (err) {
703		mlog_errno(err);
704		return err;
705	}
706
707	err = ocfs2_double_lock(osb, &old_dir_bh, old_dir,
708			&parent_fe_bh, dir, 0);
709	if (err < 0) {
710		if (err != -ENOENT)
711			mlog_errno(err);
712		return err;
713	}
714
715	/* make sure both dirs have bhs
716	 * get an extra ref on old_dir_bh if old==new */
717	if (!parent_fe_bh) {
718		if (old_dir_bh) {
719			parent_fe_bh = old_dir_bh;
720			get_bh(parent_fe_bh);
721		} else {
722			mlog(ML_ERROR, "%s: no old_dir_bh!\n", osb->uuid_str);
723			err = -EIO;
724			goto out;
725		}
726	}
727
728	if (!dir->i_nlink) {
729		err = -ENOENT;
730		goto out;
731	}
732
733	err = ocfs2_lookup_ino_from_name(old_dir, old_dentry->d_name.name,
734			old_dentry->d_name.len, &old_de_ino);
735	if (err) {
736		err = -ENOENT;
737		goto out;
738	}
739
740	/*
741	 * Check whether another node removed the source inode while we
742	 * were in the vfs.
743	 */
744	if (old_de_ino != OCFS2_I(inode)->ip_blkno) {
745		err = -ENOENT;
746		goto out;
747	}
748
749	err = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
750					dentry->d_name.len);
751	if (err)
752		goto out;
753
754	err = ocfs2_prepare_dir_for_insert(osb, dir, parent_fe_bh,
755					   dentry->d_name.name,
756					   dentry->d_name.len, &lookup);
757	if (err < 0) {
758		mlog_errno(err);
759		goto out;
760	}
761
762	err = ocfs2_inode_lock(inode, &fe_bh, 1);
763	if (err < 0) {
764		if (err != -ENOENT)
765			mlog_errno(err);
766		goto out;
767	}
768
769	fe = (struct ocfs2_dinode *) fe_bh->b_data;
770	if (ocfs2_read_links_count(fe) >= ocfs2_link_max(osb)) {
771		err = -EMLINK;
772		goto out_unlock_inode;
773	}
774
775	handle = ocfs2_start_trans(osb, ocfs2_link_credits(osb->sb));
776	if (IS_ERR(handle)) {
777		err = PTR_ERR(handle);
778		handle = NULL;
779		mlog_errno(err);
780		goto out_unlock_inode;
781	}
782
783	/* Starting to change things, restart is no longer possible. */
784	ocfs2_block_signals(&oldset);
785
786	err = ocfs2_journal_access_di(handle, INODE_CACHE(inode), fe_bh,
787				      OCFS2_JOURNAL_ACCESS_WRITE);
788	if (err < 0) {
789		mlog_errno(err);
790		goto out_commit;
791	}
792
793	inc_nlink(inode);
794	inode->i_ctime = current_time(inode);
795	ocfs2_set_links_count(fe, inode->i_nlink);
796	fe->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec);
797	fe->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec);
798	ocfs2_journal_dirty(handle, fe_bh);
799
800	err = ocfs2_add_entry(handle, dentry, inode,
801			      OCFS2_I(inode)->ip_blkno,
802			      parent_fe_bh, &lookup);
803	if (err) {
804		ocfs2_add_links_count(fe, -1);
805		drop_nlink(inode);
806		mlog_errno(err);
807		goto out_commit;
808	}
809
810	err = ocfs2_dentry_attach_lock(dentry, inode, OCFS2_I(dir)->ip_blkno);
811	if (err) {
812		mlog_errno(err);
813		goto out_commit;
814	}
815
816	ihold(inode);
817	d_instantiate(dentry, inode);
818
819out_commit:
820	ocfs2_commit_trans(osb, handle);
821	ocfs2_unblock_signals(&oldset);
822out_unlock_inode:
823	ocfs2_inode_unlock(inode, 1);
824
825out:
826	ocfs2_double_unlock(old_dir, dir);
827
828	brelse(fe_bh);
829	brelse(parent_fe_bh);
830	brelse(old_dir_bh);
831
832	ocfs2_free_dir_lookup_result(&lookup);
833
834	if (err)
835		mlog_errno(err);
836
837	return err;
838}
839
840/*
841 * Takes and drops an exclusive lock on the given dentry. This will
842 * force other nodes to drop it.
843 */
844static int ocfs2_remote_dentry_delete(struct dentry *dentry)
845{
846	int ret;
847
848	ret = ocfs2_dentry_lock(dentry, 1);
849	if (ret)
850		mlog_errno(ret);
851	else
852		ocfs2_dentry_unlock(dentry, 1);
853
854	return ret;
855}
856
857static inline int ocfs2_inode_is_unlinkable(struct inode *inode)
858{
859	if (S_ISDIR(inode->i_mode)) {
860		if (inode->i_nlink == 2)
861			return 1;
862		return 0;
863	}
864
865	if (inode->i_nlink == 1)
866		return 1;
867	return 0;
868}
869
870static int ocfs2_unlink(struct inode *dir,
871			struct dentry *dentry)
872{
873	int status;
874	int child_locked = 0;
875	bool is_unlinkable = false;
876	struct inode *inode = d_inode(dentry);
877	struct inode *orphan_dir = NULL;
878	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
879	u64 blkno;
880	struct ocfs2_dinode *fe = NULL;
881	struct buffer_head *fe_bh = NULL;
882	struct buffer_head *parent_node_bh = NULL;
883	handle_t *handle = NULL;
884	char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
885	struct ocfs2_dir_lookup_result lookup = { NULL, };
886	struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
887
888	trace_ocfs2_unlink(dir, dentry, dentry->d_name.len,
889			   dentry->d_name.name,
890			   (unsigned long long)OCFS2_I(dir)->ip_blkno,
891			   (unsigned long long)OCFS2_I(inode)->ip_blkno);
892
893	status = dquot_initialize(dir);
894	if (status) {
895		mlog_errno(status);
896		return status;
897	}
898
899	BUG_ON(d_inode(dentry->d_parent) != dir);
900
901	if (inode == osb->root_inode)
902		return -EPERM;
903
904	status = ocfs2_inode_lock_nested(dir, &parent_node_bh, 1,
905					 OI_LS_PARENT);
906	if (status < 0) {
907		if (status != -ENOENT)
908			mlog_errno(status);
909		return status;
910	}
911
912	status = ocfs2_find_files_on_disk(dentry->d_name.name,
913					  dentry->d_name.len, &blkno, dir,
914					  &lookup);
915	if (status < 0) {
916		if (status != -ENOENT)
917			mlog_errno(status);
918		goto leave;
919	}
920
921	if (OCFS2_I(inode)->ip_blkno != blkno) {
922		status = -ENOENT;
923
924		trace_ocfs2_unlink_noent(
925				(unsigned long long)OCFS2_I(inode)->ip_blkno,
926				(unsigned long long)blkno,
927				OCFS2_I(inode)->ip_flags);
928		goto leave;
929	}
930
931	status = ocfs2_inode_lock(inode, &fe_bh, 1);
932	if (status < 0) {
933		if (status != -ENOENT)
934			mlog_errno(status);
935		goto leave;
936	}
937	child_locked = 1;
938
939	if (S_ISDIR(inode->i_mode)) {
940		if (inode->i_nlink != 2 || !ocfs2_empty_dir(inode)) {
941			status = -ENOTEMPTY;
942			goto leave;
943		}
944	}
945
946	status = ocfs2_remote_dentry_delete(dentry);
947	if (status < 0) {
948		/* This remote delete should succeed under all normal
949		 * circumstances. */
950		mlog_errno(status);
951		goto leave;
952	}
953
954	if (ocfs2_inode_is_unlinkable(inode)) {
955		status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
956						  OCFS2_I(inode)->ip_blkno,
957						  orphan_name, &orphan_insert,
958						  false);
959		if (status < 0) {
960			mlog_errno(status);
961			goto leave;
962		}
963		is_unlinkable = true;
964	}
965
966	handle = ocfs2_start_trans(osb, ocfs2_unlink_credits(osb->sb));
967	if (IS_ERR(handle)) {
968		status = PTR_ERR(handle);
969		handle = NULL;
970		mlog_errno(status);
971		goto leave;
972	}
973
974	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), fe_bh,
975					 OCFS2_JOURNAL_ACCESS_WRITE);
976	if (status < 0) {
977		mlog_errno(status);
978		goto leave;
979	}
980
981	fe = (struct ocfs2_dinode *) fe_bh->b_data;
982
983	/* delete the name from the parent dir */
984	status = ocfs2_delete_entry(handle, dir, &lookup);
985	if (status < 0) {
986		mlog_errno(status);
987		goto leave;
988	}
989
990	if (S_ISDIR(inode->i_mode))
991		drop_nlink(inode);
992	drop_nlink(inode);
993	ocfs2_set_links_count(fe, inode->i_nlink);
994	ocfs2_journal_dirty(handle, fe_bh);
995
996	dir->i_ctime = dir->i_mtime = current_time(dir);
997	if (S_ISDIR(inode->i_mode))
998		drop_nlink(dir);
999
1000	status = ocfs2_mark_inode_dirty(handle, dir, parent_node_bh);
1001	if (status < 0) {
1002		mlog_errno(status);
1003		if (S_ISDIR(inode->i_mode))
1004			inc_nlink(dir);
1005		goto leave;
1006	}
1007
1008	if (is_unlinkable) {
1009		status = ocfs2_orphan_add(osb, handle, inode, fe_bh,
1010				orphan_name, &orphan_insert, orphan_dir, false);
1011		if (status < 0)
1012			mlog_errno(status);
1013	}
1014
1015leave:
1016	if (handle)
1017		ocfs2_commit_trans(osb, handle);
1018
1019	if (orphan_dir) {
1020		/* This was locked for us in ocfs2_prepare_orphan_dir() */
1021		ocfs2_inode_unlock(orphan_dir, 1);
1022		inode_unlock(orphan_dir);
1023		iput(orphan_dir);
1024	}
1025
1026	if (child_locked)
1027		ocfs2_inode_unlock(inode, 1);
1028
1029	ocfs2_inode_unlock(dir, 1);
1030
1031	brelse(fe_bh);
1032	brelse(parent_node_bh);
1033
1034	ocfs2_free_dir_lookup_result(&orphan_insert);
1035	ocfs2_free_dir_lookup_result(&lookup);
1036
1037	if (status && (status != -ENOTEMPTY) && (status != -ENOENT))
1038		mlog_errno(status);
1039
1040	return status;
1041}
1042
1043static int ocfs2_check_if_ancestor(struct ocfs2_super *osb,
1044		u64 src_inode_no, u64 dest_inode_no)
1045{
1046	int ret = 0, i = 0;
1047	u64 parent_inode_no = 0;
1048	u64 child_inode_no = src_inode_no;
1049	struct inode *child_inode;
1050
1051#define MAX_LOOKUP_TIMES 32
1052	while (1) {
1053		child_inode = ocfs2_iget(osb, child_inode_no, 0, 0);
1054		if (IS_ERR(child_inode)) {
1055			ret = PTR_ERR(child_inode);
1056			break;
1057		}
1058
1059		ret = ocfs2_inode_lock(child_inode, NULL, 0);
1060		if (ret < 0) {
1061			iput(child_inode);
1062			if (ret != -ENOENT)
1063				mlog_errno(ret);
1064			break;
1065		}
1066
1067		ret = ocfs2_lookup_ino_from_name(child_inode, "..", 2,
1068				&parent_inode_no);
1069		ocfs2_inode_unlock(child_inode, 0);
1070		iput(child_inode);
1071		if (ret < 0) {
1072			ret = -ENOENT;
1073			break;
1074		}
1075
1076		if (parent_inode_no == dest_inode_no) {
1077			ret = 1;
1078			break;
1079		}
1080
1081		if (parent_inode_no == osb->root_inode->i_ino) {
1082			ret = 0;
1083			break;
1084		}
1085
1086		child_inode_no = parent_inode_no;
1087
1088		if (++i >= MAX_LOOKUP_TIMES) {
1089			mlog(ML_NOTICE, "max lookup times reached, filesystem "
1090					"may have nested directories, "
1091					"src inode: %llu, dest inode: %llu.\n",
1092					(unsigned long long)src_inode_no,
1093					(unsigned long long)dest_inode_no);
1094			ret = 0;
1095			break;
1096		}
1097	}
1098
1099	return ret;
1100}
1101
1102/*
1103 * The only place this should be used is rename and link!
1104 * if they have the same id, then the 1st one is the only one locked.
1105 */
1106static int ocfs2_double_lock(struct ocfs2_super *osb,
1107			     struct buffer_head **bh1,
1108			     struct inode *inode1,
1109			     struct buffer_head **bh2,
1110			     struct inode *inode2,
1111			     int rename)
1112{
1113	int status;
1114	int inode1_is_ancestor, inode2_is_ancestor;
1115	struct ocfs2_inode_info *oi1 = OCFS2_I(inode1);
1116	struct ocfs2_inode_info *oi2 = OCFS2_I(inode2);
1117
1118	trace_ocfs2_double_lock((unsigned long long)oi1->ip_blkno,
1119				(unsigned long long)oi2->ip_blkno);
1120
1121	if (*bh1)
1122		*bh1 = NULL;
1123	if (*bh2)
1124		*bh2 = NULL;
1125
1126	/* we always want to lock the one with the lower lockid first.
1127	 * and if they are nested, we lock ancestor first */
1128	if (oi1->ip_blkno != oi2->ip_blkno) {
1129		inode1_is_ancestor = ocfs2_check_if_ancestor(osb, oi2->ip_blkno,
1130				oi1->ip_blkno);
1131		if (inode1_is_ancestor < 0) {
1132			status = inode1_is_ancestor;
1133			goto bail;
1134		}
1135
1136		inode2_is_ancestor = ocfs2_check_if_ancestor(osb, oi1->ip_blkno,
1137				oi2->ip_blkno);
1138		if (inode2_is_ancestor < 0) {
1139			status = inode2_is_ancestor;
1140			goto bail;
1141		}
1142
1143		if ((inode1_is_ancestor == 1) ||
1144				(oi1->ip_blkno < oi2->ip_blkno &&
1145				inode2_is_ancestor == 0)) {
1146			/* switch id1 and id2 around */
1147			swap(bh2, bh1);
1148			swap(inode2, inode1);
1149		}
1150		/* lock id2 */
1151		status = ocfs2_inode_lock_nested(inode2, bh2, 1,
1152				rename == 1 ? OI_LS_RENAME1 : OI_LS_PARENT);
1153		if (status < 0) {
1154			if (status != -ENOENT)
1155				mlog_errno(status);
1156			goto bail;
1157		}
1158	}
1159
1160	/* lock id1 */
1161	status = ocfs2_inode_lock_nested(inode1, bh1, 1,
1162			rename == 1 ?  OI_LS_RENAME2 : OI_LS_PARENT);
1163	if (status < 0) {
1164		/*
1165		 * An error return must mean that no cluster locks
1166		 * were held on function exit.
1167		 */
1168		if (oi1->ip_blkno != oi2->ip_blkno) {
1169			ocfs2_inode_unlock(inode2, 1);
1170			brelse(*bh2);
1171			*bh2 = NULL;
1172		}
1173
1174		if (status != -ENOENT)
1175			mlog_errno(status);
1176	}
1177
1178	trace_ocfs2_double_lock_end(
1179			(unsigned long long)oi1->ip_blkno,
1180			(unsigned long long)oi2->ip_blkno);
1181
1182bail:
1183	if (status)
1184		mlog_errno(status);
1185	return status;
1186}
1187
1188static void ocfs2_double_unlock(struct inode *inode1, struct inode *inode2)
1189{
1190	ocfs2_inode_unlock(inode1, 1);
1191
1192	if (inode1 != inode2)
1193		ocfs2_inode_unlock(inode2, 1);
1194}
1195
1196static int ocfs2_rename(struct inode *old_dir,
1197			struct dentry *old_dentry,
1198			struct inode *new_dir,
1199			struct dentry *new_dentry,
1200			unsigned int flags)
1201{
1202	int status = 0, rename_lock = 0, parents_locked = 0, target_exists = 0;
1203	int old_child_locked = 0, new_child_locked = 0, update_dot_dot = 0;
1204	struct inode *old_inode = d_inode(old_dentry);
1205	struct inode *new_inode = d_inode(new_dentry);
1206	struct inode *orphan_dir = NULL;
1207	struct ocfs2_dinode *newfe = NULL;
1208	char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
1209	struct buffer_head *newfe_bh = NULL;
1210	struct buffer_head *old_inode_bh = NULL;
1211	struct ocfs2_super *osb = NULL;
1212	u64 newfe_blkno, old_de_ino;
1213	handle_t *handle = NULL;
1214	struct buffer_head *old_dir_bh = NULL;
1215	struct buffer_head *new_dir_bh = NULL;
1216	u32 old_dir_nlink = old_dir->i_nlink;
1217	struct ocfs2_dinode *old_di;
1218	struct ocfs2_dir_lookup_result old_inode_dot_dot_res = { NULL, };
1219	struct ocfs2_dir_lookup_result target_lookup_res = { NULL, };
1220	struct ocfs2_dir_lookup_result old_entry_lookup = { NULL, };
1221	struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
1222	struct ocfs2_dir_lookup_result target_insert = { NULL, };
1223	bool should_add_orphan = false;
1224
1225	if (flags)
1226		return -EINVAL;
1227
1228	/* At some point it might be nice to break this function up a
1229	 * bit. */
1230
1231	trace_ocfs2_rename(old_dir, old_dentry, new_dir, new_dentry,
1232			   old_dentry->d_name.len, old_dentry->d_name.name,
1233			   new_dentry->d_name.len, new_dentry->d_name.name);
1234
1235	status = dquot_initialize(old_dir);
1236	if (status) {
1237		mlog_errno(status);
1238		goto bail;
1239	}
1240	status = dquot_initialize(new_dir);
1241	if (status) {
1242		mlog_errno(status);
1243		goto bail;
1244	}
1245
1246	osb = OCFS2_SB(old_dir->i_sb);
1247
1248	if (new_inode) {
1249		if (!igrab(new_inode))
1250			BUG();
1251	}
1252
1253	/* Assume a directory hierarchy thusly:
1254	 * a/b/c
1255	 * a/d
1256	 * a,b,c, and d are all directories.
1257	 *
1258	 * from cwd of 'a' on both nodes:
1259	 * node1: mv b/c d
1260	 * node2: mv d   b/c
1261	 *
1262	 * And that's why, just like the VFS, we need a file system
1263	 * rename lock. */
1264	if (old_dir != new_dir && S_ISDIR(old_inode->i_mode)) {
1265		status = ocfs2_rename_lock(osb);
1266		if (status < 0) {
1267			mlog_errno(status);
1268			goto bail;
1269		}
1270		rename_lock = 1;
1271
1272		/* here we cannot guarantee the inodes haven't just been
1273		 * changed, so check if they are nested again */
1274		status = ocfs2_check_if_ancestor(osb, new_dir->i_ino,
1275				old_inode->i_ino);
1276		if (status < 0) {
1277			mlog_errno(status);
1278			goto bail;
1279		} else if (status == 1) {
1280			status = -EPERM;
1281			trace_ocfs2_rename_not_permitted(
1282					(unsigned long long)old_inode->i_ino,
1283					(unsigned long long)new_dir->i_ino);
1284			goto bail;
1285		}
1286	}
1287
1288	/* if old and new are the same, this'll just do one lock. */
1289	status = ocfs2_double_lock(osb, &old_dir_bh, old_dir,
1290				   &new_dir_bh, new_dir, 1);
1291	if (status < 0) {
1292		mlog_errno(status);
1293		goto bail;
1294	}
1295	parents_locked = 1;
1296
1297	if (!new_dir->i_nlink) {
1298		status = -EACCES;
1299		goto bail;
1300	}
1301
1302	/* make sure both dirs have bhs
1303	 * get an extra ref on old_dir_bh if old==new */
1304	if (!new_dir_bh) {
1305		if (old_dir_bh) {
1306			new_dir_bh = old_dir_bh;
1307			get_bh(new_dir_bh);
1308		} else {
1309			mlog(ML_ERROR, "no old_dir_bh!\n");
1310			status = -EIO;
1311			goto bail;
1312		}
1313	}
1314
1315	/*
1316	 * Aside from allowing a meta data update, the locking here
1317	 * also ensures that the downconvert thread on other nodes
1318	 * won't have to concurrently downconvert the inode and the
1319	 * dentry locks.
1320	 */
1321	status = ocfs2_inode_lock_nested(old_inode, &old_inode_bh, 1,
1322					 OI_LS_PARENT);
1323	if (status < 0) {
1324		if (status != -ENOENT)
1325			mlog_errno(status);
1326		goto bail;
1327	}
1328	old_child_locked = 1;
1329
1330	status = ocfs2_remote_dentry_delete(old_dentry);
1331	if (status < 0) {
1332		mlog_errno(status);
1333		goto bail;
1334	}
1335
1336	if (S_ISDIR(old_inode->i_mode)) {
1337		u64 old_inode_parent;
1338
1339		update_dot_dot = 1;
1340		status = ocfs2_find_files_on_disk("..", 2, &old_inode_parent,
1341						  old_inode,
1342						  &old_inode_dot_dot_res);
1343		if (status) {
1344			status = -EIO;
1345			goto bail;
1346		}
1347
1348		if (old_inode_parent != OCFS2_I(old_dir)->ip_blkno) {
1349			status = -EIO;
1350			goto bail;
1351		}
1352
1353		if (!new_inode && new_dir != old_dir &&
1354		    new_dir->i_nlink >= ocfs2_link_max(osb)) {
1355			status = -EMLINK;
1356			goto bail;
1357		}
1358	}
1359
1360	status = ocfs2_lookup_ino_from_name(old_dir, old_dentry->d_name.name,
1361					    old_dentry->d_name.len,
1362					    &old_de_ino);
1363	if (status) {
1364		status = -ENOENT;
1365		goto bail;
1366	}
1367
1368	/*
1369	 *  Check for inode number is _not_ due to possible IO errors.
1370	 *  We might rmdir the source, keep it as pwd of some process
1371	 *  and merrily kill the link to whatever was created under the
1372	 *  same name. Goodbye sticky bit ;-<
1373	 */
1374	if (old_de_ino != OCFS2_I(old_inode)->ip_blkno) {
1375		status = -ENOENT;
1376		goto bail;
1377	}
1378
1379	/* check if the target already exists (in which case we need
1380	 * to delete it */
1381	status = ocfs2_find_files_on_disk(new_dentry->d_name.name,
1382					  new_dentry->d_name.len,
1383					  &newfe_blkno, new_dir,
1384					  &target_lookup_res);
1385	/* The only error we allow here is -ENOENT because the new
1386	 * file not existing is perfectly valid. */
1387	if ((status < 0) && (status != -ENOENT)) {
1388		/* If we cannot find the file specified we should just */
1389		/* return the error... */
1390		mlog_errno(status);
1391		goto bail;
1392	}
1393	if (status == 0)
1394		target_exists = 1;
1395
1396	if (!target_exists && new_inode) {
1397		/*
1398		 * Target was unlinked by another node while we were
1399		 * waiting to get to ocfs2_rename(). There isn't
1400		 * anything we can do here to help the situation, so
1401		 * bubble up the appropriate error.
1402		 */
1403		status = -ENOENT;
1404		goto bail;
1405	}
1406
1407	/* In case we need to overwrite an existing file, we blow it
1408	 * away first */
1409	if (target_exists) {
1410		/* VFS didn't think there existed an inode here, but
1411		 * someone else in the cluster must have raced our
1412		 * rename to create one. Today we error cleanly, in
1413		 * the future we should consider calling iget to build
1414		 * a new struct inode for this entry. */
1415		if (!new_inode) {
1416			status = -EACCES;
1417
1418			trace_ocfs2_rename_target_exists(new_dentry->d_name.len,
1419						new_dentry->d_name.name);
1420			goto bail;
1421		}
1422
1423		if (OCFS2_I(new_inode)->ip_blkno != newfe_blkno) {
1424			status = -EACCES;
1425
1426			trace_ocfs2_rename_disagree(
1427			     (unsigned long long)OCFS2_I(new_inode)->ip_blkno,
1428			     (unsigned long long)newfe_blkno,
1429			     OCFS2_I(new_inode)->ip_flags);
1430			goto bail;
1431		}
1432
1433		status = ocfs2_inode_lock(new_inode, &newfe_bh, 1);
1434		if (status < 0) {
1435			if (status != -ENOENT)
1436				mlog_errno(status);
1437			goto bail;
1438		}
1439		new_child_locked = 1;
1440
1441		status = ocfs2_remote_dentry_delete(new_dentry);
1442		if (status < 0) {
1443			mlog_errno(status);
1444			goto bail;
1445		}
1446
1447		newfe = (struct ocfs2_dinode *) newfe_bh->b_data;
1448
1449		trace_ocfs2_rename_over_existing(
1450		     (unsigned long long)newfe_blkno, newfe_bh, newfe_bh ?
1451		     (unsigned long long)newfe_bh->b_blocknr : 0ULL);
1452
1453		if (S_ISDIR(new_inode->i_mode) || (new_inode->i_nlink == 1)) {
1454			status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
1455						OCFS2_I(new_inode)->ip_blkno,
1456						orphan_name, &orphan_insert,
1457						false);
1458			if (status < 0) {
1459				mlog_errno(status);
1460				goto bail;
1461			}
1462			should_add_orphan = true;
1463		}
1464	} else {
1465		BUG_ON(d_inode(new_dentry->d_parent) != new_dir);
1466
1467		status = ocfs2_check_dir_for_entry(new_dir,
1468						   new_dentry->d_name.name,
1469						   new_dentry->d_name.len);
1470		if (status)
1471			goto bail;
1472
1473		status = ocfs2_prepare_dir_for_insert(osb, new_dir, new_dir_bh,
1474						      new_dentry->d_name.name,
1475						      new_dentry->d_name.len,
1476						      &target_insert);
1477		if (status < 0) {
1478			mlog_errno(status);
1479			goto bail;
1480		}
1481	}
1482
1483	handle = ocfs2_start_trans(osb, ocfs2_rename_credits(osb->sb));
1484	if (IS_ERR(handle)) {
1485		status = PTR_ERR(handle);
1486		handle = NULL;
1487		mlog_errno(status);
1488		goto bail;
1489	}
1490
1491	if (target_exists) {
1492		if (S_ISDIR(new_inode->i_mode)) {
1493			if (new_inode->i_nlink != 2 ||
1494			    !ocfs2_empty_dir(new_inode)) {
1495				status = -ENOTEMPTY;
1496				goto bail;
1497			}
1498		}
1499		status = ocfs2_journal_access_di(handle, INODE_CACHE(new_inode),
1500						 newfe_bh,
1501						 OCFS2_JOURNAL_ACCESS_WRITE);
1502		if (status < 0) {
1503			mlog_errno(status);
1504			goto bail;
1505		}
1506
1507		/* change the dirent to point to the correct inode */
1508		status = ocfs2_update_entry(new_dir, handle, &target_lookup_res,
1509					    old_inode);
1510		if (status < 0) {
1511			mlog_errno(status);
1512			goto bail;
1513		}
1514		inode_inc_iversion(new_dir);
1515
1516		if (S_ISDIR(new_inode->i_mode))
1517			ocfs2_set_links_count(newfe, 0);
1518		else
1519			ocfs2_add_links_count(newfe, -1);
1520		ocfs2_journal_dirty(handle, newfe_bh);
1521		if (should_add_orphan) {
1522			status = ocfs2_orphan_add(osb, handle, new_inode,
1523					newfe_bh, orphan_name,
1524					&orphan_insert, orphan_dir, false);
1525			if (status < 0) {
1526				mlog_errno(status);
1527				goto bail;
1528			}
1529		}
1530	} else {
1531		/* if the name was not found in new_dir, add it now */
1532		status = ocfs2_add_entry(handle, new_dentry, old_inode,
1533					 OCFS2_I(old_inode)->ip_blkno,
1534					 new_dir_bh, &target_insert);
1535		if (status < 0) {
1536			mlog_errno(status);
1537			goto bail;
1538		}
1539	}
1540
1541	old_inode->i_ctime = current_time(old_inode);
1542	mark_inode_dirty(old_inode);
1543
1544	status = ocfs2_journal_access_di(handle, INODE_CACHE(old_inode),
1545					 old_inode_bh,
1546					 OCFS2_JOURNAL_ACCESS_WRITE);
1547	if (status >= 0) {
1548		old_di = (struct ocfs2_dinode *) old_inode_bh->b_data;
1549
1550		old_di->i_ctime = cpu_to_le64(old_inode->i_ctime.tv_sec);
1551		old_di->i_ctime_nsec = cpu_to_le32(old_inode->i_ctime.tv_nsec);
1552		ocfs2_journal_dirty(handle, old_inode_bh);
1553	} else
1554		mlog_errno(status);
1555
1556	/*
1557	 * Now that the name has been added to new_dir, remove the old name.
1558	 *
1559	 * We don't keep any directory entry context around until now
1560	 * because the insert might have changed the type of directory
1561	 * we're dealing with.
1562	 */
1563	status = ocfs2_find_entry(old_dentry->d_name.name,
1564				  old_dentry->d_name.len, old_dir,
1565				  &old_entry_lookup);
1566	if (status) {
1567		if (!is_journal_aborted(osb->journal->j_journal)) {
1568			ocfs2_error(osb->sb, "new entry %.*s is added, but old entry %.*s "
1569					"is not deleted.",
1570					new_dentry->d_name.len, new_dentry->d_name.name,
1571					old_dentry->d_name.len, old_dentry->d_name.name);
1572		}
1573		goto bail;
1574	}
1575
1576	status = ocfs2_delete_entry(handle, old_dir, &old_entry_lookup);
1577	if (status < 0) {
1578		mlog_errno(status);
1579		if (!is_journal_aborted(osb->journal->j_journal)) {
1580			ocfs2_error(osb->sb, "new entry %.*s is added, but old entry %.*s "
1581					"is not deleted.",
1582					new_dentry->d_name.len, new_dentry->d_name.name,
1583					old_dentry->d_name.len, old_dentry->d_name.name);
1584		}
1585		goto bail;
1586	}
1587
1588	if (new_inode) {
1589		drop_nlink(new_inode);
1590		new_inode->i_ctime = current_time(new_inode);
1591	}
1592	old_dir->i_ctime = old_dir->i_mtime = current_time(old_dir);
1593
1594	if (update_dot_dot) {
1595		status = ocfs2_update_entry(old_inode, handle,
1596					    &old_inode_dot_dot_res, new_dir);
1597		drop_nlink(old_dir);
1598		if (new_inode) {
1599			drop_nlink(new_inode);
1600		} else {
1601			inc_nlink(new_dir);
1602			mark_inode_dirty(new_dir);
1603		}
1604	}
1605	mark_inode_dirty(old_dir);
1606	ocfs2_mark_inode_dirty(handle, old_dir, old_dir_bh);
1607	if (new_inode) {
1608		mark_inode_dirty(new_inode);
1609		ocfs2_mark_inode_dirty(handle, new_inode, newfe_bh);
1610	}
1611
1612	if (old_dir != new_dir) {
1613		/* Keep the same times on both directories.*/
1614		new_dir->i_ctime = new_dir->i_mtime = old_dir->i_ctime;
1615
1616		/*
1617		 * This will also pick up the i_nlink change from the
1618		 * block above.
1619		 */
1620		ocfs2_mark_inode_dirty(handle, new_dir, new_dir_bh);
1621	}
1622
1623	if (old_dir_nlink != old_dir->i_nlink) {
1624		if (!old_dir_bh) {
1625			mlog(ML_ERROR, "need to change nlink for old dir "
1626			     "%llu from %d to %d but bh is NULL!\n",
1627			     (unsigned long long)OCFS2_I(old_dir)->ip_blkno,
1628			     (int)old_dir_nlink, old_dir->i_nlink);
1629		} else {
1630			struct ocfs2_dinode *fe;
1631			status = ocfs2_journal_access_di(handle,
1632							 INODE_CACHE(old_dir),
1633							 old_dir_bh,
1634							 OCFS2_JOURNAL_ACCESS_WRITE);
1635			fe = (struct ocfs2_dinode *) old_dir_bh->b_data;
1636			ocfs2_set_links_count(fe, old_dir->i_nlink);
1637			ocfs2_journal_dirty(handle, old_dir_bh);
1638		}
1639	}
1640	ocfs2_dentry_move(old_dentry, new_dentry, old_dir, new_dir);
1641	status = 0;
1642bail:
1643	if (handle)
1644		ocfs2_commit_trans(osb, handle);
1645
1646	if (orphan_dir) {
1647		/* This was locked for us in ocfs2_prepare_orphan_dir() */
1648		ocfs2_inode_unlock(orphan_dir, 1);
1649		inode_unlock(orphan_dir);
1650		iput(orphan_dir);
1651	}
1652
1653	if (new_child_locked)
1654		ocfs2_inode_unlock(new_inode, 1);
1655
1656	if (old_child_locked)
1657		ocfs2_inode_unlock(old_inode, 1);
1658
1659	if (parents_locked)
1660		ocfs2_double_unlock(old_dir, new_dir);
1661
1662	if (rename_lock)
1663		ocfs2_rename_unlock(osb);
1664
1665	if (new_inode)
1666		sync_mapping_buffers(old_inode->i_mapping);
1667
1668	iput(new_inode);
1669
1670	ocfs2_free_dir_lookup_result(&target_lookup_res);
1671	ocfs2_free_dir_lookup_result(&old_entry_lookup);
1672	ocfs2_free_dir_lookup_result(&old_inode_dot_dot_res);
1673	ocfs2_free_dir_lookup_result(&orphan_insert);
1674	ocfs2_free_dir_lookup_result(&target_insert);
1675
1676	brelse(newfe_bh);
1677	brelse(old_inode_bh);
1678	brelse(old_dir_bh);
1679	brelse(new_dir_bh);
1680
1681	if (status)
1682		mlog_errno(status);
1683
1684	return status;
1685}
1686
1687/*
1688 * we expect i_size = strlen(symname). Copy symname into the file
1689 * data, including the null terminator.
1690 */
1691static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
1692				     handle_t *handle,
1693				     struct inode *inode,
1694				     const char *symname)
1695{
1696	struct buffer_head **bhs = NULL;
1697	const char *c;
1698	struct super_block *sb = osb->sb;
1699	u64 p_blkno, p_blocks;
1700	int virtual, blocks, status, i, bytes_left;
1701
1702	bytes_left = i_size_read(inode) + 1;
1703	/* we can't trust i_blocks because we're actually going to
1704	 * write i_size + 1 bytes. */
1705	blocks = (bytes_left + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
1706
1707	trace_ocfs2_create_symlink_data((unsigned long long)inode->i_blocks,
1708					i_size_read(inode), blocks);
1709
1710	/* Sanity check -- make sure we're going to fit. */
1711	if (bytes_left >
1712	    ocfs2_clusters_to_bytes(sb, OCFS2_I(inode)->ip_clusters)) {
1713		status = -EIO;
1714		mlog_errno(status);
1715		goto bail;
1716	}
1717
1718	bhs = kcalloc(blocks, sizeof(struct buffer_head *), GFP_KERNEL);
1719	if (!bhs) {
1720		status = -ENOMEM;
1721		mlog_errno(status);
1722		goto bail;
1723	}
1724
1725	status = ocfs2_extent_map_get_blocks(inode, 0, &p_blkno, &p_blocks,
1726					     NULL);
1727	if (status < 0) {
1728		mlog_errno(status);
1729		goto bail;
1730	}
1731
1732	/* links can never be larger than one cluster so we know this
1733	 * is all going to be contiguous, but do a sanity check
1734	 * anyway. */
1735	if ((p_blocks << sb->s_blocksize_bits) < bytes_left) {
1736		status = -EIO;
1737		mlog_errno(status);
1738		goto bail;
1739	}
1740
1741	virtual = 0;
1742	while(bytes_left > 0) {
1743		c = &symname[virtual * sb->s_blocksize];
1744
1745		bhs[virtual] = sb_getblk(sb, p_blkno);
1746		if (!bhs[virtual]) {
1747			status = -ENOMEM;
1748			mlog_errno(status);
1749			goto bail;
1750		}
1751		ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode),
1752					      bhs[virtual]);
1753
1754		status = ocfs2_journal_access(handle, INODE_CACHE(inode),
1755					      bhs[virtual],
1756					      OCFS2_JOURNAL_ACCESS_CREATE);
1757		if (status < 0) {
1758			mlog_errno(status);
1759			goto bail;
1760		}
1761
1762		memset(bhs[virtual]->b_data, 0, sb->s_blocksize);
1763
1764		memcpy(bhs[virtual]->b_data, c,
1765		       (bytes_left > sb->s_blocksize) ? sb->s_blocksize :
1766		       bytes_left);
1767
1768		ocfs2_journal_dirty(handle, bhs[virtual]);
1769
1770		virtual++;
1771		p_blkno++;
1772		bytes_left -= sb->s_blocksize;
1773	}
1774
1775	status = 0;
1776bail:
1777
1778	if (bhs) {
1779		for(i = 0; i < blocks; i++)
1780			brelse(bhs[i]);
1781		kfree(bhs);
1782	}
1783
1784	if (status)
1785		mlog_errno(status);
1786	return status;
1787}
1788
1789static int ocfs2_symlink(struct inode *dir,
1790			 struct dentry *dentry,
1791			 const char *symname)
1792{
1793	int status, l, credits;
1794	u64 newsize;
1795	struct ocfs2_super *osb = NULL;
1796	struct inode *inode = NULL;
1797	struct super_block *sb;
1798	struct buffer_head *new_fe_bh = NULL;
1799	struct buffer_head *parent_fe_bh = NULL;
1800	struct ocfs2_dinode *fe = NULL;
1801	struct ocfs2_dinode *dirfe;
1802	handle_t *handle = NULL;
1803	struct ocfs2_alloc_context *inode_ac = NULL;
1804	struct ocfs2_alloc_context *data_ac = NULL;
1805	struct ocfs2_alloc_context *xattr_ac = NULL;
1806	int want_clusters = 0;
1807	int xattr_credits = 0;
1808	struct ocfs2_security_xattr_info si = {
1809		.name = NULL,
1810		.enable = 1,
1811	};
1812	int did_quota = 0, did_quota_inode = 0;
1813	struct ocfs2_dir_lookup_result lookup = { NULL, };
1814	sigset_t oldset;
1815	int did_block_signals = 0;
1816	struct ocfs2_dentry_lock *dl = NULL;
1817
1818	trace_ocfs2_symlink_begin(dir, dentry, symname,
1819				  dentry->d_name.len, dentry->d_name.name);
1820
1821	status = dquot_initialize(dir);
1822	if (status) {
1823		mlog_errno(status);
1824		goto bail;
1825	}
1826
1827	sb = dir->i_sb;
1828	osb = OCFS2_SB(sb);
1829
1830	l = strlen(symname) + 1;
1831
1832	credits = ocfs2_calc_symlink_credits(sb);
1833
1834	/* lock the parent directory */
1835	status = ocfs2_inode_lock(dir, &parent_fe_bh, 1);
1836	if (status < 0) {
1837		if (status != -ENOENT)
1838			mlog_errno(status);
1839		return status;
1840	}
1841
1842	dirfe = (struct ocfs2_dinode *) parent_fe_bh->b_data;
1843	if (!ocfs2_read_links_count(dirfe)) {
1844		/* can't make a file in a deleted directory. */
1845		status = -ENOENT;
1846		goto bail;
1847	}
1848
1849	status = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
1850					   dentry->d_name.len);
1851	if (status)
1852		goto bail;
1853
1854	status = ocfs2_prepare_dir_for_insert(osb, dir, parent_fe_bh,
1855					      dentry->d_name.name,
1856					      dentry->d_name.len, &lookup);
1857	if (status < 0) {
1858		mlog_errno(status);
1859		goto bail;
1860	}
1861
1862	status = ocfs2_reserve_new_inode(osb, &inode_ac);
1863	if (status < 0) {
1864		if (status != -ENOSPC)
1865			mlog_errno(status);
1866		goto bail;
1867	}
1868
1869	inode = ocfs2_get_init_inode(dir, S_IFLNK | S_IRWXUGO);
1870	if (IS_ERR(inode)) {
1871		status = PTR_ERR(inode);
1872		inode = NULL;
1873		mlog_errno(status);
1874		goto bail;
1875	}
1876
1877	/* get security xattr */
1878	status = ocfs2_init_security_get(inode, dir, &dentry->d_name, &si);
1879	if (status) {
1880		if (status == -EOPNOTSUPP)
1881			si.enable = 0;
1882		else {
1883			mlog_errno(status);
1884			goto bail;
1885		}
1886	}
1887
1888	/* calculate meta data/clusters for setting security xattr */
1889	if (si.enable) {
1890		status = ocfs2_calc_security_init(dir, &si, &want_clusters,
1891						  &xattr_credits, &xattr_ac);
1892		if (status < 0) {
1893			mlog_errno(status);
1894			goto bail;
1895		}
1896	}
1897
1898	/* don't reserve bitmap space for fast symlinks. */
1899	if (l > ocfs2_fast_symlink_chars(sb))
1900		want_clusters += 1;
1901
1902	status = ocfs2_reserve_clusters(osb, want_clusters, &data_ac);
1903	if (status < 0) {
1904		if (status != -ENOSPC)
1905			mlog_errno(status);
1906		goto bail;
1907	}
1908
1909	handle = ocfs2_start_trans(osb, credits + xattr_credits);
1910	if (IS_ERR(handle)) {
1911		status = PTR_ERR(handle);
1912		handle = NULL;
1913		mlog_errno(status);
1914		goto bail;
1915	}
1916
1917	/* Starting to change things, restart is no longer possible. */
1918	ocfs2_block_signals(&oldset);
1919	did_block_signals = 1;
1920
1921	status = dquot_alloc_inode(inode);
1922	if (status)
1923		goto bail;
1924	did_quota_inode = 1;
1925
1926	trace_ocfs2_symlink_create(dir, dentry, dentry->d_name.len,
1927				   dentry->d_name.name,
1928				   (unsigned long long)OCFS2_I(dir)->ip_blkno,
1929				   inode->i_mode);
1930
1931	status = ocfs2_mknod_locked(osb, dir, inode,
1932				    0, &new_fe_bh, parent_fe_bh, handle,
1933				    inode_ac);
1934	if (status < 0) {
1935		mlog_errno(status);
1936		goto bail;
1937	}
1938
1939	fe = (struct ocfs2_dinode *) new_fe_bh->b_data;
1940	inode->i_rdev = 0;
1941	newsize = l - 1;
1942	inode->i_op = &ocfs2_symlink_inode_operations;
1943	inode_nohighmem(inode);
1944	if (l > ocfs2_fast_symlink_chars(sb)) {
1945		u32 offset = 0;
1946
1947		status = dquot_alloc_space_nodirty(inode,
1948		    ocfs2_clusters_to_bytes(osb->sb, 1));
1949		if (status)
1950			goto bail;
1951		did_quota = 1;
1952		inode->i_mapping->a_ops = &ocfs2_aops;
1953		status = ocfs2_add_inode_data(osb, inode, &offset, 1, 0,
1954					      new_fe_bh,
1955					      handle, data_ac, NULL,
1956					      NULL);
1957		if (status < 0) {
1958			if (status != -ENOSPC && status != -EINTR) {
1959				mlog(ML_ERROR,
1960				     "Failed to extend file to %llu\n",
1961				     (unsigned long long)newsize);
1962				mlog_errno(status);
1963				status = -ENOSPC;
1964			}
1965			goto bail;
1966		}
1967		i_size_write(inode, newsize);
1968		inode->i_blocks = ocfs2_inode_sector_count(inode);
1969	} else {
1970		inode->i_mapping->a_ops = &ocfs2_fast_symlink_aops;
1971		memcpy((char *) fe->id2.i_symlink, symname, l);
1972		i_size_write(inode, newsize);
1973		inode->i_blocks = 0;
1974	}
1975
1976	status = ocfs2_mark_inode_dirty(handle, inode, new_fe_bh);
1977	if (status < 0) {
1978		mlog_errno(status);
1979		goto bail;
1980	}
1981
1982	if (!ocfs2_inode_is_fast_symlink(inode)) {
1983		status = ocfs2_create_symlink_data(osb, handle, inode,
1984						   symname);
1985		if (status < 0) {
1986			mlog_errno(status);
1987			goto bail;
1988		}
1989	}
1990
1991	if (si.enable) {
1992		status = ocfs2_init_security_set(handle, inode, new_fe_bh, &si,
1993						 xattr_ac, data_ac);
1994		if (status < 0) {
1995			mlog_errno(status);
1996			goto bail;
1997		}
1998	}
1999
2000	/*
2001	 * Do this before adding the entry to the directory. We add
2002	 * also set d_op after success so that ->d_iput() will cleanup
2003	 * the dentry lock even if ocfs2_add_entry() fails below.
2004	 */
2005	status = ocfs2_dentry_attach_lock(dentry, inode, OCFS2_I(dir)->ip_blkno);
2006	if (status) {
2007		mlog_errno(status);
2008		goto bail;
2009	}
2010
2011	dl = dentry->d_fsdata;
2012
2013	status = ocfs2_add_entry(handle, dentry, inode,
2014				 le64_to_cpu(fe->i_blkno), parent_fe_bh,
2015				 &lookup);
2016	if (status < 0) {
2017		mlog_errno(status);
2018		goto bail;
2019	}
2020
2021	insert_inode_hash(inode);
2022	d_instantiate(dentry, inode);
2023bail:
2024	if (status < 0 && did_quota)
2025		dquot_free_space_nodirty(inode,
2026					ocfs2_clusters_to_bytes(osb->sb, 1));
2027	if (status < 0 && did_quota_inode)
2028		dquot_free_inode(inode);
2029	if (handle) {
2030		if (status < 0 && fe)
2031			ocfs2_set_links_count(fe, 0);
2032		ocfs2_commit_trans(osb, handle);
2033	}
2034
2035	ocfs2_inode_unlock(dir, 1);
2036	if (did_block_signals)
2037		ocfs2_unblock_signals(&oldset);
2038
2039	brelse(new_fe_bh);
2040	brelse(parent_fe_bh);
2041	kfree(si.value);
2042	ocfs2_free_dir_lookup_result(&lookup);
2043	if (inode_ac)
2044		ocfs2_free_alloc_context(inode_ac);
2045	if (data_ac)
2046		ocfs2_free_alloc_context(data_ac);
2047	if (xattr_ac)
2048		ocfs2_free_alloc_context(xattr_ac);
2049	if ((status < 0) && inode) {
2050		if (dl)
2051			ocfs2_cleanup_add_entry_failure(osb, dentry, inode);
2052
2053		OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR;
2054		clear_nlink(inode);
2055		iput(inode);
2056	}
2057
2058	if (status)
2059		mlog_errno(status);
2060
2061	return status;
2062}
2063
2064static int ocfs2_blkno_stringify(u64 blkno, char *name)
2065{
2066	int status, namelen;
2067
2068	namelen = snprintf(name, OCFS2_ORPHAN_NAMELEN + 1, "%016llx",
2069			   (long long)blkno);
2070	if (namelen <= 0) {
2071		if (namelen)
2072			status = namelen;
2073		else
2074			status = -EINVAL;
2075		mlog_errno(status);
2076		goto bail;
2077	}
2078	if (namelen != OCFS2_ORPHAN_NAMELEN) {
2079		status = -EINVAL;
2080		mlog_errno(status);
2081		goto bail;
2082	}
2083
2084	trace_ocfs2_blkno_stringify(blkno, name, namelen);
2085
2086	status = 0;
2087bail:
2088	if (status < 0)
2089		mlog_errno(status);
2090	return status;
2091}
2092
2093static int ocfs2_lookup_lock_orphan_dir(struct ocfs2_super *osb,
2094					struct inode **ret_orphan_dir,
2095					struct buffer_head **ret_orphan_dir_bh)
2096{
2097	struct inode *orphan_dir_inode;
2098	struct buffer_head *orphan_dir_bh = NULL;
2099	int ret = 0;
2100
2101	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
2102						       ORPHAN_DIR_SYSTEM_INODE,
2103						       osb->slot_num);
2104	if (!orphan_dir_inode) {
2105		ret = -ENOENT;
2106		mlog_errno(ret);
2107		return ret;
2108	}
2109
2110	inode_lock(orphan_dir_inode);
2111
2112	ret = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
2113	if (ret < 0) {
2114		inode_unlock(orphan_dir_inode);
2115		iput(orphan_dir_inode);
2116
2117		mlog_errno(ret);
2118		return ret;
2119	}
2120
2121	*ret_orphan_dir = orphan_dir_inode;
2122	*ret_orphan_dir_bh = orphan_dir_bh;
2123
2124	return 0;
2125}
2126
2127static int __ocfs2_prepare_orphan_dir(struct inode *orphan_dir_inode,
2128				      struct buffer_head *orphan_dir_bh,
2129				      u64 blkno,
2130				      char *name,
2131				      struct ocfs2_dir_lookup_result *lookup,
2132				      bool dio)
2133{
2134	int ret;
2135	struct ocfs2_super *osb = OCFS2_SB(orphan_dir_inode->i_sb);
2136	int namelen = dio ?
2137			(OCFS2_DIO_ORPHAN_PREFIX_LEN + OCFS2_ORPHAN_NAMELEN) :
2138			OCFS2_ORPHAN_NAMELEN;
2139
2140	if (dio) {
2141		ret = snprintf(name, OCFS2_DIO_ORPHAN_PREFIX_LEN + 1, "%s",
2142				OCFS2_DIO_ORPHAN_PREFIX);
2143		if (ret != OCFS2_DIO_ORPHAN_PREFIX_LEN) {
2144			ret = -EINVAL;
2145			mlog_errno(ret);
2146			return ret;
2147		}
2148
2149		ret = ocfs2_blkno_stringify(blkno,
2150				name + OCFS2_DIO_ORPHAN_PREFIX_LEN);
2151	} else
2152		ret = ocfs2_blkno_stringify(blkno, name);
2153	if (ret < 0) {
2154		mlog_errno(ret);
2155		return ret;
2156	}
2157
2158	ret = ocfs2_prepare_dir_for_insert(osb, orphan_dir_inode,
2159					   orphan_dir_bh, name,
2160					   namelen, lookup);
2161	if (ret < 0) {
2162		mlog_errno(ret);
2163		return ret;
2164	}
2165
2166	return 0;
2167}
2168
2169/**
2170 * ocfs2_prepare_orphan_dir() - Prepare an orphan directory for
2171 * insertion of an orphan.
2172 * @osb: ocfs2 file system
2173 * @ret_orphan_dir: Orphan dir inode - returned locked!
2174 * @blkno: Actual block number of the inode to be inserted into orphan dir.
2175 * @lookup: dir lookup result, to be passed back into functions like
2176 *          ocfs2_orphan_add
2177 *
2178 * Returns zero on success and the ret_orphan_dir, name and lookup
2179 * fields will be populated.
2180 *
2181 * Returns non-zero on failure.
2182 */
2183static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
2184				    struct inode **ret_orphan_dir,
2185				    u64 blkno,
2186				    char *name,
2187				    struct ocfs2_dir_lookup_result *lookup,
2188				    bool dio)
2189{
2190	struct inode *orphan_dir_inode = NULL;
2191	struct buffer_head *orphan_dir_bh = NULL;
2192	int ret = 0;
2193
2194	ret = ocfs2_lookup_lock_orphan_dir(osb, &orphan_dir_inode,
2195					   &orphan_dir_bh);
2196	if (ret < 0) {
2197		mlog_errno(ret);
2198		return ret;
2199	}
2200
2201	ret = __ocfs2_prepare_orphan_dir(orphan_dir_inode, orphan_dir_bh,
2202					 blkno, name, lookup, dio);
2203	if (ret < 0) {
2204		mlog_errno(ret);
2205		goto out;
2206	}
2207
2208	*ret_orphan_dir = orphan_dir_inode;
2209
2210out:
2211	brelse(orphan_dir_bh);
2212
2213	if (ret) {
2214		ocfs2_inode_unlock(orphan_dir_inode, 1);
2215		inode_unlock(orphan_dir_inode);
2216		iput(orphan_dir_inode);
2217	}
2218
2219	if (ret)
2220		mlog_errno(ret);
2221	return ret;
2222}
2223
2224static int ocfs2_orphan_add(struct ocfs2_super *osb,
2225			    handle_t *handle,
2226			    struct inode *inode,
2227			    struct buffer_head *fe_bh,
2228			    char *name,
2229			    struct ocfs2_dir_lookup_result *lookup,
2230			    struct inode *orphan_dir_inode,
2231			    bool dio)
2232{
2233	struct buffer_head *orphan_dir_bh = NULL;
2234	int status = 0;
2235	struct ocfs2_dinode *orphan_fe;
2236	struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
2237	int namelen = dio ?
2238			(OCFS2_DIO_ORPHAN_PREFIX_LEN + OCFS2_ORPHAN_NAMELEN) :
2239			OCFS2_ORPHAN_NAMELEN;
2240
2241	trace_ocfs2_orphan_add_begin(
2242				(unsigned long long)OCFS2_I(inode)->ip_blkno);
2243
2244	status = ocfs2_read_inode_block(orphan_dir_inode, &orphan_dir_bh);
2245	if (status < 0) {
2246		mlog_errno(status);
2247		goto leave;
2248	}
2249
2250	status = ocfs2_journal_access_di(handle,
2251					 INODE_CACHE(orphan_dir_inode),
2252					 orphan_dir_bh,
2253					 OCFS2_JOURNAL_ACCESS_WRITE);
2254	if (status < 0) {
2255		mlog_errno(status);
2256		goto leave;
2257	}
2258
2259	/*
2260	 * We're going to journal the change of i_flags and i_orphaned_slot.
2261	 * It's safe anyway, though some callers may duplicate the journaling.
2262	 * Journaling within the func just make the logic look more
2263	 * straightforward.
2264	 */
2265	status = ocfs2_journal_access_di(handle,
2266					 INODE_CACHE(inode),
2267					 fe_bh,
2268					 OCFS2_JOURNAL_ACCESS_WRITE);
2269	if (status < 0) {
2270		mlog_errno(status);
2271		goto leave;
2272	}
2273
2274	/* we're a cluster, and nlink can change on disk from
2275	 * underneath us... */
2276	orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data;
2277	if (S_ISDIR(inode->i_mode))
2278		ocfs2_add_links_count(orphan_fe, 1);
2279	set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe));
2280	ocfs2_journal_dirty(handle, orphan_dir_bh);
2281
2282	status = __ocfs2_add_entry(handle, orphan_dir_inode, name,
2283				   namelen, inode,
2284				   OCFS2_I(inode)->ip_blkno,
2285				   orphan_dir_bh, lookup);
2286	if (status < 0) {
2287		mlog_errno(status);
2288		goto rollback;
2289	}
2290
2291	if (dio) {
2292		/* Update flag OCFS2_DIO_ORPHANED_FL and record the orphan
2293		 * slot.
2294		 */
2295		fe->i_flags |= cpu_to_le32(OCFS2_DIO_ORPHANED_FL);
2296		fe->i_dio_orphaned_slot = cpu_to_le16(osb->slot_num);
2297	} else {
2298		fe->i_flags |= cpu_to_le32(OCFS2_ORPHANED_FL);
2299		OCFS2_I(inode)->ip_flags &= ~OCFS2_INODE_SKIP_ORPHAN_DIR;
2300
2301		/* Record which orphan dir our inode now resides
2302		 * in. delete_inode will use this to determine which orphan
2303		 * dir to lock. */
2304		fe->i_orphaned_slot = cpu_to_le16(osb->slot_num);
2305	}
2306
2307	ocfs2_journal_dirty(handle, fe_bh);
2308
2309	trace_ocfs2_orphan_add_end((unsigned long long)OCFS2_I(inode)->ip_blkno,
2310				   osb->slot_num);
2311
2312rollback:
2313	if (status < 0) {
2314		if (S_ISDIR(inode->i_mode))
2315			ocfs2_add_links_count(orphan_fe, -1);
2316		set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe));
2317	}
2318
2319leave:
2320	brelse(orphan_dir_bh);
2321
2322	return status;
2323}
2324
2325/* unlike orphan_add, we expect the orphan dir to already be locked here. */
2326int ocfs2_orphan_del(struct ocfs2_super *osb,
2327		     handle_t *handle,
2328		     struct inode *orphan_dir_inode,
2329		     struct inode *inode,
2330		     struct buffer_head *orphan_dir_bh,
2331		     bool dio)
2332{
2333	char name[OCFS2_DIO_ORPHAN_PREFIX_LEN + OCFS2_ORPHAN_NAMELEN + 1];
2334	struct ocfs2_dinode *orphan_fe;
2335	int status = 0;
2336	struct ocfs2_dir_lookup_result lookup = { NULL, };
2337
2338	if (dio) {
2339		status = snprintf(name, OCFS2_DIO_ORPHAN_PREFIX_LEN + 1, "%s",
2340				OCFS2_DIO_ORPHAN_PREFIX);
2341		if (status != OCFS2_DIO_ORPHAN_PREFIX_LEN) {
2342			status = -EINVAL;
2343			mlog_errno(status);
2344			return status;
2345		}
2346
2347		status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno,
2348				name + OCFS2_DIO_ORPHAN_PREFIX_LEN);
2349	} else
2350		status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, name);
2351	if (status < 0) {
2352		mlog_errno(status);
2353		goto leave;
2354	}
2355
2356	trace_ocfs2_orphan_del(
2357	     (unsigned long long)OCFS2_I(orphan_dir_inode)->ip_blkno,
2358	     name, strlen(name));
2359
2360	status = ocfs2_journal_access_di(handle,
2361					 INODE_CACHE(orphan_dir_inode),
2362					 orphan_dir_bh,
2363					 OCFS2_JOURNAL_ACCESS_WRITE);
2364	if (status < 0) {
2365		mlog_errno(status);
2366		goto leave;
2367	}
2368
2369	/* find it's spot in the orphan directory */
2370	status = ocfs2_find_entry(name, strlen(name), orphan_dir_inode,
2371				  &lookup);
2372	if (status) {
2373		mlog_errno(status);
2374		goto leave;
2375	}
2376
2377	/* remove it from the orphan directory */
2378	status = ocfs2_delete_entry(handle, orphan_dir_inode, &lookup);
2379	if (status < 0) {
2380		mlog_errno(status);
2381		goto leave;
2382	}
2383
2384	/* do the i_nlink dance! :) */
2385	orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data;
2386	if (S_ISDIR(inode->i_mode))
2387		ocfs2_add_links_count(orphan_fe, -1);
2388	set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe));
2389	ocfs2_journal_dirty(handle, orphan_dir_bh);
2390
2391leave:
2392	ocfs2_free_dir_lookup_result(&lookup);
2393
2394	if (status)
2395		mlog_errno(status);
2396	return status;
2397}
2398
2399/**
2400 * ocfs2_prep_new_orphaned_file() - Prepare the orphan dir to receive a newly
2401 * allocated file. This is different from the typical 'add to orphan dir'
2402 * operation in that the inode does not yet exist. This is a problem because
2403 * the orphan dir stringifies the inode block number to come up with it's
2404 * dirent. Obviously if the inode does not yet exist we have a chicken and egg
2405 * problem. This function works around it by calling deeper into the orphan
2406 * and suballoc code than other callers. Use this only by necessity.
2407 * @dir: The directory which this inode will ultimately wind up under - not the
2408 * orphan dir!
2409 * @dir_bh: buffer_head the @dir inode block
2410 * @orphan_name: string of length (CFS2_ORPHAN_NAMELEN + 1). Will be filled
2411 * with the string to be used for orphan dirent. Pass back to the orphan dir
2412 * code.
2413 * @ret_orphan_dir: orphan dir inode returned to be passed back into orphan
2414 * dir code.
2415 * @ret_di_blkno: block number where the new inode will be allocated.
2416 * @orphan_insert: Dir insert context to be passed back into orphan dir code.
2417 * @ret_inode_ac: Inode alloc context to be passed back to the allocator.
2418 *
2419 * Returns zero on success and the ret_orphan_dir, name and lookup
2420 * fields will be populated.
2421 *
2422 * Returns non-zero on failure.
2423 */
2424static int ocfs2_prep_new_orphaned_file(struct inode *dir,
2425					struct buffer_head *dir_bh,
2426					char *orphan_name,
2427					struct inode **ret_orphan_dir,
2428					u64 *ret_di_blkno,
2429					struct ocfs2_dir_lookup_result *orphan_insert,
2430					struct ocfs2_alloc_context **ret_inode_ac)
2431{
2432	int ret;
2433	u64 di_blkno;
2434	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
2435	struct inode *orphan_dir = NULL;
2436	struct buffer_head *orphan_dir_bh = NULL;
2437	struct ocfs2_alloc_context *inode_ac = NULL;
2438
2439	ret = ocfs2_lookup_lock_orphan_dir(osb, &orphan_dir, &orphan_dir_bh);
2440	if (ret < 0) {
2441		mlog_errno(ret);
2442		return ret;
2443	}
2444
2445	/* reserve an inode spot */
2446	ret = ocfs2_reserve_new_inode(osb, &inode_ac);
2447	if (ret < 0) {
2448		if (ret != -ENOSPC)
2449			mlog_errno(ret);
2450		goto out;
2451	}
2452
2453	ret = ocfs2_find_new_inode_loc(dir, dir_bh, inode_ac,
2454				       &di_blkno);
2455	if (ret) {
2456		mlog_errno(ret);
2457		goto out;
2458	}
2459
2460	ret = __ocfs2_prepare_orphan_dir(orphan_dir, orphan_dir_bh,
2461					 di_blkno, orphan_name, orphan_insert,
2462					 false);
2463	if (ret < 0) {
2464		mlog_errno(ret);
2465		goto out;
2466	}
2467
2468out:
2469	if (ret == 0) {
2470		*ret_orphan_dir = orphan_dir;
2471		*ret_di_blkno = di_blkno;
2472		*ret_inode_ac = inode_ac;
2473		/*
2474		 * orphan_name and orphan_insert are already up to
2475		 * date via prepare_orphan_dir
2476		 */
2477	} else {
2478		/* Unroll reserve_new_inode* */
2479		if (inode_ac)
2480			ocfs2_free_alloc_context(inode_ac);
2481
2482		/* Unroll orphan dir locking */
2483		inode_unlock(orphan_dir);
2484		ocfs2_inode_unlock(orphan_dir, 1);
2485		iput(orphan_dir);
2486	}
2487
2488	brelse(orphan_dir_bh);
2489
2490	return ret;
2491}
2492
2493int ocfs2_create_inode_in_orphan(struct inode *dir,
2494				 int mode,
2495				 struct inode **new_inode)
2496{
2497	int status, did_quota_inode = 0;
2498	struct inode *inode = NULL;
2499	struct inode *orphan_dir = NULL;
2500	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
2501	handle_t *handle = NULL;
2502	char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
2503	struct buffer_head *parent_di_bh = NULL;
2504	struct buffer_head *new_di_bh = NULL;
2505	struct ocfs2_alloc_context *inode_ac = NULL;
2506	struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
2507	u64 di_blkno, suballoc_loc;
2508	u16 suballoc_bit;
2509
2510	status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
2511	if (status < 0) {
2512		if (status != -ENOENT)
2513			mlog_errno(status);
2514		return status;
2515	}
2516
2517	status = ocfs2_prep_new_orphaned_file(dir, parent_di_bh,
2518					      orphan_name, &orphan_dir,
2519					      &di_blkno, &orphan_insert, &inode_ac);
2520	if (status < 0) {
2521		if (status != -ENOSPC)
2522			mlog_errno(status);
2523		goto leave;
2524	}
2525
2526	inode = ocfs2_get_init_inode(dir, mode);
2527	if (IS_ERR(inode)) {
2528		status = PTR_ERR(inode);
2529		inode = NULL;
2530		mlog_errno(status);
2531		goto leave;
2532	}
2533
2534	handle = ocfs2_start_trans(osb, ocfs2_mknod_credits(osb->sb, 0, 0));
2535	if (IS_ERR(handle)) {
2536		status = PTR_ERR(handle);
2537		handle = NULL;
2538		mlog_errno(status);
2539		goto leave;
2540	}
2541
2542	status = dquot_alloc_inode(inode);
2543	if (status)
2544		goto leave;
2545	did_quota_inode = 1;
2546
2547	status = ocfs2_claim_new_inode_at_loc(handle, dir, inode_ac,
2548					      &suballoc_loc,
2549					      &suballoc_bit, di_blkno);
2550	if (status < 0) {
2551		mlog_errno(status);
2552		goto leave;
2553	}
2554
2555	clear_nlink(inode);
2556	/* do the real work now. */
2557	status = __ocfs2_mknod_locked(dir, inode,
2558				      0, &new_di_bh, parent_di_bh, handle,
2559				      inode_ac, di_blkno, suballoc_loc,
2560				      suballoc_bit);
2561	if (status < 0) {
2562		mlog_errno(status);
2563		goto leave;
2564	}
2565
2566	status = ocfs2_orphan_add(osb, handle, inode, new_di_bh, orphan_name,
2567				  &orphan_insert, orphan_dir, false);
2568	if (status < 0) {
2569		mlog_errno(status);
2570		goto leave;
2571	}
2572
2573	/* get open lock so that only nodes can't remove it from orphan dir. */
2574	status = ocfs2_open_lock(inode);
2575	if (status < 0)
2576		mlog_errno(status);
2577
2578	insert_inode_hash(inode);
2579leave:
2580	if (status < 0 && did_quota_inode)
2581		dquot_free_inode(inode);
2582	if (handle)
2583		ocfs2_commit_trans(osb, handle);
2584
2585	if (orphan_dir) {
2586		/* This was locked for us in ocfs2_prepare_orphan_dir() */
2587		ocfs2_inode_unlock(orphan_dir, 1);
2588		inode_unlock(orphan_dir);
2589		iput(orphan_dir);
2590	}
2591
2592	if ((status < 0) && inode) {
2593		clear_nlink(inode);
2594		iput(inode);
2595	}
2596
2597	if (inode_ac)
2598		ocfs2_free_alloc_context(inode_ac);
2599
2600	brelse(new_di_bh);
2601
2602	if (!status)
2603		*new_inode = inode;
2604
2605	ocfs2_free_dir_lookup_result(&orphan_insert);
2606
2607	ocfs2_inode_unlock(dir, 1);
2608	brelse(parent_di_bh);
2609	return status;
2610}
2611
2612int ocfs2_add_inode_to_orphan(struct ocfs2_super *osb,
2613	struct inode *inode)
2614{
2615	char orphan_name[OCFS2_DIO_ORPHAN_PREFIX_LEN + OCFS2_ORPHAN_NAMELEN + 1];
2616	struct inode *orphan_dir_inode = NULL;
2617	struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
2618	struct buffer_head *di_bh = NULL;
2619	int status = 0;
2620	handle_t *handle = NULL;
2621	struct ocfs2_dinode *di = NULL;
2622
2623	status = ocfs2_inode_lock(inode, &di_bh, 1);
2624	if (status < 0) {
2625		mlog_errno(status);
2626		goto bail;
2627	}
2628
2629	di = (struct ocfs2_dinode *) di_bh->b_data;
2630	/*
2631	 * Another append dio crashed?
2632	 * If so, manually recover it first.
2633	 */
2634	if (unlikely(di->i_flags & cpu_to_le32(OCFS2_DIO_ORPHANED_FL))) {
2635		status = ocfs2_truncate_file(inode, di_bh, i_size_read(inode));
2636		if (status < 0) {
2637			if (status != -ENOSPC)
2638				mlog_errno(status);
2639			goto bail_unlock_inode;
2640		}
2641
2642		status = ocfs2_del_inode_from_orphan(osb, inode, di_bh, 0, 0);
2643		if (status < 0) {
2644			mlog_errno(status);
2645			goto bail_unlock_inode;
2646		}
2647	}
2648
2649	status = ocfs2_prepare_orphan_dir(osb, &orphan_dir_inode,
2650			OCFS2_I(inode)->ip_blkno,
2651			orphan_name,
2652			&orphan_insert,
2653			true);
2654	if (status < 0) {
2655		mlog_errno(status);
2656		goto bail_unlock_inode;
2657	}
2658
2659	handle = ocfs2_start_trans(osb,
2660			OCFS2_INODE_ADD_TO_ORPHAN_CREDITS);
2661	if (IS_ERR(handle)) {
2662		status = PTR_ERR(handle);
2663		goto bail_unlock_orphan;
2664	}
2665
2666	status = ocfs2_orphan_add(osb, handle, inode, di_bh, orphan_name,
2667			&orphan_insert, orphan_dir_inode, true);
2668	if (status)
2669		mlog_errno(status);
2670
2671	ocfs2_commit_trans(osb, handle);
2672
2673bail_unlock_orphan:
2674	ocfs2_inode_unlock(orphan_dir_inode, 1);
2675	inode_unlock(orphan_dir_inode);
2676	iput(orphan_dir_inode);
2677
2678	ocfs2_free_dir_lookup_result(&orphan_insert);
2679
2680bail_unlock_inode:
2681	ocfs2_inode_unlock(inode, 1);
2682	brelse(di_bh);
2683
2684bail:
2685	return status;
2686}
2687
2688int ocfs2_del_inode_from_orphan(struct ocfs2_super *osb,
2689		struct inode *inode, struct buffer_head *di_bh,
2690		int update_isize, loff_t end)
2691{
2692	struct inode *orphan_dir_inode = NULL;
2693	struct buffer_head *orphan_dir_bh = NULL;
2694	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
2695	handle_t *handle = NULL;
2696	int status = 0;
2697
2698	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
2699			ORPHAN_DIR_SYSTEM_INODE,
2700			le16_to_cpu(di->i_dio_orphaned_slot));
2701	if (!orphan_dir_inode) {
2702		status = -ENOENT;
2703		mlog_errno(status);
2704		goto bail;
2705	}
2706
2707	inode_lock(orphan_dir_inode);
2708	status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
2709	if (status < 0) {
2710		inode_unlock(orphan_dir_inode);
2711		iput(orphan_dir_inode);
2712		mlog_errno(status);
2713		goto bail;
2714	}
2715
2716	handle = ocfs2_start_trans(osb,
2717			OCFS2_INODE_DEL_FROM_ORPHAN_CREDITS);
2718	if (IS_ERR(handle)) {
2719		status = PTR_ERR(handle);
2720		goto bail_unlock_orphan;
2721	}
2722
2723	BUG_ON(!(di->i_flags & cpu_to_le32(OCFS2_DIO_ORPHANED_FL)));
2724
2725	status = ocfs2_orphan_del(osb, handle, orphan_dir_inode,
2726				inode, orphan_dir_bh, true);
2727	if (status < 0) {
2728		mlog_errno(status);
2729		goto bail_commit;
2730	}
2731
2732	status = ocfs2_journal_access_di(handle,
2733			INODE_CACHE(inode),
2734			di_bh,
2735			OCFS2_JOURNAL_ACCESS_WRITE);
2736	if (status < 0) {
2737		mlog_errno(status);
2738		goto bail_commit;
2739	}
2740
2741	di->i_flags &= ~cpu_to_le32(OCFS2_DIO_ORPHANED_FL);
2742	di->i_dio_orphaned_slot = 0;
2743
2744	if (update_isize) {
2745		status = ocfs2_set_inode_size(handle, inode, di_bh, end);
2746		if (status)
2747			mlog_errno(status);
2748	} else
2749		ocfs2_journal_dirty(handle, di_bh);
2750
2751bail_commit:
2752	ocfs2_commit_trans(osb, handle);
2753
2754bail_unlock_orphan:
2755	ocfs2_inode_unlock(orphan_dir_inode, 1);
2756	inode_unlock(orphan_dir_inode);
2757	brelse(orphan_dir_bh);
2758	iput(orphan_dir_inode);
2759
2760bail:
2761	return status;
2762}
2763
2764int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
2765				   struct inode *inode,
2766				   struct dentry *dentry)
2767{
2768	int status = 0;
2769	struct buffer_head *parent_di_bh = NULL;
2770	handle_t *handle = NULL;
2771	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
2772	struct ocfs2_dinode *dir_di, *di;
2773	struct inode *orphan_dir_inode = NULL;
2774	struct buffer_head *orphan_dir_bh = NULL;
2775	struct buffer_head *di_bh = NULL;
2776	struct ocfs2_dir_lookup_result lookup = { NULL, };
2777
2778	trace_ocfs2_mv_orphaned_inode_to_new(dir, dentry,
2779				dentry->d_name.len, dentry->d_name.name,
2780				(unsigned long long)OCFS2_I(dir)->ip_blkno,
2781				(unsigned long long)OCFS2_I(inode)->ip_blkno);
2782
2783	status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
2784	if (status < 0) {
2785		if (status != -ENOENT)
2786			mlog_errno(status);
2787		return status;
2788	}
2789
2790	dir_di = (struct ocfs2_dinode *) parent_di_bh->b_data;
2791	if (!dir_di->i_links_count) {
2792		/* can't make a file in a deleted directory. */
2793		status = -ENOENT;
2794		goto leave;
2795	}
2796
2797	status = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
2798					   dentry->d_name.len);
2799	if (status)
2800		goto leave;
2801
2802	/* get a spot inside the dir. */
2803	status = ocfs2_prepare_dir_for_insert(osb, dir, parent_di_bh,
2804					      dentry->d_name.name,
2805					      dentry->d_name.len, &lookup);
2806	if (status < 0) {
2807		mlog_errno(status);
2808		goto leave;
2809	}
2810
2811	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
2812						       ORPHAN_DIR_SYSTEM_INODE,
2813						       osb->slot_num);
2814	if (!orphan_dir_inode) {
2815		status = -ENOENT;
2816		mlog_errno(status);
2817		goto leave;
2818	}
2819
2820	inode_lock(orphan_dir_inode);
2821
2822	status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
2823	if (status < 0) {
2824		mlog_errno(status);
2825		inode_unlock(orphan_dir_inode);
2826		iput(orphan_dir_inode);
2827		goto leave;
2828	}
2829
2830	status = ocfs2_read_inode_block(inode, &di_bh);
2831	if (status < 0) {
2832		mlog_errno(status);
2833		goto orphan_unlock;
2834	}
2835
2836	handle = ocfs2_start_trans(osb, ocfs2_rename_credits(osb->sb));
2837	if (IS_ERR(handle)) {
2838		status = PTR_ERR(handle);
2839		handle = NULL;
2840		mlog_errno(status);
2841		goto orphan_unlock;
2842	}
2843
2844	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
2845					 di_bh, OCFS2_JOURNAL_ACCESS_WRITE);
2846	if (status < 0) {
2847		mlog_errno(status);
2848		goto out_commit;
2849	}
2850
2851	status = ocfs2_orphan_del(osb, handle, orphan_dir_inode, inode,
2852				  orphan_dir_bh, false);
2853	if (status < 0) {
2854		mlog_errno(status);
2855		goto out_commit;
2856	}
2857
2858	di = (struct ocfs2_dinode *)di_bh->b_data;
2859	di->i_flags &= ~cpu_to_le32(OCFS2_ORPHANED_FL);
2860	di->i_orphaned_slot = 0;
2861	set_nlink(inode, 1);
2862	ocfs2_set_links_count(di, inode->i_nlink);
2863	ocfs2_update_inode_fsync_trans(handle, inode, 1);
2864	ocfs2_journal_dirty(handle, di_bh);
2865
2866	status = ocfs2_add_entry(handle, dentry, inode,
2867				 OCFS2_I(inode)->ip_blkno, parent_di_bh,
2868				 &lookup);
2869	if (status < 0) {
2870		mlog_errno(status);
2871		goto out_commit;
2872	}
2873
2874	status = ocfs2_dentry_attach_lock(dentry, inode,
2875					  OCFS2_I(dir)->ip_blkno);
2876	if (status) {
2877		mlog_errno(status);
2878		goto out_commit;
2879	}
2880
2881	d_instantiate(dentry, inode);
2882	status = 0;
2883out_commit:
2884	ocfs2_commit_trans(osb, handle);
2885orphan_unlock:
2886	ocfs2_inode_unlock(orphan_dir_inode, 1);
2887	inode_unlock(orphan_dir_inode);
2888	iput(orphan_dir_inode);
2889leave:
2890
2891	ocfs2_inode_unlock(dir, 1);
2892
2893	brelse(di_bh);
2894	brelse(parent_di_bh);
2895	brelse(orphan_dir_bh);
2896
2897	ocfs2_free_dir_lookup_result(&lookup);
2898
2899	if (status)
2900		mlog_errno(status);
2901
2902	return status;
2903}
2904
2905const struct inode_operations ocfs2_dir_iops = {
2906	.create		= ocfs2_create,
2907	.lookup		= ocfs2_lookup,
2908	.link		= ocfs2_link,
2909	.unlink		= ocfs2_unlink,
2910	.rmdir		= ocfs2_unlink,
2911	.symlink	= ocfs2_symlink,
2912	.mkdir		= ocfs2_mkdir,
2913	.mknod		= ocfs2_mknod,
2914	.rename		= ocfs2_rename,
2915	.setattr	= ocfs2_setattr,
2916	.getattr	= ocfs2_getattr,
2917	.permission	= ocfs2_permission,
2918	.listxattr	= ocfs2_listxattr,
2919	.fiemap         = ocfs2_fiemap,
2920	.get_acl	= ocfs2_iop_get_acl,
2921	.set_acl	= ocfs2_iop_set_acl,
2922};
2923