4 * Complete reimplementation
5 * (C) 1997 Thomas Schoebel-Theuer,
6 * with heavy changes by Linus Torvalds
10 * Notes on the allocation strategy:
12 * The dcache is a master of the icache - whenever a dcache entry
13 * exists, the inode will always exist. "iput()" is done either when
14 * the dcache entry is deleted or garbage collected.
17 #include <linux/syscalls.h>
18 #include <linux/string.h>
21 #include <linux/fsnotify.h>
22 #include <linux/slab.h>
23 #include <linux/init.h>
24 #include <linux/hash.h>
25 #include <linux/cache.h>
26 #include <linux/export.h>
27 #include <linux/mount.h>
28 #include <linux/file.h>
29 #include <asm/uaccess.h>
30 #include <linux/security.h>
31 #include <linux/seqlock.h>
32 #include <linux/swap.h>
33 #include <linux/bootmem.h>
34 #include <linux/fs_struct.h>
35 #include <linux/hardirq.h>
36 #include <linux/bit_spinlock.h>
37 #include <linux/rculist_bl.h>
38 #include <linux/prefetch.h>
39 #include <linux/ratelimit.h>
45 * dcache->d_inode->i_lock protects:
46 * - i_dentry, d_alias, d_inode of aliases
47 * dcache_hash_bucket lock protects:
48 * - the dcache hash table
49 * s_anon bl list spinlock protects:
50 * - the s_anon list (see __d_drop)
51 * dcache_lru_lock protects:
52 * - the dcache lru lists and counters
59 * - d_parent and d_subdirs
60 * - childrens' d_child and d_parent
64 * dentry->d_inode->i_lock
67 * dcache_hash_bucket lock
70 * If there is an ancestor relationship:
71 * dentry->d_parent->...->d_parent->d_lock
73 * dentry->d_parent->d_lock
76 * If no ancestor relationship:
77 * if (dentry1 < dentry2)
81 int sysctl_vfs_cache_pressure __read_mostly
= 100;
82 EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure
);
84 static __cacheline_aligned_in_smp
DEFINE_SPINLOCK(dcache_lru_lock
);
85 __cacheline_aligned_in_smp
DEFINE_SEQLOCK(rename_lock
);
87 EXPORT_SYMBOL(rename_lock
);
89 static struct kmem_cache
*dentry_cache __read_mostly
;
92 * This is the single most critical data structure when it comes
93 * to the dcache: the hashtable for lookups. Somebody should try
94 * to make this good - I've just made it work.
96 * This hash-function tries to avoid losing too many bits of hash
97 * information, yet avoid using a prime hash-size or similar.
99 #define D_HASHBITS d_hash_shift
100 #define D_HASHMASK d_hash_mask
102 static unsigned int d_hash_mask __read_mostly
;
103 static unsigned int d_hash_shift __read_mostly
;
105 static struct hlist_bl_head
*dentry_hashtable __read_mostly
;
107 static inline struct hlist_bl_head
*d_hash(const struct dentry
*parent
,
110 hash
+= (unsigned long) parent
/ L1_CACHE_BYTES
;
111 hash
= hash
+ (hash
>> D_HASHBITS
);
112 return dentry_hashtable
+ (hash
& D_HASHMASK
);
115 /* Statistics gathering. */
116 struct dentry_stat_t dentry_stat
= {
120 static DEFINE_PER_CPU(unsigned int, nr_dentry
);
122 #if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
123 static int get_nr_dentry(void)
127 for_each_possible_cpu(i
)
128 sum
+= per_cpu(nr_dentry
, i
);
129 return sum
< 0 ? 0 : sum
;
132 int proc_nr_dentry(ctl_table
*table
, int write
, void __user
*buffer
,
133 size_t *lenp
, loff_t
*ppos
)
135 dentry_stat
.nr_dentry
= get_nr_dentry();
136 return proc_dointvec(table
, write
, buffer
, lenp
, ppos
);
141 * Compare 2 name strings, return 0 if they match, otherwise non-zero.
142 * The strings are both count bytes long, and count is non-zero.
144 #ifdef CONFIG_DCACHE_WORD_ACCESS
146 #include <asm/word-at-a-time.h>
148 * NOTE! 'cs' and 'scount' come from a dentry, so it has a
149 * aligned allocation for this particular component. We don't
150 * strictly need the load_unaligned_zeropad() safety, but it
151 * doesn't hurt either.
153 * In contrast, 'ct' and 'tcount' can be from a pathname, and do
154 * need the careful unaligned handling.
156 static inline int dentry_string_cmp(const unsigned char *cs
, const unsigned char *ct
, unsigned tcount
)
158 unsigned long a
,b
,mask
;
161 a
= *(unsigned long *)cs
;
162 b
= load_unaligned_zeropad(ct
);
163 if (tcount
< sizeof(unsigned long))
165 if (unlikely(a
!= b
))
167 cs
+= sizeof(unsigned long);
168 ct
+= sizeof(unsigned long);
169 tcount
-= sizeof(unsigned long);
173 mask
= ~(~0ul << tcount
*8);
174 return unlikely(!!((a
^ b
) & mask
));
179 static inline int dentry_string_cmp(const unsigned char *cs
, const unsigned char *ct
, unsigned tcount
)
193 static inline int dentry_cmp(const struct dentry
*dentry
, const unsigned char *ct
, unsigned tcount
)
195 const unsigned char *cs
;
197 * Be careful about RCU walk racing with rename:
198 * use ACCESS_ONCE to fetch the name pointer.
200 * NOTE! Even if a rename will mean that the length
201 * was not loaded atomically, we don't care. The
202 * RCU walk will check the sequence count eventually,
203 * and catch it. And we won't overrun the buffer,
204 * because we're reading the name pointer atomically,
205 * and a dentry name is guaranteed to be properly
206 * terminated with a NUL byte.
208 * End result: even if 'len' is wrong, we'll exit
209 * early because the data cannot match (there can
210 * be no NUL in the ct/tcount data)
212 cs
= ACCESS_ONCE(dentry
->d_name
.name
);
213 smp_read_barrier_depends();
214 return dentry_string_cmp(cs
, ct
, tcount
);
217 static void __d_free(struct rcu_head
*head
)
219 struct dentry
*dentry
= container_of(head
, struct dentry
, d_u
.d_rcu
);
221 WARN_ON(!hlist_unhashed(&dentry
->d_alias
));
222 if (dname_external(dentry
))
223 kfree(dentry
->d_name
.name
);
224 kmem_cache_free(dentry_cache
, dentry
);
230 static void d_free(struct dentry
*dentry
)
232 BUG_ON(dentry
->d_lockref
.count
);
233 this_cpu_dec(nr_dentry
);
234 if (dentry
->d_op
&& dentry
->d_op
->d_release
)
235 dentry
->d_op
->d_release(dentry
);
237 /* if dentry was never visible to RCU, immediate free is OK */
238 if (!(dentry
->d_flags
& DCACHE_RCUACCESS
))
239 __d_free(&dentry
->d_u
.d_rcu
);
241 call_rcu(&dentry
->d_u
.d_rcu
, __d_free
);
245 * dentry_rcuwalk_barrier - invalidate in-progress rcu-walk lookups
246 * @dentry: the target dentry
247 * After this call, in-progress rcu-walk path lookup will fail. This
248 * should be called after unhashing, and after changing d_inode (if
249 * the dentry has not already been unhashed).
251 static inline void dentry_rcuwalk_barrier(struct dentry
*dentry
)
253 assert_spin_locked(&dentry
->d_lock
);
254 /* Go through a barrier */
255 write_seqcount_barrier(&dentry
->d_seq
);
259 * Release the dentry's inode, using the filesystem
260 * d_iput() operation if defined. Dentry has no refcount
263 static void dentry_iput(struct dentry
* dentry
)
264 __releases(dentry
->d_lock
)
265 __releases(dentry
->d_inode
->i_lock
)
267 struct inode
*inode
= dentry
->d_inode
;
269 dentry
->d_inode
= NULL
;
270 hlist_del_init(&dentry
->d_alias
);
271 spin_unlock(&dentry
->d_lock
);
272 spin_unlock(&inode
->i_lock
);
274 fsnotify_inoderemove(inode
);
275 if (dentry
->d_op
&& dentry
->d_op
->d_iput
)
276 dentry
->d_op
->d_iput(dentry
, inode
);
280 spin_unlock(&dentry
->d_lock
);
285 * Release the dentry's inode, using the filesystem
286 * d_iput() operation if defined. dentry remains in-use.
288 static void dentry_unlink_inode(struct dentry
* dentry
)
289 __releases(dentry
->d_lock
)
290 __releases(dentry
->d_inode
->i_lock
)
292 struct inode
*inode
= dentry
->d_inode
;
293 dentry
->d_inode
= NULL
;
294 hlist_del_init(&dentry
->d_alias
);
295 dentry_rcuwalk_barrier(dentry
);
296 spin_unlock(&dentry
->d_lock
);
297 spin_unlock(&inode
->i_lock
);
299 fsnotify_inoderemove(inode
);
300 if (dentry
->d_op
&& dentry
->d_op
->d_iput
)
301 dentry
->d_op
->d_iput(dentry
, inode
);
307 * dentry_lru_(add|del|prune|move_tail) must be called with d_lock held.
309 static void dentry_lru_add(struct dentry
*dentry
)
311 if (list_empty(&dentry
->d_lru
)) {
312 spin_lock(&dcache_lru_lock
);
313 list_add(&dentry
->d_lru
, &dentry
->d_sb
->s_dentry_lru
);
314 dentry
->d_sb
->s_nr_dentry_unused
++;
315 dentry_stat
.nr_unused
++;
316 spin_unlock(&dcache_lru_lock
);
320 static void __dentry_lru_del(struct dentry
*dentry
)
322 list_del_init(&dentry
->d_lru
);
323 dentry
->d_flags
&= ~DCACHE_SHRINK_LIST
;
324 dentry
->d_sb
->s_nr_dentry_unused
--;
325 dentry_stat
.nr_unused
--;
329 * Remove a dentry with references from the LRU.
331 static void dentry_lru_del(struct dentry
*dentry
)
333 if (!list_empty(&dentry
->d_lru
)) {
334 spin_lock(&dcache_lru_lock
);
335 __dentry_lru_del(dentry
);
336 spin_unlock(&dcache_lru_lock
);
340 static void dentry_lru_move_list(struct dentry
*dentry
, struct list_head
*list
)
342 spin_lock(&dcache_lru_lock
);
343 if (list_empty(&dentry
->d_lru
)) {
344 list_add_tail(&dentry
->d_lru
, list
);
345 dentry
->d_sb
->s_nr_dentry_unused
++;
346 dentry_stat
.nr_unused
++;
348 list_move_tail(&dentry
->d_lru
, list
);
350 spin_unlock(&dcache_lru_lock
);
354 * d_kill - kill dentry and return parent
355 * @dentry: dentry to kill
356 * @parent: parent dentry
358 * The dentry must already be unhashed and removed from the LRU.
360 * If this is the root of the dentry tree, return NULL.
362 * dentry->d_lock and parent->d_lock must be held by caller, and are dropped by
365 static struct dentry
*d_kill(struct dentry
*dentry
, struct dentry
*parent
)
366 __releases(dentry
->d_lock
)
367 __releases(parent
->d_lock
)
368 __releases(dentry
->d_inode
->i_lock
)
370 list_del(&dentry
->d_u
.d_child
);
372 * Inform try_to_ascend() that we are no longer attached to the
375 dentry
->d_flags
|= DCACHE_DENTRY_KILLED
;
377 spin_unlock(&parent
->d_lock
);
380 * dentry_iput drops the locks, at which point nobody (except
381 * transient RCU lookups) can reach this dentry.
388 * Unhash a dentry without inserting an RCU walk barrier or checking that
389 * dentry->d_lock is locked. The caller must take care of that, if
392 static void __d_shrink(struct dentry
*dentry
)
394 if (!d_unhashed(dentry
)) {
395 struct hlist_bl_head
*b
;
396 if (unlikely(dentry
->d_flags
& DCACHE_DISCONNECTED
))
397 b
= &dentry
->d_sb
->s_anon
;
399 b
= d_hash(dentry
->d_parent
, dentry
->d_name
.hash
);
402 __hlist_bl_del(&dentry
->d_hash
);
403 dentry
->d_hash
.pprev
= NULL
;
409 * d_drop - drop a dentry
410 * @dentry: dentry to drop
412 * d_drop() unhashes the entry from the parent dentry hashes, so that it won't
413 * be found through a VFS lookup any more. Note that this is different from
414 * deleting the dentry - d_delete will try to mark the dentry negative if
415 * possible, giving a successful _negative_ lookup, while d_drop will
416 * just make the cache lookup fail.
418 * d_drop() is used mainly for stuff that wants to invalidate a dentry for some
419 * reason (NFS timeouts or autofs deletes).
421 * __d_drop requires dentry->d_lock.
423 void __d_drop(struct dentry
*dentry
)
425 if (!d_unhashed(dentry
)) {
427 dentry_rcuwalk_barrier(dentry
);
430 EXPORT_SYMBOL(__d_drop
);
432 void d_drop(struct dentry
*dentry
)
434 spin_lock(&dentry
->d_lock
);
436 spin_unlock(&dentry
->d_lock
);
438 EXPORT_SYMBOL(d_drop
);
441 * Finish off a dentry we've decided to kill.
442 * dentry->d_lock must be held, returns with it unlocked.
443 * If ref is non-zero, then decrement the refcount too.
444 * Returns dentry requiring refcount drop, or NULL if we're done.
446 static inline struct dentry
*dentry_kill(struct dentry
*dentry
, int ref
)
447 __releases(dentry
->d_lock
)
450 struct dentry
*parent
;
452 inode
= dentry
->d_inode
;
453 if (inode
&& !spin_trylock(&inode
->i_lock
)) {
455 spin_unlock(&dentry
->d_lock
);
457 return dentry
; /* try again with same dentry */
462 parent
= dentry
->d_parent
;
463 if (parent
&& !spin_trylock(&parent
->d_lock
)) {
465 spin_unlock(&inode
->i_lock
);
470 dentry
->d_lockref
.count
--;
472 * inform the fs via d_prune that this dentry is about to be
473 * unhashed and destroyed.
475 if (dentry
->d_flags
& DCACHE_OP_PRUNE
)
476 dentry
->d_op
->d_prune(dentry
);
478 dentry_lru_del(dentry
);
479 /* if it was on the hash then remove it */
481 return d_kill(dentry
, parent
);
487 * This is complicated by the fact that we do not want to put
488 * dentries that are no longer on any hash chain on the unused
489 * list: we'd much rather just get rid of them immediately.
491 * However, that implies that we have to traverse the dentry
492 * tree upwards to the parents which might _also_ now be
493 * scheduled for deletion (it may have been only waiting for
494 * its last child to go away).
496 * This tail recursion is done by hand as we don't want to depend
497 * on the compiler to always get this right (gcc generally doesn't).
498 * Real recursion would eat up our stack space.
502 * dput - release a dentry
503 * @dentry: dentry to release
505 * Release a dentry. This will drop the usage count and if appropriate
506 * call the dentry unlink method as well as removing it from the queues and
507 * releasing its resources. If the parent dentries were scheduled for release
508 * they too may now get deleted.
510 void dput(struct dentry
*dentry
)
516 if (dentry
->d_lockref
.count
== 1)
518 if (lockref_put_or_lock(&dentry
->d_lockref
))
521 if (dentry
->d_flags
& DCACHE_OP_DELETE
) {
522 if (dentry
->d_op
->d_delete(dentry
))
526 /* Unreachable? Get rid of it */
527 if (d_unhashed(dentry
))
530 dentry
->d_flags
|= DCACHE_REFERENCED
;
531 dentry_lru_add(dentry
);
533 dentry
->d_lockref
.count
--;
534 spin_unlock(&dentry
->d_lock
);
538 dentry
= dentry_kill(dentry
, 1);
545 * d_invalidate - invalidate a dentry
546 * @dentry: dentry to invalidate
548 * Try to invalidate the dentry if it turns out to be
549 * possible. If there are other dentries that can be
550 * reached through this one we can't delete it and we
551 * return -EBUSY. On success we return 0.
556 int d_invalidate(struct dentry
* dentry
)
559 * If it's already been dropped, return OK.
561 spin_lock(&dentry
->d_lock
);
562 if (d_unhashed(dentry
)) {
563 spin_unlock(&dentry
->d_lock
);
567 * Check whether to do a partial shrink_dcache
568 * to get rid of unused child entries.
570 if (!list_empty(&dentry
->d_subdirs
)) {
571 spin_unlock(&dentry
->d_lock
);
572 shrink_dcache_parent(dentry
);
573 spin_lock(&dentry
->d_lock
);
577 * Somebody else still using it?
579 * If it's a directory, we can't drop it
580 * for fear of somebody re-populating it
581 * with children (even though dropping it
582 * would make it unreachable from the root,
583 * we might still populate it if it was a
584 * working directory or similar).
585 * We also need to leave mountpoints alone,
588 if (dentry
->d_lockref
.count
> 1 && dentry
->d_inode
) {
589 if (S_ISDIR(dentry
->d_inode
->i_mode
) || d_mountpoint(dentry
)) {
590 spin_unlock(&dentry
->d_lock
);
596 spin_unlock(&dentry
->d_lock
);
599 EXPORT_SYMBOL(d_invalidate
);
601 /* This must be called with d_lock held */
602 static inline void __dget_dlock(struct dentry
*dentry
)
604 dentry
->d_lockref
.count
++;
607 static inline void __dget(struct dentry
*dentry
)
609 lockref_get(&dentry
->d_lockref
);
612 struct dentry
*dget_parent(struct dentry
*dentry
)
618 * Don't need rcu_dereference because we re-check it was correct under
622 ret
= dentry
->d_parent
;
623 spin_lock(&ret
->d_lock
);
624 if (unlikely(ret
!= dentry
->d_parent
)) {
625 spin_unlock(&ret
->d_lock
);
630 BUG_ON(!ret
->d_lockref
.count
);
631 ret
->d_lockref
.count
++;
632 spin_unlock(&ret
->d_lock
);
635 EXPORT_SYMBOL(dget_parent
);
638 * d_find_alias - grab a hashed alias of inode
639 * @inode: inode in question
640 * @want_discon: flag, used by d_splice_alias, to request
641 * that only a DISCONNECTED alias be returned.
643 * If inode has a hashed alias, or is a directory and has any alias,
644 * acquire the reference to alias and return it. Otherwise return NULL.
645 * Notice that if inode is a directory there can be only one alias and
646 * it can be unhashed only if it has no children, or if it is the root
649 * If the inode has an IS_ROOT, DCACHE_DISCONNECTED alias, then prefer
650 * any other hashed alias over that one unless @want_discon is set,
651 * in which case only return an IS_ROOT, DCACHE_DISCONNECTED alias.
653 static struct dentry
*__d_find_alias(struct inode
*inode
, int want_discon
)
655 struct dentry
*alias
, *discon_alias
;
659 hlist_for_each_entry(alias
, &inode
->i_dentry
, d_alias
) {
660 spin_lock(&alias
->d_lock
);
661 if (S_ISDIR(inode
->i_mode
) || !d_unhashed(alias
)) {
662 if (IS_ROOT(alias
) &&
663 (alias
->d_flags
& DCACHE_DISCONNECTED
)) {
664 discon_alias
= alias
;
665 } else if (!want_discon
) {
667 spin_unlock(&alias
->d_lock
);
671 spin_unlock(&alias
->d_lock
);
674 alias
= discon_alias
;
675 spin_lock(&alias
->d_lock
);
676 if (S_ISDIR(inode
->i_mode
) || !d_unhashed(alias
)) {
677 if (IS_ROOT(alias
) &&
678 (alias
->d_flags
& DCACHE_DISCONNECTED
)) {
680 spin_unlock(&alias
->d_lock
);
684 spin_unlock(&alias
->d_lock
);
690 struct dentry
*d_find_alias(struct inode
*inode
)
692 struct dentry
*de
= NULL
;
694 if (!hlist_empty(&inode
->i_dentry
)) {
695 spin_lock(&inode
->i_lock
);
696 de
= __d_find_alias(inode
, 0);
697 spin_unlock(&inode
->i_lock
);
701 EXPORT_SYMBOL(d_find_alias
);
704 * Try to kill dentries associated with this inode.
705 * WARNING: you must own a reference to inode.
707 void d_prune_aliases(struct inode
*inode
)
709 struct dentry
*dentry
;
711 spin_lock(&inode
->i_lock
);
712 hlist_for_each_entry(dentry
, &inode
->i_dentry
, d_alias
) {
713 spin_lock(&dentry
->d_lock
);
714 if (!dentry
->d_lockref
.count
) {
715 __dget_dlock(dentry
);
717 spin_unlock(&dentry
->d_lock
);
718 spin_unlock(&inode
->i_lock
);
722 spin_unlock(&dentry
->d_lock
);
724 spin_unlock(&inode
->i_lock
);
726 EXPORT_SYMBOL(d_prune_aliases
);
729 * Try to throw away a dentry - free the inode, dput the parent.
730 * Requires dentry->d_lock is held, and dentry->d_count == 0.
731 * Releases dentry->d_lock.
733 * This may fail if locks cannot be acquired no problem, just try again.
735 static void try_prune_one_dentry(struct dentry
*dentry
)
736 __releases(dentry
->d_lock
)
738 struct dentry
*parent
;
740 parent
= dentry_kill(dentry
, 0);
742 * If dentry_kill returns NULL, we have nothing more to do.
743 * if it returns the same dentry, trylocks failed. In either
744 * case, just loop again.
746 * Otherwise, we need to prune ancestors too. This is necessary
747 * to prevent quadratic behavior of shrink_dcache_parent(), but
748 * is also expected to be beneficial in reducing dentry cache
753 if (parent
== dentry
)
756 /* Prune ancestors. */
759 if (lockref_put_or_lock(&dentry
->d_lockref
))
761 dentry
= dentry_kill(dentry
, 1);
765 static void shrink_dentry_list(struct list_head
*list
)
767 struct dentry
*dentry
;
771 dentry
= list_entry_rcu(list
->prev
, struct dentry
, d_lru
);
772 if (&dentry
->d_lru
== list
)
774 spin_lock(&dentry
->d_lock
);
775 if (dentry
!= list_entry(list
->prev
, struct dentry
, d_lru
)) {
776 spin_unlock(&dentry
->d_lock
);
781 * We found an inuse dentry which was not removed from
782 * the LRU because of laziness during lookup. Do not free
783 * it - just keep it off the LRU list.
785 if (dentry
->d_lockref
.count
) {
786 dentry_lru_del(dentry
);
787 spin_unlock(&dentry
->d_lock
);
793 try_prune_one_dentry(dentry
);
801 * prune_dcache_sb - shrink the dcache
803 * @count: number of entries to try to free
805 * Attempt to shrink the superblock dcache LRU by @count entries. This is
806 * done when we need more memory an called from the superblock shrinker
809 * This function may fail to free any resources if all the dentries are in
812 void prune_dcache_sb(struct super_block
*sb
, int count
)
814 struct dentry
*dentry
;
815 LIST_HEAD(referenced
);
819 spin_lock(&dcache_lru_lock
);
820 while (!list_empty(&sb
->s_dentry_lru
)) {
821 dentry
= list_entry(sb
->s_dentry_lru
.prev
,
822 struct dentry
, d_lru
);
823 BUG_ON(dentry
->d_sb
!= sb
);
825 if (!spin_trylock(&dentry
->d_lock
)) {
826 spin_unlock(&dcache_lru_lock
);
831 if (dentry
->d_flags
& DCACHE_REFERENCED
) {
832 dentry
->d_flags
&= ~DCACHE_REFERENCED
;
833 list_move(&dentry
->d_lru
, &referenced
);
834 spin_unlock(&dentry
->d_lock
);
836 list_move_tail(&dentry
->d_lru
, &tmp
);
837 dentry
->d_flags
|= DCACHE_SHRINK_LIST
;
838 spin_unlock(&dentry
->d_lock
);
842 cond_resched_lock(&dcache_lru_lock
);
844 if (!list_empty(&referenced
))
845 list_splice(&referenced
, &sb
->s_dentry_lru
);
846 spin_unlock(&dcache_lru_lock
);
848 shrink_dentry_list(&tmp
);
852 * shrink_dcache_sb - shrink dcache for a superblock
855 * Shrink the dcache for the specified super block. This is used to free
856 * the dcache before unmounting a file system.
858 void shrink_dcache_sb(struct super_block
*sb
)
862 spin_lock(&dcache_lru_lock
);
863 while (!list_empty(&sb
->s_dentry_lru
)) {
864 list_splice_init(&sb
->s_dentry_lru
, &tmp
);
865 spin_unlock(&dcache_lru_lock
);
866 shrink_dentry_list(&tmp
);
867 spin_lock(&dcache_lru_lock
);
869 spin_unlock(&dcache_lru_lock
);
871 EXPORT_SYMBOL(shrink_dcache_sb
);
874 * destroy a single subtree of dentries for unmount
875 * - see the comments on shrink_dcache_for_umount() for a description of the
878 static void shrink_dcache_for_umount_subtree(struct dentry
*dentry
)
880 struct dentry
*parent
;
882 BUG_ON(!IS_ROOT(dentry
));
885 /* descend to the first leaf in the current subtree */
886 while (!list_empty(&dentry
->d_subdirs
))
887 dentry
= list_entry(dentry
->d_subdirs
.next
,
888 struct dentry
, d_u
.d_child
);
890 /* consume the dentries from this leaf up through its parents
891 * until we find one with children or run out altogether */
896 * inform the fs that this dentry is about to be
897 * unhashed and destroyed.
899 if (dentry
->d_flags
& DCACHE_OP_PRUNE
)
900 dentry
->d_op
->d_prune(dentry
);
902 dentry_lru_del(dentry
);
905 if (dentry
->d_lockref
.count
!= 0) {
907 "BUG: Dentry %p{i=%lx,n=%s}"
909 " [unmount of %s %s]\n",
912 dentry
->d_inode
->i_ino
: 0UL,
914 dentry
->d_lockref
.count
,
915 dentry
->d_sb
->s_type
->name
,
920 if (IS_ROOT(dentry
)) {
922 list_del(&dentry
->d_u
.d_child
);
924 parent
= dentry
->d_parent
;
925 parent
->d_lockref
.count
--;
926 list_del(&dentry
->d_u
.d_child
);
929 inode
= dentry
->d_inode
;
931 dentry
->d_inode
= NULL
;
932 hlist_del_init(&dentry
->d_alias
);
933 if (dentry
->d_op
&& dentry
->d_op
->d_iput
)
934 dentry
->d_op
->d_iput(dentry
, inode
);
941 /* finished when we fall off the top of the tree,
942 * otherwise we ascend to the parent and move to the
943 * next sibling if there is one */
947 } while (list_empty(&dentry
->d_subdirs
));
949 dentry
= list_entry(dentry
->d_subdirs
.next
,
950 struct dentry
, d_u
.d_child
);
955 * destroy the dentries attached to a superblock on unmounting
956 * - we don't need to use dentry->d_lock because:
957 * - the superblock is detached from all mountings and open files, so the
958 * dentry trees will not be rearranged by the VFS
959 * - s_umount is write-locked, so the memory pressure shrinker will ignore
960 * any dentries belonging to this superblock that it comes across
961 * - the filesystem itself is no longer permitted to rearrange the dentries
964 void shrink_dcache_for_umount(struct super_block
*sb
)
966 struct dentry
*dentry
;
968 if (down_read_trylock(&sb
->s_umount
))
973 dentry
->d_lockref
.count
--;
974 shrink_dcache_for_umount_subtree(dentry
);
976 while (!hlist_bl_empty(&sb
->s_anon
)) {
977 dentry
= hlist_bl_entry(hlist_bl_first(&sb
->s_anon
), struct dentry
, d_hash
);
978 shrink_dcache_for_umount_subtree(dentry
);
983 * This tries to ascend one level of parenthood, but
984 * we can race with renaming, so we need to re-check
985 * the parenthood after dropping the lock and check
986 * that the sequence number still matches.
988 static struct dentry
*try_to_ascend(struct dentry
*old
, int locked
, unsigned seq
)
990 struct dentry
*new = old
->d_parent
;
993 spin_unlock(&old
->d_lock
);
994 spin_lock(&new->d_lock
);
997 * might go back up the wrong parent if we have had a rename
1000 if (new != old
->d_parent
||
1001 (old
->d_flags
& DCACHE_DENTRY_KILLED
) ||
1002 (!locked
&& read_seqretry(&rename_lock
, seq
))) {
1003 spin_unlock(&new->d_lock
);
1012 * Search for at least 1 mount point in the dentry's subdirs.
1013 * We descend to the next level whenever the d_subdirs
1014 * list is non-empty and continue searching.
1018 * have_submounts - check for mounts over a dentry
1019 * @parent: dentry to check.
1021 * Return true if the parent or its subdirectories contain
1024 int have_submounts(struct dentry
*parent
)
1026 struct dentry
*this_parent
;
1027 struct list_head
*next
;
1031 seq
= read_seqbegin(&rename_lock
);
1033 this_parent
= parent
;
1035 if (d_mountpoint(parent
))
1037 spin_lock(&this_parent
->d_lock
);
1039 next
= this_parent
->d_subdirs
.next
;
1041 while (next
!= &this_parent
->d_subdirs
) {
1042 struct list_head
*tmp
= next
;
1043 struct dentry
*dentry
= list_entry(tmp
, struct dentry
, d_u
.d_child
);
1046 spin_lock_nested(&dentry
->d_lock
, DENTRY_D_LOCK_NESTED
);
1047 /* Have we found a mount point ? */
1048 if (d_mountpoint(dentry
)) {
1049 spin_unlock(&dentry
->d_lock
);
1050 spin_unlock(&this_parent
->d_lock
);
1053 if (!list_empty(&dentry
->d_subdirs
)) {
1054 spin_unlock(&this_parent
->d_lock
);
1055 spin_release(&dentry
->d_lock
.dep_map
, 1, _RET_IP_
);
1056 this_parent
= dentry
;
1057 spin_acquire(&this_parent
->d_lock
.dep_map
, 0, 1, _RET_IP_
);
1060 spin_unlock(&dentry
->d_lock
);
1063 * All done at this level ... ascend and resume the search.
1065 if (this_parent
!= parent
) {
1066 struct dentry
*child
= this_parent
;
1067 this_parent
= try_to_ascend(this_parent
, locked
, seq
);
1070 next
= child
->d_u
.d_child
.next
;
1073 spin_unlock(&this_parent
->d_lock
);
1074 if (!locked
&& read_seqretry(&rename_lock
, seq
))
1077 write_sequnlock(&rename_lock
);
1078 return 0; /* No mount points found in tree */
1080 if (!locked
&& read_seqretry(&rename_lock
, seq
))
1083 write_sequnlock(&rename_lock
);
1090 write_seqlock(&rename_lock
);
1093 EXPORT_SYMBOL(have_submounts
);
1096 * Search the dentry child list of the specified parent,
1097 * and move any unused dentries to the end of the unused
1098 * list for prune_dcache(). We descend to the next level
1099 * whenever the d_subdirs list is non-empty and continue
1102 * It returns zero iff there are no unused children,
1103 * otherwise it returns the number of children moved to
1104 * the end of the unused list. This may not be the total
1105 * number of unused children, because select_parent can
1106 * drop the lock and return early due to latency
1109 static int select_parent(struct dentry
*parent
, struct list_head
*dispose
)
1111 struct dentry
*this_parent
;
1112 struct list_head
*next
;
1117 seq
= read_seqbegin(&rename_lock
);
1119 this_parent
= parent
;
1120 spin_lock(&this_parent
->d_lock
);
1122 next
= this_parent
->d_subdirs
.next
;
1124 while (next
!= &this_parent
->d_subdirs
) {
1125 struct list_head
*tmp
= next
;
1126 struct dentry
*dentry
= list_entry(tmp
, struct dentry
, d_u
.d_child
);
1129 spin_lock_nested(&dentry
->d_lock
, DENTRY_D_LOCK_NESTED
);
1132 * move only zero ref count dentries to the dispose list.
1134 * Those which are presently on the shrink list, being processed
1135 * by shrink_dentry_list(), shouldn't be moved. Otherwise the
1136 * loop in shrink_dcache_parent() might not make any progress
1139 if (dentry
->d_lockref
.count
) {
1140 dentry_lru_del(dentry
);
1141 } else if (!(dentry
->d_flags
& DCACHE_SHRINK_LIST
)) {
1142 dentry_lru_move_list(dentry
, dispose
);
1143 dentry
->d_flags
|= DCACHE_SHRINK_LIST
;
1147 * We can return to the caller if we have found some (this
1148 * ensures forward progress). We'll be coming back to find
1151 if (found
&& need_resched()) {
1152 spin_unlock(&dentry
->d_lock
);
1157 * Descend a level if the d_subdirs list is non-empty.
1159 if (!list_empty(&dentry
->d_subdirs
)) {
1160 spin_unlock(&this_parent
->d_lock
);
1161 spin_release(&dentry
->d_lock
.dep_map
, 1, _RET_IP_
);
1162 this_parent
= dentry
;
1163 spin_acquire(&this_parent
->d_lock
.dep_map
, 0, 1, _RET_IP_
);
1167 spin_unlock(&dentry
->d_lock
);
1170 * All done at this level ... ascend and resume the search.
1172 if (this_parent
!= parent
) {
1173 struct dentry
*child
= this_parent
;
1174 this_parent
= try_to_ascend(this_parent
, locked
, seq
);
1177 next
= child
->d_u
.d_child
.next
;
1181 spin_unlock(&this_parent
->d_lock
);
1182 if (!locked
&& read_seqretry(&rename_lock
, seq
))
1185 write_sequnlock(&rename_lock
);
1194 write_seqlock(&rename_lock
);
1199 * shrink_dcache_parent - prune dcache
1200 * @parent: parent of entries to prune
1202 * Prune the dcache to remove unused children of the parent dentry.
1204 void shrink_dcache_parent(struct dentry
* parent
)
1209 while ((found
= select_parent(parent
, &dispose
)) != 0) {
1210 shrink_dentry_list(&dispose
);
1214 EXPORT_SYMBOL(shrink_dcache_parent
);
1217 * __d_alloc - allocate a dcache entry
1218 * @sb: filesystem it will belong to
1219 * @name: qstr of the name
1221 * Allocates a dentry. It returns %NULL if there is insufficient memory
1222 * available. On a success the dentry is returned. The name passed in is
1223 * copied and the copy passed in may be reused after this call.
1226 struct dentry
*__d_alloc(struct super_block
*sb
, const struct qstr
*name
)
1228 struct dentry
*dentry
;
1231 dentry
= kmem_cache_alloc(dentry_cache
, GFP_KERNEL
);
1236 * We guarantee that the inline name is always NUL-terminated.
1237 * This way the memcpy() done by the name switching in rename
1238 * will still always have a NUL at the end, even if we might
1239 * be overwriting an internal NUL character
1241 dentry
->d_iname
[DNAME_INLINE_LEN
-1] = 0;
1242 if (name
->len
> DNAME_INLINE_LEN
-1) {
1243 dname
= kmalloc(name
->len
+ 1, GFP_KERNEL
);
1245 kmem_cache_free(dentry_cache
, dentry
);
1249 dname
= dentry
->d_iname
;
1252 dentry
->d_name
.len
= name
->len
;
1253 dentry
->d_name
.hash
= name
->hash
;
1254 memcpy(dname
, name
->name
, name
->len
);
1255 dname
[name
->len
] = 0;
1257 /* Make sure we always see the terminating NUL character */
1259 dentry
->d_name
.name
= dname
;
1261 dentry
->d_lockref
.count
= 1;
1262 dentry
->d_flags
= 0;
1263 spin_lock_init(&dentry
->d_lock
);
1264 seqcount_init(&dentry
->d_seq
);
1265 dentry
->d_inode
= NULL
;
1266 dentry
->d_parent
= dentry
;
1268 dentry
->d_op
= NULL
;
1269 dentry
->d_fsdata
= NULL
;
1270 INIT_HLIST_BL_NODE(&dentry
->d_hash
);
1271 INIT_LIST_HEAD(&dentry
->d_lru
);
1272 INIT_LIST_HEAD(&dentry
->d_subdirs
);
1273 INIT_HLIST_NODE(&dentry
->d_alias
);
1274 INIT_LIST_HEAD(&dentry
->d_u
.d_child
);
1275 d_set_d_op(dentry
, dentry
->d_sb
->s_d_op
);
1277 this_cpu_inc(nr_dentry
);
1283 * d_alloc - allocate a dcache entry
1284 * @parent: parent of entry to allocate
1285 * @name: qstr of the name
1287 * Allocates a dentry. It returns %NULL if there is insufficient memory
1288 * available. On a success the dentry is returned. The name passed in is
1289 * copied and the copy passed in may be reused after this call.
1291 struct dentry
*d_alloc(struct dentry
* parent
, const struct qstr
*name
)
1293 struct dentry
*dentry
= __d_alloc(parent
->d_sb
, name
);
1297 spin_lock(&parent
->d_lock
);
1299 * don't need child lock because it is not subject
1300 * to concurrency here
1302 __dget_dlock(parent
);
1303 dentry
->d_parent
= parent
;
1304 list_add(&dentry
->d_u
.d_child
, &parent
->d_subdirs
);
1305 spin_unlock(&parent
->d_lock
);
1309 EXPORT_SYMBOL(d_alloc
);
1311 struct dentry
*d_alloc_pseudo(struct super_block
*sb
, const struct qstr
*name
)
1313 struct dentry
*dentry
= __d_alloc(sb
, name
);
1315 dentry
->d_flags
|= DCACHE_DISCONNECTED
;
1318 EXPORT_SYMBOL(d_alloc_pseudo
);
1320 struct dentry
*d_alloc_name(struct dentry
*parent
, const char *name
)
1325 q
.len
= strlen(name
);
1326 q
.hash
= full_name_hash(q
.name
, q
.len
);
1327 return d_alloc(parent
, &q
);
1329 EXPORT_SYMBOL(d_alloc_name
);
1331 void d_set_d_op(struct dentry
*dentry
, const struct dentry_operations
*op
)
1333 WARN_ON_ONCE(dentry
->d_op
);
1334 WARN_ON_ONCE(dentry
->d_flags
& (DCACHE_OP_HASH
|
1336 DCACHE_OP_REVALIDATE
|
1337 DCACHE_OP_WEAK_REVALIDATE
|
1338 DCACHE_OP_DELETE
));
1343 dentry
->d_flags
|= DCACHE_OP_HASH
;
1345 dentry
->d_flags
|= DCACHE_OP_COMPARE
;
1346 if (op
->d_revalidate
)
1347 dentry
->d_flags
|= DCACHE_OP_REVALIDATE
;
1348 if (op
->d_weak_revalidate
)
1349 dentry
->d_flags
|= DCACHE_OP_WEAK_REVALIDATE
;
1351 dentry
->d_flags
|= DCACHE_OP_DELETE
;
1353 dentry
->d_flags
|= DCACHE_OP_PRUNE
;
1356 EXPORT_SYMBOL(d_set_d_op
);
1358 static void __d_instantiate(struct dentry
*dentry
, struct inode
*inode
)
1360 spin_lock(&dentry
->d_lock
);
1362 if (unlikely(IS_AUTOMOUNT(inode
)))
1363 dentry
->d_flags
|= DCACHE_NEED_AUTOMOUNT
;
1364 hlist_add_head(&dentry
->d_alias
, &inode
->i_dentry
);
1366 dentry
->d_inode
= inode
;
1367 dentry_rcuwalk_barrier(dentry
);
1368 spin_unlock(&dentry
->d_lock
);
1369 fsnotify_d_instantiate(dentry
, inode
);
1373 * d_instantiate - fill in inode information for a dentry
1374 * @entry: dentry to complete
1375 * @inode: inode to attach to this dentry
1377 * Fill in inode information in the entry.
1379 * This turns negative dentries into productive full members
1382 * NOTE! This assumes that the inode count has been incremented
1383 * (or otherwise set) by the caller to indicate that it is now
1384 * in use by the dcache.
1387 void d_instantiate(struct dentry
*entry
, struct inode
* inode
)
1389 BUG_ON(!hlist_unhashed(&entry
->d_alias
));
1391 spin_lock(&inode
->i_lock
);
1392 __d_instantiate(entry
, inode
);
1394 spin_unlock(&inode
->i_lock
);
1395 security_d_instantiate(entry
, inode
);
1397 EXPORT_SYMBOL(d_instantiate
);
1400 * d_instantiate_unique - instantiate a non-aliased dentry
1401 * @entry: dentry to instantiate
1402 * @inode: inode to attach to this dentry
1404 * Fill in inode information in the entry. On success, it returns NULL.
1405 * If an unhashed alias of "entry" already exists, then we return the
1406 * aliased dentry instead and drop one reference to inode.
1408 * Note that in order to avoid conflicts with rename() etc, the caller
1409 * had better be holding the parent directory semaphore.
1411 * This also assumes that the inode count has been incremented
1412 * (or otherwise set) by the caller to indicate that it is now
1413 * in use by the dcache.
1415 static struct dentry
*__d_instantiate_unique(struct dentry
*entry
,
1416 struct inode
*inode
)
1418 struct dentry
*alias
;
1419 int len
= entry
->d_name
.len
;
1420 const char *name
= entry
->d_name
.name
;
1421 unsigned int hash
= entry
->d_name
.hash
;
1424 __d_instantiate(entry
, NULL
);
1428 hlist_for_each_entry(alias
, &inode
->i_dentry
, d_alias
) {
1430 * Don't need alias->d_lock here, because aliases with
1431 * d_parent == entry->d_parent are not subject to name or
1432 * parent changes, because the parent inode i_mutex is held.
1434 if (alias
->d_name
.hash
!= hash
)
1436 if (alias
->d_parent
!= entry
->d_parent
)
1438 if (alias
->d_name
.len
!= len
)
1440 if (dentry_cmp(alias
, name
, len
))
1446 __d_instantiate(entry
, inode
);
1450 struct dentry
*d_instantiate_unique(struct dentry
*entry
, struct inode
*inode
)
1452 struct dentry
*result
;
1454 BUG_ON(!hlist_unhashed(&entry
->d_alias
));
1457 spin_lock(&inode
->i_lock
);
1458 result
= __d_instantiate_unique(entry
, inode
);
1460 spin_unlock(&inode
->i_lock
);
1463 security_d_instantiate(entry
, inode
);
1467 BUG_ON(!d_unhashed(result
));
1472 EXPORT_SYMBOL(d_instantiate_unique
);
1474 struct dentry
*d_make_root(struct inode
*root_inode
)
1476 struct dentry
*res
= NULL
;
1479 static const struct qstr name
= QSTR_INIT("/", 1);
1481 res
= __d_alloc(root_inode
->i_sb
, &name
);
1483 d_instantiate(res
, root_inode
);
1489 EXPORT_SYMBOL(d_make_root
);
1491 static struct dentry
* __d_find_any_alias(struct inode
*inode
)
1493 struct dentry
*alias
;
1495 if (hlist_empty(&inode
->i_dentry
))
1497 alias
= hlist_entry(inode
->i_dentry
.first
, struct dentry
, d_alias
);
1503 * d_find_any_alias - find any alias for a given inode
1504 * @inode: inode to find an alias for
1506 * If any aliases exist for the given inode, take and return a
1507 * reference for one of them. If no aliases exist, return %NULL.
1509 struct dentry
*d_find_any_alias(struct inode
*inode
)
1513 spin_lock(&inode
->i_lock
);
1514 de
= __d_find_any_alias(inode
);
1515 spin_unlock(&inode
->i_lock
);
1518 EXPORT_SYMBOL(d_find_any_alias
);
1521 * d_obtain_alias - find or allocate a dentry for a given inode
1522 * @inode: inode to allocate the dentry for
1524 * Obtain a dentry for an inode resulting from NFS filehandle conversion or
1525 * similar open by handle operations. The returned dentry may be anonymous,
1526 * or may have a full name (if the inode was already in the cache).
1528 * When called on a directory inode, we must ensure that the inode only ever
1529 * has one dentry. If a dentry is found, that is returned instead of
1530 * allocating a new one.
1532 * On successful return, the reference to the inode has been transferred
1533 * to the dentry. In case of an error the reference on the inode is released.
1534 * To make it easier to use in export operations a %NULL or IS_ERR inode may
1535 * be passed in and will be the error will be propagate to the return value,
1536 * with a %NULL @inode replaced by ERR_PTR(-ESTALE).
1538 struct dentry
*d_obtain_alias(struct inode
*inode
)
1540 static const struct qstr anonstring
= QSTR_INIT("/", 1);
1545 return ERR_PTR(-ESTALE
);
1547 return ERR_CAST(inode
);
1549 res
= d_find_any_alias(inode
);
1553 tmp
= __d_alloc(inode
->i_sb
, &anonstring
);
1555 res
= ERR_PTR(-ENOMEM
);
1559 spin_lock(&inode
->i_lock
);
1560 res
= __d_find_any_alias(inode
);
1562 spin_unlock(&inode
->i_lock
);
1567 /* attach a disconnected dentry */
1568 spin_lock(&tmp
->d_lock
);
1569 tmp
->d_inode
= inode
;
1570 tmp
->d_flags
|= DCACHE_DISCONNECTED
;
1571 hlist_add_head(&tmp
->d_alias
, &inode
->i_dentry
);
1572 hlist_bl_lock(&tmp
->d_sb
->s_anon
);
1573 hlist_bl_add_head(&tmp
->d_hash
, &tmp
->d_sb
->s_anon
);
1574 hlist_bl_unlock(&tmp
->d_sb
->s_anon
);
1575 spin_unlock(&tmp
->d_lock
);
1576 spin_unlock(&inode
->i_lock
);
1577 security_d_instantiate(tmp
, inode
);
1582 if (res
&& !IS_ERR(res
))
1583 security_d_instantiate(res
, inode
);
1587 EXPORT_SYMBOL(d_obtain_alias
);
1590 * d_splice_alias - splice a disconnected dentry into the tree if one exists
1591 * @inode: the inode which may have a disconnected dentry
1592 * @dentry: a negative dentry which we want to point to the inode.
1594 * If inode is a directory and has a 'disconnected' dentry (i.e. IS_ROOT and
1595 * DCACHE_DISCONNECTED), then d_move that in place of the given dentry
1596 * and return it, else simply d_add the inode to the dentry and return NULL.
1598 * This is needed in the lookup routine of any filesystem that is exportable
1599 * (via knfsd) so that we can build dcache paths to directories effectively.
1601 * If a dentry was found and moved, then it is returned. Otherwise NULL
1602 * is returned. This matches the expected return value of ->lookup.
1604 * Cluster filesystems may call this function with a negative, hashed dentry.
1605 * In that case, we know that the inode will be a regular file, and also this
1606 * will only occur during atomic_open. So we need to check for the dentry
1607 * being already hashed only in the final case.
1609 struct dentry
*d_splice_alias(struct inode
*inode
, struct dentry
*dentry
)
1611 struct dentry
*new = NULL
;
1614 return ERR_CAST(inode
);
1616 if (inode
&& S_ISDIR(inode
->i_mode
)) {
1617 spin_lock(&inode
->i_lock
);
1618 new = __d_find_alias(inode
, 1);
1620 BUG_ON(!(new->d_flags
& DCACHE_DISCONNECTED
));
1621 spin_unlock(&inode
->i_lock
);
1622 security_d_instantiate(new, inode
);
1623 d_move(new, dentry
);
1626 /* already taking inode->i_lock, so d_add() by hand */
1627 __d_instantiate(dentry
, inode
);
1628 spin_unlock(&inode
->i_lock
);
1629 security_d_instantiate(dentry
, inode
);
1633 d_instantiate(dentry
, inode
);
1634 if (d_unhashed(dentry
))
1639 EXPORT_SYMBOL(d_splice_alias
);
1642 * d_add_ci - lookup or allocate new dentry with case-exact name
1643 * @inode: the inode case-insensitive lookup has found
1644 * @dentry: the negative dentry that was passed to the parent's lookup func
1645 * @name: the case-exact name to be associated with the returned dentry
1647 * This is to avoid filling the dcache with case-insensitive names to the
1648 * same inode, only the actual correct case is stored in the dcache for
1649 * case-insensitive filesystems.
1651 * For a case-insensitive lookup match and if the the case-exact dentry
1652 * already exists in in the dcache, use it and return it.
1654 * If no entry exists with the exact case name, allocate new dentry with
1655 * the exact case, and return the spliced entry.
1657 struct dentry
*d_add_ci(struct dentry
*dentry
, struct inode
*inode
,
1660 struct dentry
*found
;
1664 * First check if a dentry matching the name already exists,
1665 * if not go ahead and create it now.
1667 found
= d_hash_and_lookup(dentry
->d_parent
, name
);
1668 if (unlikely(IS_ERR(found
)))
1671 new = d_alloc(dentry
->d_parent
, name
);
1673 found
= ERR_PTR(-ENOMEM
);
1677 found
= d_splice_alias(inode
, new);
1686 * If a matching dentry exists, and it's not negative use it.
1688 * Decrement the reference count to balance the iget() done
1691 if (found
->d_inode
) {
1692 if (unlikely(found
->d_inode
!= inode
)) {
1693 /* This can't happen because bad inodes are unhashed. */
1694 BUG_ON(!is_bad_inode(inode
));
1695 BUG_ON(!is_bad_inode(found
->d_inode
));
1702 * Negative dentry: instantiate it unless the inode is a directory and
1703 * already has a dentry.
1705 new = d_splice_alias(inode
, found
);
1716 EXPORT_SYMBOL(d_add_ci
);
1719 * Do the slow-case of the dentry name compare.
1721 * Unlike the dentry_cmp() function, we need to atomically
1722 * load the name and length information, so that the
1723 * filesystem can rely on them, and can use the 'name' and
1724 * 'len' information without worrying about walking off the
1725 * end of memory etc.
1727 * Thus the read_seqcount_retry() and the "duplicate" info
1728 * in arguments (the low-level filesystem should not look
1729 * at the dentry inode or name contents directly, since
1730 * rename can change them while we're in RCU mode).
1732 enum slow_d_compare
{
1738 static noinline
enum slow_d_compare
slow_dentry_cmp(
1739 const struct dentry
*parent
,
1740 struct dentry
*dentry
,
1742 const struct qstr
*name
)
1744 int tlen
= dentry
->d_name
.len
;
1745 const char *tname
= dentry
->d_name
.name
;
1747 if (read_seqcount_retry(&dentry
->d_seq
, seq
)) {
1749 return D_COMP_SEQRETRY
;
1751 if (parent
->d_op
->d_compare(parent
, dentry
, tlen
, tname
, name
))
1752 return D_COMP_NOMATCH
;
1757 * __d_lookup_rcu - search for a dentry (racy, store-free)
1758 * @parent: parent dentry
1759 * @name: qstr of name we wish to find
1760 * @seqp: returns d_seq value at the point where the dentry was found
1761 * Returns: dentry, or NULL
1763 * __d_lookup_rcu is the dcache lookup function for rcu-walk name
1764 * resolution (store-free path walking) design described in
1765 * Documentation/filesystems/path-lookup.txt.
1767 * This is not to be used outside core vfs.
1769 * __d_lookup_rcu must only be used in rcu-walk mode, ie. with vfsmount lock
1770 * held, and rcu_read_lock held. The returned dentry must not be stored into
1771 * without taking d_lock and checking d_seq sequence count against @seq
1774 * A refcount may be taken on the found dentry with the __d_rcu_to_refcount
1777 * Alternatively, __d_lookup_rcu may be called again to look up the child of
1778 * the returned dentry, so long as its parent's seqlock is checked after the
1779 * child is looked up. Thus, an interlocking stepping of sequence lock checks
1780 * is formed, giving integrity down the path walk.
1782 * NOTE! The caller *has* to check the resulting dentry against the sequence
1783 * number we've returned before using any of the resulting dentry state!
1785 struct dentry
*__d_lookup_rcu(const struct dentry
*parent
,
1786 const struct qstr
*name
,
1789 u64 hashlen
= name
->hash_len
;
1790 const unsigned char *str
= name
->name
;
1791 struct hlist_bl_head
*b
= d_hash(parent
, hashlen_hash(hashlen
));
1792 struct hlist_bl_node
*node
;
1793 struct dentry
*dentry
;
1796 * Note: There is significant duplication with __d_lookup_rcu which is
1797 * required to prevent single threaded performance regressions
1798 * especially on architectures where smp_rmb (in seqcounts) are costly.
1799 * Keep the two functions in sync.
1803 * The hash list is protected using RCU.
1805 * Carefully use d_seq when comparing a candidate dentry, to avoid
1806 * races with d_move().
1808 * It is possible that concurrent renames can mess up our list
1809 * walk here and result in missing our dentry, resulting in the
1810 * false-negative result. d_lookup() protects against concurrent
1811 * renames using rename_lock seqlock.
1813 * See Documentation/filesystems/path-lookup.txt for more details.
1815 hlist_bl_for_each_entry_rcu(dentry
, node
, b
, d_hash
) {
1820 * The dentry sequence count protects us from concurrent
1821 * renames, and thus protects parent and name fields.
1823 * The caller must perform a seqcount check in order
1824 * to do anything useful with the returned dentry.
1826 * NOTE! We do a "raw" seqcount_begin here. That means that
1827 * we don't wait for the sequence count to stabilize if it
1828 * is in the middle of a sequence change. If we do the slow
1829 * dentry compare, we will do seqretries until it is stable,
1830 * and if we end up with a successful lookup, we actually
1831 * want to exit RCU lookup anyway.
1833 seq
= raw_seqcount_begin(&dentry
->d_seq
);
1834 if (dentry
->d_parent
!= parent
)
1836 if (d_unhashed(dentry
))
1839 if (unlikely(parent
->d_flags
& DCACHE_OP_COMPARE
)) {
1840 if (dentry
->d_name
.hash
!= hashlen_hash(hashlen
))
1843 switch (slow_dentry_cmp(parent
, dentry
, seq
, name
)) {
1846 case D_COMP_NOMATCH
:
1853 if (dentry
->d_name
.hash_len
!= hashlen
)
1856 if (!dentry_cmp(dentry
, str
, hashlen_len(hashlen
)))
1863 * d_lookup - search for a dentry
1864 * @parent: parent dentry
1865 * @name: qstr of name we wish to find
1866 * Returns: dentry, or NULL
1868 * d_lookup searches the children of the parent dentry for the name in
1869 * question. If the dentry is found its reference count is incremented and the
1870 * dentry is returned. The caller must use dput to free the entry when it has
1871 * finished using it. %NULL is returned if the dentry does not exist.
1873 struct dentry
*d_lookup(const struct dentry
*parent
, const struct qstr
*name
)
1875 struct dentry
*dentry
;
1879 seq
= read_seqbegin(&rename_lock
);
1880 dentry
= __d_lookup(parent
, name
);
1883 } while (read_seqretry(&rename_lock
, seq
));
1886 EXPORT_SYMBOL(d_lookup
);
1889 * __d_lookup - search for a dentry (racy)
1890 * @parent: parent dentry
1891 * @name: qstr of name we wish to find
1892 * Returns: dentry, or NULL
1894 * __d_lookup is like d_lookup, however it may (rarely) return a
1895 * false-negative result due to unrelated rename activity.
1897 * __d_lookup is slightly faster by avoiding rename_lock read seqlock,
1898 * however it must be used carefully, eg. with a following d_lookup in
1899 * the case of failure.
1901 * __d_lookup callers must be commented.
1903 struct dentry
*__d_lookup(const struct dentry
*parent
, const struct qstr
*name
)
1905 unsigned int len
= name
->len
;
1906 unsigned int hash
= name
->hash
;
1907 const unsigned char *str
= name
->name
;
1908 struct hlist_bl_head
*b
= d_hash(parent
, hash
);
1909 struct hlist_bl_node
*node
;
1910 struct dentry
*found
= NULL
;
1911 struct dentry
*dentry
;
1914 * Note: There is significant duplication with __d_lookup_rcu which is
1915 * required to prevent single threaded performance regressions
1916 * especially on architectures where smp_rmb (in seqcounts) are costly.
1917 * Keep the two functions in sync.
1921 * The hash list is protected using RCU.
1923 * Take d_lock when comparing a candidate dentry, to avoid races
1926 * It is possible that concurrent renames can mess up our list
1927 * walk here and result in missing our dentry, resulting in the
1928 * false-negative result. d_lookup() protects against concurrent
1929 * renames using rename_lock seqlock.
1931 * See Documentation/filesystems/path-lookup.txt for more details.
1935 hlist_bl_for_each_entry_rcu(dentry
, node
, b
, d_hash
) {
1937 if (dentry
->d_name
.hash
!= hash
)
1940 spin_lock(&dentry
->d_lock
);
1941 if (dentry
->d_parent
!= parent
)
1943 if (d_unhashed(dentry
))
1947 * It is safe to compare names since d_move() cannot
1948 * change the qstr (protected by d_lock).
1950 if (parent
->d_flags
& DCACHE_OP_COMPARE
) {
1951 int tlen
= dentry
->d_name
.len
;
1952 const char *tname
= dentry
->d_name
.name
;
1953 if (parent
->d_op
->d_compare(parent
, dentry
, tlen
, tname
, name
))
1956 if (dentry
->d_name
.len
!= len
)
1958 if (dentry_cmp(dentry
, str
, len
))
1962 dentry
->d_lockref
.count
++;
1964 spin_unlock(&dentry
->d_lock
);
1967 spin_unlock(&dentry
->d_lock
);
1975 * d_hash_and_lookup - hash the qstr then search for a dentry
1976 * @dir: Directory to search in
1977 * @name: qstr of name we wish to find
1979 * On lookup failure NULL is returned; on bad name - ERR_PTR(-error)
1981 struct dentry
*d_hash_and_lookup(struct dentry
*dir
, struct qstr
*name
)
1984 * Check for a fs-specific hash function. Note that we must
1985 * calculate the standard hash first, as the d_op->d_hash()
1986 * routine may choose to leave the hash value unchanged.
1988 name
->hash
= full_name_hash(name
->name
, name
->len
);
1989 if (dir
->d_flags
& DCACHE_OP_HASH
) {
1990 int err
= dir
->d_op
->d_hash(dir
, name
);
1991 if (unlikely(err
< 0))
1992 return ERR_PTR(err
);
1994 return d_lookup(dir
, name
);
1996 EXPORT_SYMBOL(d_hash_and_lookup
);
1999 * d_validate - verify dentry provided from insecure source (deprecated)
2000 * @dentry: The dentry alleged to be valid child of @dparent
2001 * @dparent: The parent dentry (known to be valid)
2003 * An insecure source has sent us a dentry, here we verify it and dget() it.
2004 * This is used by ncpfs in its readdir implementation.
2005 * Zero is returned in the dentry is invalid.
2007 * This function is slow for big directories, and deprecated, do not use it.
2009 int d_validate(struct dentry
*dentry
, struct dentry
*dparent
)
2011 struct dentry
*child
;
2013 spin_lock(&dparent
->d_lock
);
2014 list_for_each_entry(child
, &dparent
->d_subdirs
, d_u
.d_child
) {
2015 if (dentry
== child
) {
2016 spin_lock_nested(&dentry
->d_lock
, DENTRY_D_LOCK_NESTED
);
2017 __dget_dlock(dentry
);
2018 spin_unlock(&dentry
->d_lock
);
2019 spin_unlock(&dparent
->d_lock
);
2023 spin_unlock(&dparent
->d_lock
);
2027 EXPORT_SYMBOL(d_validate
);
2030 * When a file is deleted, we have two options:
2031 * - turn this dentry into a negative dentry
2032 * - unhash this dentry and free it.
2034 * Usually, we want to just turn this into
2035 * a negative dentry, but if anybody else is
2036 * currently using the dentry or the inode
2037 * we can't do that and we fall back on removing
2038 * it from the hash queues and waiting for
2039 * it to be deleted later when it has no users
2043 * d_delete - delete a dentry
2044 * @dentry: The dentry to delete
2046 * Turn the dentry into a negative dentry if possible, otherwise
2047 * remove it from the hash queues so it can be deleted later
2050 void d_delete(struct dentry
* dentry
)
2052 struct inode
*inode
;
2055 * Are we the only user?
2058 spin_lock(&dentry
->d_lock
);
2059 inode
= dentry
->d_inode
;
2060 isdir
= S_ISDIR(inode
->i_mode
);
2061 if (dentry
->d_lockref
.count
== 1) {
2062 if (!spin_trylock(&inode
->i_lock
)) {
2063 spin_unlock(&dentry
->d_lock
);
2067 dentry
->d_flags
&= ~DCACHE_CANT_MOUNT
;
2068 dentry_unlink_inode(dentry
);
2069 fsnotify_nameremove(dentry
, isdir
);
2073 if (!d_unhashed(dentry
))
2076 spin_unlock(&dentry
->d_lock
);
2078 fsnotify_nameremove(dentry
, isdir
);
2080 EXPORT_SYMBOL(d_delete
);
2082 static void __d_rehash(struct dentry
* entry
, struct hlist_bl_head
*b
)
2084 BUG_ON(!d_unhashed(entry
));
2086 entry
->d_flags
|= DCACHE_RCUACCESS
;
2087 hlist_bl_add_head_rcu(&entry
->d_hash
, b
);
2091 static void _d_rehash(struct dentry
* entry
)
2093 __d_rehash(entry
, d_hash(entry
->d_parent
, entry
->d_name
.hash
));
2097 * d_rehash - add an entry back to the hash
2098 * @entry: dentry to add to the hash
2100 * Adds a dentry to the hash according to its name.
2103 void d_rehash(struct dentry
* entry
)
2105 spin_lock(&entry
->d_lock
);
2107 spin_unlock(&entry
->d_lock
);
2109 EXPORT_SYMBOL(d_rehash
);
2112 * dentry_update_name_case - update case insensitive dentry with a new name
2113 * @dentry: dentry to be updated
2116 * Update a case insensitive dentry with new case of name.
2118 * dentry must have been returned by d_lookup with name @name. Old and new
2119 * name lengths must match (ie. no d_compare which allows mismatched name
2122 * Parent inode i_mutex must be held over d_lookup and into this call (to
2123 * keep renames and concurrent inserts, and readdir(2) away).
2125 void dentry_update_name_case(struct dentry
*dentry
, struct qstr
*name
)
2127 BUG_ON(!mutex_is_locked(&dentry
->d_parent
->d_inode
->i_mutex
));
2128 BUG_ON(dentry
->d_name
.len
!= name
->len
); /* d_lookup gives this */
2130 spin_lock(&dentry
->d_lock
);
2131 write_seqcount_begin(&dentry
->d_seq
);
2132 memcpy((unsigned char *)dentry
->d_name
.name
, name
->name
, name
->len
);
2133 write_seqcount_end(&dentry
->d_seq
);
2134 spin_unlock(&dentry
->d_lock
);
2136 EXPORT_SYMBOL(dentry_update_name_case
);
2138 static void switch_names(struct dentry
*dentry
, struct dentry
*target
)
2140 if (dname_external(target
)) {
2141 if (dname_external(dentry
)) {
2143 * Both external: swap the pointers
2145 swap(target
->d_name
.name
, dentry
->d_name
.name
);
2148 * dentry:internal, target:external. Steal target's
2149 * storage and make target internal.
2151 memcpy(target
->d_iname
, dentry
->d_name
.name
,
2152 dentry
->d_name
.len
+ 1);
2153 dentry
->d_name
.name
= target
->d_name
.name
;
2154 target
->d_name
.name
= target
->d_iname
;
2157 if (dname_external(dentry
)) {
2159 * dentry:external, target:internal. Give dentry's
2160 * storage to target and make dentry internal
2162 memcpy(dentry
->d_iname
, target
->d_name
.name
,
2163 target
->d_name
.len
+ 1);
2164 target
->d_name
.name
= dentry
->d_name
.name
;
2165 dentry
->d_name
.name
= dentry
->d_iname
;
2168 * Both are internal. Just copy target to dentry
2170 memcpy(dentry
->d_iname
, target
->d_name
.name
,
2171 target
->d_name
.len
+ 1);
2172 dentry
->d_name
.len
= target
->d_name
.len
;
2176 swap(dentry
->d_name
.len
, target
->d_name
.len
);
2179 static void dentry_lock_for_move(struct dentry
*dentry
, struct dentry
*target
)
2182 * XXXX: do we really need to take target->d_lock?
2184 if (IS_ROOT(dentry
) || dentry
->d_parent
== target
->d_parent
)
2185 spin_lock(&target
->d_parent
->d_lock
);
2187 if (d_ancestor(dentry
->d_parent
, target
->d_parent
)) {
2188 spin_lock(&dentry
->d_parent
->d_lock
);
2189 spin_lock_nested(&target
->d_parent
->d_lock
,
2190 DENTRY_D_LOCK_NESTED
);
2192 spin_lock(&target
->d_parent
->d_lock
);
2193 spin_lock_nested(&dentry
->d_parent
->d_lock
,
2194 DENTRY_D_LOCK_NESTED
);
2197 if (target
< dentry
) {
2198 spin_lock_nested(&target
->d_lock
, 2);
2199 spin_lock_nested(&dentry
->d_lock
, 3);
2201 spin_lock_nested(&dentry
->d_lock
, 2);
2202 spin_lock_nested(&target
->d_lock
, 3);
2206 static void dentry_unlock_parents_for_move(struct dentry
*dentry
,
2207 struct dentry
*target
)
2209 if (target
->d_parent
!= dentry
->d_parent
)
2210 spin_unlock(&dentry
->d_parent
->d_lock
);
2211 if (target
->d_parent
!= target
)
2212 spin_unlock(&target
->d_parent
->d_lock
);
2216 * When switching names, the actual string doesn't strictly have to
2217 * be preserved in the target - because we're dropping the target
2218 * anyway. As such, we can just do a simple memcpy() to copy over
2219 * the new name before we switch.
2221 * Note that we have to be a lot more careful about getting the hash
2222 * switched - we have to switch the hash value properly even if it
2223 * then no longer matches the actual (corrupted) string of the target.
2224 * The hash value has to match the hash queue that the dentry is on..
2227 * __d_move - move a dentry
2228 * @dentry: entry to move
2229 * @target: new dentry
2231 * Update the dcache to reflect the move of a file name. Negative
2232 * dcache entries should not be moved in this way. Caller must hold
2233 * rename_lock, the i_mutex of the source and target directories,
2234 * and the sb->s_vfs_rename_mutex if they differ. See lock_rename().
2236 static void __d_move(struct dentry
* dentry
, struct dentry
* target
)
2238 if (!dentry
->d_inode
)
2239 printk(KERN_WARNING
"VFS: moving negative dcache entry\n");
2241 BUG_ON(d_ancestor(dentry
, target
));
2242 BUG_ON(d_ancestor(target
, dentry
));
2244 dentry_lock_for_move(dentry
, target
);
2246 write_seqcount_begin(&dentry
->d_seq
);
2247 write_seqcount_begin(&target
->d_seq
);
2249 /* __d_drop does write_seqcount_barrier, but they're OK to nest. */
2252 * Move the dentry to the target hash queue. Don't bother checking
2253 * for the same hash queue because of how unlikely it is.
2256 __d_rehash(dentry
, d_hash(target
->d_parent
, target
->d_name
.hash
));
2258 /* Unhash the target: dput() will then get rid of it */
2261 list_del(&dentry
->d_u
.d_child
);
2262 list_del(&target
->d_u
.d_child
);
2264 /* Switch the names.. */
2265 switch_names(dentry
, target
);
2266 swap(dentry
->d_name
.hash
, target
->d_name
.hash
);
2268 /* ... and switch the parents */
2269 if (IS_ROOT(dentry
)) {
2270 dentry
->d_parent
= target
->d_parent
;
2271 target
->d_parent
= target
;
2272 INIT_LIST_HEAD(&target
->d_u
.d_child
);
2274 swap(dentry
->d_parent
, target
->d_parent
);
2276 /* And add them back to the (new) parent lists */
2277 list_add(&target
->d_u
.d_child
, &target
->d_parent
->d_subdirs
);
2280 list_add(&dentry
->d_u
.d_child
, &dentry
->d_parent
->d_subdirs
);
2282 write_seqcount_end(&target
->d_seq
);
2283 write_seqcount_end(&dentry
->d_seq
);
2285 dentry_unlock_parents_for_move(dentry
, target
);
2286 spin_unlock(&target
->d_lock
);
2287 fsnotify_d_move(dentry
);
2288 spin_unlock(&dentry
->d_lock
);
2292 * d_move - move a dentry
2293 * @dentry: entry to move
2294 * @target: new dentry
2296 * Update the dcache to reflect the move of a file name. Negative
2297 * dcache entries should not be moved in this way. See the locking
2298 * requirements for __d_move.
2300 void d_move(struct dentry
*dentry
, struct dentry
*target
)
2302 write_seqlock(&rename_lock
);
2303 __d_move(dentry
, target
);
2304 write_sequnlock(&rename_lock
);
2306 EXPORT_SYMBOL(d_move
);
2309 * d_ancestor - search for an ancestor
2310 * @p1: ancestor dentry
2313 * Returns the ancestor dentry of p2 which is a child of p1, if p1 is
2314 * an ancestor of p2, else NULL.
2316 struct dentry
*d_ancestor(struct dentry
*p1
, struct dentry
*p2
)
2320 for (p
= p2
; !IS_ROOT(p
); p
= p
->d_parent
) {
2321 if (p
->d_parent
== p1
)
2328 * This helper attempts to cope with remotely renamed directories
2330 * It assumes that the caller is already holding
2331 * dentry->d_parent->d_inode->i_mutex, inode->i_lock and rename_lock
2333 * Note: If ever the locking in lock_rename() changes, then please
2334 * remember to update this too...
2336 static struct dentry
*__d_unalias(struct inode
*inode
,
2337 struct dentry
*dentry
, struct dentry
*alias
)
2339 struct mutex
*m1
= NULL
, *m2
= NULL
;
2340 struct dentry
*ret
= ERR_PTR(-EBUSY
);
2342 /* If alias and dentry share a parent, then no extra locks required */
2343 if (alias
->d_parent
== dentry
->d_parent
)
2346 /* See lock_rename() */
2347 if (!mutex_trylock(&dentry
->d_sb
->s_vfs_rename_mutex
))
2349 m1
= &dentry
->d_sb
->s_vfs_rename_mutex
;
2350 if (!mutex_trylock(&alias
->d_parent
->d_inode
->i_mutex
))
2352 m2
= &alias
->d_parent
->d_inode
->i_mutex
;
2354 if (likely(!d_mountpoint(alias
))) {
2355 __d_move(alias
, dentry
);
2359 spin_unlock(&inode
->i_lock
);
2368 * Prepare an anonymous dentry for life in the superblock's dentry tree as a
2369 * named dentry in place of the dentry to be replaced.
2370 * returns with anon->d_lock held!
2372 static void __d_materialise_dentry(struct dentry
*dentry
, struct dentry
*anon
)
2374 struct dentry
*dparent
;
2376 dentry_lock_for_move(anon
, dentry
);
2378 write_seqcount_begin(&dentry
->d_seq
);
2379 write_seqcount_begin(&anon
->d_seq
);
2381 dparent
= dentry
->d_parent
;
2383 switch_names(dentry
, anon
);
2384 swap(dentry
->d_name
.hash
, anon
->d_name
.hash
);
2386 dentry
->d_parent
= dentry
;
2387 list_del_init(&dentry
->d_u
.d_child
);
2388 anon
->d_parent
= dparent
;
2389 list_move(&anon
->d_u
.d_child
, &dparent
->d_subdirs
);
2391 write_seqcount_end(&dentry
->d_seq
);
2392 write_seqcount_end(&anon
->d_seq
);
2394 dentry_unlock_parents_for_move(anon
, dentry
);
2395 spin_unlock(&dentry
->d_lock
);
2397 /* anon->d_lock still locked, returns locked */
2398 anon
->d_flags
&= ~DCACHE_DISCONNECTED
;
2402 * d_materialise_unique - introduce an inode into the tree
2403 * @dentry: candidate dentry
2404 * @inode: inode to bind to the dentry, to which aliases may be attached
2406 * Introduces an dentry into the tree, substituting an extant disconnected
2407 * root directory alias in its place if there is one. Caller must hold the
2408 * i_mutex of the parent directory.
2410 struct dentry
*d_materialise_unique(struct dentry
*dentry
, struct inode
*inode
)
2412 struct dentry
*actual
;
2414 BUG_ON(!d_unhashed(dentry
));
2418 __d_instantiate(dentry
, NULL
);
2423 spin_lock(&inode
->i_lock
);
2425 if (S_ISDIR(inode
->i_mode
)) {
2426 struct dentry
*alias
;
2428 /* Does an aliased dentry already exist? */
2429 alias
= __d_find_alias(inode
, 0);
2432 write_seqlock(&rename_lock
);
2434 if (d_ancestor(alias
, dentry
)) {
2435 /* Check for loops */
2436 actual
= ERR_PTR(-ELOOP
);
2437 spin_unlock(&inode
->i_lock
);
2438 } else if (IS_ROOT(alias
)) {
2439 /* Is this an anonymous mountpoint that we
2440 * could splice into our tree? */
2441 __d_materialise_dentry(dentry
, alias
);
2442 write_sequnlock(&rename_lock
);
2446 /* Nope, but we must(!) avoid directory
2447 * aliasing. This drops inode->i_lock */
2448 actual
= __d_unalias(inode
, dentry
, alias
);
2450 write_sequnlock(&rename_lock
);
2451 if (IS_ERR(actual
)) {
2452 if (PTR_ERR(actual
) == -ELOOP
)
2453 pr_warn_ratelimited(
2454 "VFS: Lookup of '%s' in %s %s"
2455 " would have caused loop\n",
2456 dentry
->d_name
.name
,
2457 inode
->i_sb
->s_type
->name
,
2465 /* Add a unique reference */
2466 actual
= __d_instantiate_unique(dentry
, inode
);
2470 BUG_ON(!d_unhashed(actual
));
2472 spin_lock(&actual
->d_lock
);
2475 spin_unlock(&actual
->d_lock
);
2476 spin_unlock(&inode
->i_lock
);
2478 if (actual
== dentry
) {
2479 security_d_instantiate(dentry
, inode
);
2486 EXPORT_SYMBOL_GPL(d_materialise_unique
);
2488 static int prepend(char **buffer
, int *buflen
, const char *str
, int namelen
)
2492 return -ENAMETOOLONG
;
2494 memcpy(*buffer
, str
, namelen
);
2498 static int prepend_name(char **buffer
, int *buflen
, struct qstr
*name
)
2500 return prepend(buffer
, buflen
, name
->name
, name
->len
);
2504 * prepend_path - Prepend path string to a buffer
2505 * @path: the dentry/vfsmount to report
2506 * @root: root vfsmnt/dentry
2507 * @buffer: pointer to the end of the buffer
2508 * @buflen: pointer to buffer length
2510 * Caller holds the rename_lock.
2512 static int prepend_path(const struct path
*path
,
2513 const struct path
*root
,
2514 char **buffer
, int *buflen
)
2516 struct dentry
*dentry
= path
->dentry
;
2517 struct vfsmount
*vfsmnt
= path
->mnt
;
2518 struct mount
*mnt
= real_mount(vfsmnt
);
2522 while (dentry
!= root
->dentry
|| vfsmnt
!= root
->mnt
) {
2523 struct dentry
* parent
;
2525 if (dentry
== vfsmnt
->mnt_root
|| IS_ROOT(dentry
)) {
2527 if (!mnt_has_parent(mnt
))
2529 dentry
= mnt
->mnt_mountpoint
;
2530 mnt
= mnt
->mnt_parent
;
2534 parent
= dentry
->d_parent
;
2536 spin_lock(&dentry
->d_lock
);
2537 error
= prepend_name(buffer
, buflen
, &dentry
->d_name
);
2538 spin_unlock(&dentry
->d_lock
);
2540 error
= prepend(buffer
, buflen
, "/", 1);
2548 if (!error
&& !slash
)
2549 error
= prepend(buffer
, buflen
, "/", 1);
2555 * Filesystems needing to implement special "root names"
2556 * should do so with ->d_dname()
2558 if (IS_ROOT(dentry
) &&
2559 (dentry
->d_name
.len
!= 1 || dentry
->d_name
.name
[0] != '/')) {
2560 WARN(1, "Root dentry has weird name <%.*s>\n",
2561 (int) dentry
->d_name
.len
, dentry
->d_name
.name
);
2564 error
= prepend(buffer
, buflen
, "/", 1);
2566 error
= is_mounted(vfsmnt
) ? 1 : 2;
2571 * __d_path - return the path of a dentry
2572 * @path: the dentry/vfsmount to report
2573 * @root: root vfsmnt/dentry
2574 * @buf: buffer to return value in
2575 * @buflen: buffer length
2577 * Convert a dentry into an ASCII path name.
2579 * Returns a pointer into the buffer or an error code if the
2580 * path was too long.
2582 * "buflen" should be positive.
2584 * If the path is not reachable from the supplied root, return %NULL.
2586 char *__d_path(const struct path
*path
,
2587 const struct path
*root
,
2588 char *buf
, int buflen
)
2590 char *res
= buf
+ buflen
;
2593 prepend(&res
, &buflen
, "\0", 1);
2594 br_read_lock(&vfsmount_lock
);
2595 write_seqlock(&rename_lock
);
2596 error
= prepend_path(path
, root
, &res
, &buflen
);
2597 write_sequnlock(&rename_lock
);
2598 br_read_unlock(&vfsmount_lock
);
2601 return ERR_PTR(error
);
2607 char *d_absolute_path(const struct path
*path
,
2608 char *buf
, int buflen
)
2610 struct path root
= {};
2611 char *res
= buf
+ buflen
;
2614 prepend(&res
, &buflen
, "\0", 1);
2615 br_read_lock(&vfsmount_lock
);
2616 write_seqlock(&rename_lock
);
2617 error
= prepend_path(path
, &root
, &res
, &buflen
);
2618 write_sequnlock(&rename_lock
);
2619 br_read_unlock(&vfsmount_lock
);
2624 return ERR_PTR(error
);
2629 * same as __d_path but appends "(deleted)" for unlinked files.
2631 static int path_with_deleted(const struct path
*path
,
2632 const struct path
*root
,
2633 char **buf
, int *buflen
)
2635 prepend(buf
, buflen
, "\0", 1);
2636 if (d_unlinked(path
->dentry
)) {
2637 int error
= prepend(buf
, buflen
, " (deleted)", 10);
2642 return prepend_path(path
, root
, buf
, buflen
);
2645 static int prepend_unreachable(char **buffer
, int *buflen
)
2647 return prepend(buffer
, buflen
, "(unreachable)", 13);
2651 * d_path - return the path of a dentry
2652 * @path: path to report
2653 * @buf: buffer to return value in
2654 * @buflen: buffer length
2656 * Convert a dentry into an ASCII path name. If the entry has been deleted
2657 * the string " (deleted)" is appended. Note that this is ambiguous.
2659 * Returns a pointer into the buffer or an error code if the path was
2660 * too long. Note: Callers should use the returned pointer, not the passed
2661 * in buffer, to use the name! The implementation often starts at an offset
2662 * into the buffer, and may leave 0 bytes at the start.
2664 * "buflen" should be positive.
2666 char *d_path(const struct path
*path
, char *buf
, int buflen
)
2668 char *res
= buf
+ buflen
;
2673 * We have various synthetic filesystems that never get mounted. On
2674 * these filesystems dentries are never used for lookup purposes, and
2675 * thus don't need to be hashed. They also don't need a name until a
2676 * user wants to identify the object in /proc/pid/fd/. The little hack
2677 * below allows us to generate a name for these objects on demand:
2679 if (path
->dentry
->d_op
&& path
->dentry
->d_op
->d_dname
)
2680 return path
->dentry
->d_op
->d_dname(path
->dentry
, buf
, buflen
);
2682 get_fs_root(current
->fs
, &root
);
2683 br_read_lock(&vfsmount_lock
);
2684 write_seqlock(&rename_lock
);
2685 error
= path_with_deleted(path
, &root
, &res
, &buflen
);
2686 write_sequnlock(&rename_lock
);
2687 br_read_unlock(&vfsmount_lock
);
2689 res
= ERR_PTR(error
);
2693 EXPORT_SYMBOL(d_path
);
2696 * Helper function for dentry_operations.d_dname() members
2698 char *dynamic_dname(struct dentry
*dentry
, char *buffer
, int buflen
,
2699 const char *fmt
, ...)
2705 va_start(args
, fmt
);
2706 sz
= vsnprintf(temp
, sizeof(temp
), fmt
, args
) + 1;
2709 if (sz
> sizeof(temp
) || sz
> buflen
)
2710 return ERR_PTR(-ENAMETOOLONG
);
2712 buffer
+= buflen
- sz
;
2713 return memcpy(buffer
, temp
, sz
);
2716 char *simple_dname(struct dentry
*dentry
, char *buffer
, int buflen
)
2718 char *end
= buffer
+ buflen
;
2719 /* these dentries are never renamed, so d_lock is not needed */
2720 if (prepend(&end
, &buflen
, " (deleted)", 11) ||
2721 prepend_name(&end
, &buflen
, &dentry
->d_name
) ||
2722 prepend(&end
, &buflen
, "/", 1))
2723 end
= ERR_PTR(-ENAMETOOLONG
);
2728 * Write full pathname from the root of the filesystem into the buffer.
2730 static char *__dentry_path(struct dentry
*dentry
, char *buf
, int buflen
)
2732 char *end
= buf
+ buflen
;
2735 prepend(&end
, &buflen
, "\0", 1);
2742 while (!IS_ROOT(dentry
)) {
2743 struct dentry
*parent
= dentry
->d_parent
;
2747 spin_lock(&dentry
->d_lock
);
2748 error
= prepend_name(&end
, &buflen
, &dentry
->d_name
);
2749 spin_unlock(&dentry
->d_lock
);
2750 if (error
!= 0 || prepend(&end
, &buflen
, "/", 1) != 0)
2758 return ERR_PTR(-ENAMETOOLONG
);
2761 char *dentry_path_raw(struct dentry
*dentry
, char *buf
, int buflen
)
2765 write_seqlock(&rename_lock
);
2766 retval
= __dentry_path(dentry
, buf
, buflen
);
2767 write_sequnlock(&rename_lock
);
2771 EXPORT_SYMBOL(dentry_path_raw
);
2773 char *dentry_path(struct dentry
*dentry
, char *buf
, int buflen
)
2778 write_seqlock(&rename_lock
);
2779 if (d_unlinked(dentry
)) {
2781 if (prepend(&p
, &buflen
, "//deleted", 10) != 0)
2785 retval
= __dentry_path(dentry
, buf
, buflen
);
2786 write_sequnlock(&rename_lock
);
2787 if (!IS_ERR(retval
) && p
)
2788 *p
= '/'; /* restore '/' overriden with '\0' */
2791 return ERR_PTR(-ENAMETOOLONG
);
2795 * NOTE! The user-level library version returns a
2796 * character pointer. The kernel system call just
2797 * returns the length of the buffer filled (which
2798 * includes the ending '\0' character), or a negative
2799 * error value. So libc would do something like
2801 * char *getcwd(char * buf, size_t size)
2805 * retval = sys_getcwd(buf, size);
2812 SYSCALL_DEFINE2(getcwd
, char __user
*, buf
, unsigned long, size
)
2815 struct path pwd
, root
;
2816 char *page
= (char *) __get_free_page(GFP_USER
);
2821 get_fs_root_and_pwd(current
->fs
, &root
, &pwd
);
2824 br_read_lock(&vfsmount_lock
);
2825 write_seqlock(&rename_lock
);
2826 if (!d_unlinked(pwd
.dentry
)) {
2828 char *cwd
= page
+ PAGE_SIZE
;
2829 int buflen
= PAGE_SIZE
;
2831 prepend(&cwd
, &buflen
, "\0", 1);
2832 error
= prepend_path(&pwd
, &root
, &cwd
, &buflen
);
2833 write_sequnlock(&rename_lock
);
2834 br_read_unlock(&vfsmount_lock
);
2839 /* Unreachable from current root */
2841 error
= prepend_unreachable(&cwd
, &buflen
);
2847 len
= PAGE_SIZE
+ page
- cwd
;
2850 if (copy_to_user(buf
, cwd
, len
))
2854 write_sequnlock(&rename_lock
);
2855 br_read_unlock(&vfsmount_lock
);
2861 free_page((unsigned long) page
);
2866 * Test whether new_dentry is a subdirectory of old_dentry.
2868 * Trivially implemented using the dcache structure
2872 * is_subdir - is new dentry a subdirectory of old_dentry
2873 * @new_dentry: new dentry
2874 * @old_dentry: old dentry
2876 * Returns 1 if new_dentry is a subdirectory of the parent (at any depth).
2877 * Returns 0 otherwise.
2878 * Caller must ensure that "new_dentry" is pinned before calling is_subdir()
2881 int is_subdir(struct dentry
*new_dentry
, struct dentry
*old_dentry
)
2886 if (new_dentry
== old_dentry
)
2890 /* for restarting inner loop in case of seq retry */
2891 seq
= read_seqbegin(&rename_lock
);
2893 * Need rcu_readlock to protect against the d_parent trashing
2897 if (d_ancestor(old_dentry
, new_dentry
))
2902 } while (read_seqretry(&rename_lock
, seq
));
2907 void d_genocide(struct dentry
*root
)
2909 struct dentry
*this_parent
;
2910 struct list_head
*next
;
2914 seq
= read_seqbegin(&rename_lock
);
2917 spin_lock(&this_parent
->d_lock
);
2919 next
= this_parent
->d_subdirs
.next
;
2921 while (next
!= &this_parent
->d_subdirs
) {
2922 struct list_head
*tmp
= next
;
2923 struct dentry
*dentry
= list_entry(tmp
, struct dentry
, d_u
.d_child
);
2926 spin_lock_nested(&dentry
->d_lock
, DENTRY_D_LOCK_NESTED
);
2927 if (d_unhashed(dentry
) || !dentry
->d_inode
) {
2928 spin_unlock(&dentry
->d_lock
);
2931 if (!list_empty(&dentry
->d_subdirs
)) {
2932 spin_unlock(&this_parent
->d_lock
);
2933 spin_release(&dentry
->d_lock
.dep_map
, 1, _RET_IP_
);
2934 this_parent
= dentry
;
2935 spin_acquire(&this_parent
->d_lock
.dep_map
, 0, 1, _RET_IP_
);
2938 if (!(dentry
->d_flags
& DCACHE_GENOCIDE
)) {
2939 dentry
->d_flags
|= DCACHE_GENOCIDE
;
2940 dentry
->d_lockref
.count
--;
2942 spin_unlock(&dentry
->d_lock
);
2944 if (this_parent
!= root
) {
2945 struct dentry
*child
= this_parent
;
2946 if (!(this_parent
->d_flags
& DCACHE_GENOCIDE
)) {
2947 this_parent
->d_flags
|= DCACHE_GENOCIDE
;
2948 this_parent
->d_lockref
.count
--;
2950 this_parent
= try_to_ascend(this_parent
, locked
, seq
);
2953 next
= child
->d_u
.d_child
.next
;
2956 spin_unlock(&this_parent
->d_lock
);
2957 if (!locked
&& read_seqretry(&rename_lock
, seq
))
2960 write_sequnlock(&rename_lock
);
2967 write_seqlock(&rename_lock
);
2971 void d_tmpfile(struct dentry
*dentry
, struct inode
*inode
)
2973 inode_dec_link_count(inode
);
2974 BUG_ON(dentry
->d_name
.name
!= dentry
->d_iname
||
2975 !hlist_unhashed(&dentry
->d_alias
) ||
2976 !d_unlinked(dentry
));
2977 spin_lock(&dentry
->d_parent
->d_lock
);
2978 spin_lock_nested(&dentry
->d_lock
, DENTRY_D_LOCK_NESTED
);
2979 dentry
->d_name
.len
= sprintf(dentry
->d_iname
, "#%llu",
2980 (unsigned long long)inode
->i_ino
);
2981 spin_unlock(&dentry
->d_lock
);
2982 spin_unlock(&dentry
->d_parent
->d_lock
);
2983 d_instantiate(dentry
, inode
);
2985 EXPORT_SYMBOL(d_tmpfile
);
2987 static __initdata
unsigned long dhash_entries
;
2988 static int __init
set_dhash_entries(char *str
)
2992 dhash_entries
= simple_strtoul(str
, &str
, 0);
2995 __setup("dhash_entries=", set_dhash_entries
);
2997 static void __init
dcache_init_early(void)
3001 /* If hashes are distributed across NUMA nodes, defer
3002 * hash allocation until vmalloc space is available.
3008 alloc_large_system_hash("Dentry cache",
3009 sizeof(struct hlist_bl_head
),
3018 for (loop
= 0; loop
< (1U << d_hash_shift
); loop
++)
3019 INIT_HLIST_BL_HEAD(dentry_hashtable
+ loop
);
3022 static void __init
dcache_init(void)
3027 * A constructor could be added for stable state like the lists,
3028 * but it is probably not worth it because of the cache nature
3031 dentry_cache
= KMEM_CACHE(dentry
,
3032 SLAB_RECLAIM_ACCOUNT
|SLAB_PANIC
|SLAB_MEM_SPREAD
);
3034 /* Hash may have been set up in dcache_init_early */
3039 alloc_large_system_hash("Dentry cache",
3040 sizeof(struct hlist_bl_head
),
3049 for (loop
= 0; loop
< (1U << d_hash_shift
); loop
++)
3050 INIT_HLIST_BL_HEAD(dentry_hashtable
+ loop
);
3053 /* SLAB cache for __getname() consumers */
3054 struct kmem_cache
*names_cachep __read_mostly
;
3055 EXPORT_SYMBOL(names_cachep
);
3057 EXPORT_SYMBOL(d_genocide
);
3059 void __init
vfs_caches_init_early(void)
3061 dcache_init_early();
3065 void __init
vfs_caches_init(unsigned long mempages
)
3067 unsigned long reserve
;
3069 /* Base hash sizes on available memory, with a reserve equal to
3070 150% of current kernel size */
3072 reserve
= min((mempages
- nr_free_pages()) * 3/2, mempages
- 1);
3073 mempages
-= reserve
;
3075 names_cachep
= kmem_cache_create("names_cache", PATH_MAX
, 0,
3076 SLAB_HWCACHE_ALIGN
|SLAB_PANIC
, NULL
);
3080 files_init(mempages
);