Doug / smarc-fsl-linux-kernel

Download as
Browse Code ยป

Commit 5762482f5496cb1dd86acd2aace3ea25d1404e1f

Authored by Linus Torvalds
1 parent b7c09ad401
Exists in smarc-imx_3.14.28_1.0.0_ga and in 1 other branch imx_3.14.28_1.0.0_ga

vfs: move get_fs_root_and_pwd() to single caller 

Let's not pollute the include files with inline functions that are only
used in a single place.  Especially not if we decide we might want to
change the semantics of said function to make it more efficient..

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 2 changed files with 11 additions and 11 deletions Inline Diff

1 /* 1 /*
2 * fs/dcache.c 2 * fs/dcache.c
3 * 3 *
4 * Complete reimplementation 4 * Complete reimplementation
5 * (C) 1997 Thomas Schoebel-Theuer, 5 * (C) 1997 Thomas Schoebel-Theuer,
6 * with heavy changes by Linus Torvalds 6 * with heavy changes by Linus Torvalds
7 */ 7 */
8 8
9 /* 9 /*
10 * Notes on the allocation strategy: 10 * Notes on the allocation strategy:
11 * 11 *
12 * The dcache is a master of the icache - whenever a dcache entry 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 13 * exists, the inode will always exist. "iput()" is done either when
14 * the dcache entry is deleted or garbage collected. 14 * the dcache entry is deleted or garbage collected.
15 */ 15 */
16 16
17 #include <linux/syscalls.h> 17 #include <linux/syscalls.h>
18 #include <linux/string.h> 18 #include <linux/string.h>
19 #include <linux/mm.h> 19 #include <linux/mm.h>
20 #include <linux/fs.h> 20 #include <linux/fs.h>
21 #include <linux/fsnotify.h> 21 #include <linux/fsnotify.h>
22 #include <linux/slab.h> 22 #include <linux/slab.h>
23 #include <linux/init.h> 23 #include <linux/init.h>
24 #include <linux/hash.h> 24 #include <linux/hash.h>
25 #include <linux/cache.h> 25 #include <linux/cache.h>
26 #include <linux/export.h> 26 #include <linux/export.h>
27 #include <linux/mount.h> 27 #include <linux/mount.h>
28 #include <linux/file.h> 28 #include <linux/file.h>
29 #include <asm/uaccess.h> 29 #include <asm/uaccess.h>
30 #include <linux/security.h> 30 #include <linux/security.h>
31 #include <linux/seqlock.h> 31 #include <linux/seqlock.h>
32 #include <linux/swap.h> 32 #include <linux/swap.h>
33 #include <linux/bootmem.h> 33 #include <linux/bootmem.h>
34 #include <linux/fs_struct.h> 34 #include <linux/fs_struct.h>
35 #include <linux/hardirq.h> 35 #include <linux/hardirq.h>
36 #include <linux/bit_spinlock.h> 36 #include <linux/bit_spinlock.h>
37 #include <linux/rculist_bl.h> 37 #include <linux/rculist_bl.h>
38 #include <linux/prefetch.h> 38 #include <linux/prefetch.h>
39 #include <linux/ratelimit.h> 39 #include <linux/ratelimit.h>
40 #include "internal.h" 40 #include "internal.h"
41 #include "mount.h" 41 #include "mount.h"
42 42
43 /* 43 /*
44 * Usage: 44 * Usage:
45 * dcache->d_inode->i_lock protects: 45 * dcache->d_inode->i_lock protects:
46 * - i_dentry, d_alias, d_inode of aliases 46 * - i_dentry, d_alias, d_inode of aliases
47 * dcache_hash_bucket lock protects: 47 * dcache_hash_bucket lock protects:
48 * - the dcache hash table 48 * - the dcache hash table
49 * s_anon bl list spinlock protects: 49 * s_anon bl list spinlock protects:
50 * - the s_anon list (see __d_drop) 50 * - the s_anon list (see __d_drop)
51 * dcache_lru_lock protects: 51 * dcache_lru_lock protects:
52 * - the dcache lru lists and counters 52 * - the dcache lru lists and counters
53 * d_lock protects: 53 * d_lock protects:
54 * - d_flags 54 * - d_flags
55 * - d_name 55 * - d_name
56 * - d_lru 56 * - d_lru
57 * - d_count 57 * - d_count
58 * - d_unhashed() 58 * - d_unhashed()
59 * - d_parent and d_subdirs 59 * - d_parent and d_subdirs
60 * - childrens' d_child and d_parent 60 * - childrens' d_child and d_parent
61 * - d_alias, d_inode 61 * - d_alias, d_inode
62 * 62 *
63 * Ordering: 63 * Ordering:
64 * dentry->d_inode->i_lock 64 * dentry->d_inode->i_lock
65 * dentry->d_lock 65 * dentry->d_lock
66 * dcache_lru_lock 66 * dcache_lru_lock
67 * dcache_hash_bucket lock 67 * dcache_hash_bucket lock
68 * s_anon lock 68 * s_anon lock
69 * 69 *
70 * If there is an ancestor relationship: 70 * If there is an ancestor relationship:
71 * dentry->d_parent->...->d_parent->d_lock 71 * dentry->d_parent->...->d_parent->d_lock
72 * ... 72 * ...
73 * dentry->d_parent->d_lock 73 * dentry->d_parent->d_lock
74 * dentry->d_lock 74 * dentry->d_lock
75 * 75 *
76 * If no ancestor relationship: 76 * If no ancestor relationship:
77 * if (dentry1 < dentry2) 77 * if (dentry1 < dentry2)
78 * dentry1->d_lock 78 * dentry1->d_lock
79 * dentry2->d_lock 79 * dentry2->d_lock
80 */ 80 */
81 int sysctl_vfs_cache_pressure __read_mostly = 100; 81 int sysctl_vfs_cache_pressure __read_mostly = 100;
82 EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure); 82 EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure);
83 83
84 static __cacheline_aligned_in_smp DEFINE_SPINLOCK(dcache_lru_lock); 84 static __cacheline_aligned_in_smp DEFINE_SPINLOCK(dcache_lru_lock);
85 __cacheline_aligned_in_smp DEFINE_SEQLOCK(rename_lock); 85 __cacheline_aligned_in_smp DEFINE_SEQLOCK(rename_lock);
86 86
87 EXPORT_SYMBOL(rename_lock); 87 EXPORT_SYMBOL(rename_lock);
88 88
89 static struct kmem_cache *dentry_cache __read_mostly; 89 static struct kmem_cache *dentry_cache __read_mostly;
90 90
91 /** 91 /**
92 * read_seqbegin_or_lock - begin a sequence number check or locking block 92 * read_seqbegin_or_lock - begin a sequence number check or locking block
93 * @lock: sequence lock 93 * @lock: sequence lock
94 * @seq : sequence number to be checked 94 * @seq : sequence number to be checked
95 * 95 *
96 * First try it once optimistically without taking the lock. If that fails, 96 * First try it once optimistically without taking the lock. If that fails,
97 * take the lock. The sequence number is also used as a marker for deciding 97 * take the lock. The sequence number is also used as a marker for deciding
98 * whether to be a reader (even) or writer (odd). 98 * whether to be a reader (even) or writer (odd).
99 * N.B. seq must be initialized to an even number to begin with. 99 * N.B. seq must be initialized to an even number to begin with.
100 */ 100 */
101 static inline void read_seqbegin_or_lock(seqlock_t *lock, int *seq) 101 static inline void read_seqbegin_or_lock(seqlock_t *lock, int *seq)
102 { 102 {
103 if (!(*seq & 1)) /* Even */ 103 if (!(*seq & 1)) /* Even */
104 *seq = read_seqbegin(lock); 104 *seq = read_seqbegin(lock);
105 else /* Odd */ 105 else /* Odd */
106 read_seqlock_excl(lock); 106 read_seqlock_excl(lock);
107 } 107 }
108 108
109 static inline int need_seqretry(seqlock_t *lock, int seq) 109 static inline int need_seqretry(seqlock_t *lock, int seq)
110 { 110 {
111 return !(seq & 1) && read_seqretry(lock, seq); 111 return !(seq & 1) && read_seqretry(lock, seq);
112 } 112 }
113 113
114 static inline void done_seqretry(seqlock_t *lock, int seq) 114 static inline void done_seqretry(seqlock_t *lock, int seq)
115 { 115 {
116 if (seq & 1) 116 if (seq & 1)
117 read_sequnlock_excl(lock); 117 read_sequnlock_excl(lock);
118 } 118 }
119 119
120 /* 120 /*
121 * This is the single most critical data structure when it comes 121 * This is the single most critical data structure when it comes
122 * to the dcache: the hashtable for lookups. Somebody should try 122 * to the dcache: the hashtable for lookups. Somebody should try
123 * to make this good - I've just made it work. 123 * to make this good - I've just made it work.
124 * 124 *
125 * This hash-function tries to avoid losing too many bits of hash 125 * This hash-function tries to avoid losing too many bits of hash
126 * information, yet avoid using a prime hash-size or similar. 126 * information, yet avoid using a prime hash-size or similar.
127 */ 127 */
128 #define D_HASHBITS d_hash_shift 128 #define D_HASHBITS d_hash_shift
129 #define D_HASHMASK d_hash_mask 129 #define D_HASHMASK d_hash_mask
130 130
131 static unsigned int d_hash_mask __read_mostly; 131 static unsigned int d_hash_mask __read_mostly;
132 static unsigned int d_hash_shift __read_mostly; 132 static unsigned int d_hash_shift __read_mostly;
133 133
134 static struct hlist_bl_head *dentry_hashtable __read_mostly; 134 static struct hlist_bl_head *dentry_hashtable __read_mostly;
135 135
136 static inline struct hlist_bl_head *d_hash(const struct dentry *parent, 136 static inline struct hlist_bl_head *d_hash(const struct dentry *parent,
137 unsigned int hash) 137 unsigned int hash)
138 { 138 {
139 hash += (unsigned long) parent / L1_CACHE_BYTES; 139 hash += (unsigned long) parent / L1_CACHE_BYTES;
140 hash = hash + (hash >> D_HASHBITS); 140 hash = hash + (hash >> D_HASHBITS);
141 return dentry_hashtable + (hash & D_HASHMASK); 141 return dentry_hashtable + (hash & D_HASHMASK);
142 } 142 }
143 143
144 /* Statistics gathering. */ 144 /* Statistics gathering. */
145 struct dentry_stat_t dentry_stat = { 145 struct dentry_stat_t dentry_stat = {
146 .age_limit = 45, 146 .age_limit = 45,
147 }; 147 };
148 148
149 static DEFINE_PER_CPU(unsigned int, nr_dentry); 149 static DEFINE_PER_CPU(unsigned int, nr_dentry);
150 150
151 #if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS) 151 #if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
152 static int get_nr_dentry(void) 152 static int get_nr_dentry(void)
153 { 153 {
154 int i; 154 int i;
155 int sum = 0; 155 int sum = 0;
156 for_each_possible_cpu(i) 156 for_each_possible_cpu(i)
157 sum += per_cpu(nr_dentry, i); 157 sum += per_cpu(nr_dentry, i);
158 return sum < 0 ? 0 : sum; 158 return sum < 0 ? 0 : sum;
159 } 159 }
160 160
161 int proc_nr_dentry(ctl_table *table, int write, void __user *buffer, 161 int proc_nr_dentry(ctl_table *table, int write, void __user *buffer,
162 size_t *lenp, loff_t *ppos) 162 size_t *lenp, loff_t *ppos)
163 { 163 {
164 dentry_stat.nr_dentry = get_nr_dentry(); 164 dentry_stat.nr_dentry = get_nr_dentry();
165 return proc_dointvec(table, write, buffer, lenp, ppos); 165 return proc_dointvec(table, write, buffer, lenp, ppos);
166 } 166 }
167 #endif 167 #endif
168 168
169 /* 169 /*
170 * Compare 2 name strings, return 0 if they match, otherwise non-zero. 170 * Compare 2 name strings, return 0 if they match, otherwise non-zero.
171 * The strings are both count bytes long, and count is non-zero. 171 * The strings are both count bytes long, and count is non-zero.
172 */ 172 */
173 #ifdef CONFIG_DCACHE_WORD_ACCESS 173 #ifdef CONFIG_DCACHE_WORD_ACCESS
174 174
175 #include <asm/word-at-a-time.h> 175 #include <asm/word-at-a-time.h>
176 /* 176 /*
177 * NOTE! 'cs' and 'scount' come from a dentry, so it has a 177 * NOTE! 'cs' and 'scount' come from a dentry, so it has a
178 * aligned allocation for this particular component. We don't 178 * aligned allocation for this particular component. We don't
179 * strictly need the load_unaligned_zeropad() safety, but it 179 * strictly need the load_unaligned_zeropad() safety, but it
180 * doesn't hurt either. 180 * doesn't hurt either.
181 * 181 *
182 * In contrast, 'ct' and 'tcount' can be from a pathname, and do 182 * In contrast, 'ct' and 'tcount' can be from a pathname, and do
183 * need the careful unaligned handling. 183 * need the careful unaligned handling.
184 */ 184 */
185 static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount) 185 static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount)
186 { 186 {
187 unsigned long a,b,mask; 187 unsigned long a,b,mask;
188 188
189 for (;;) { 189 for (;;) {
190 a = *(unsigned long *)cs; 190 a = *(unsigned long *)cs;
191 b = load_unaligned_zeropad(ct); 191 b = load_unaligned_zeropad(ct);
192 if (tcount < sizeof(unsigned long)) 192 if (tcount < sizeof(unsigned long))
193 break; 193 break;
194 if (unlikely(a != b)) 194 if (unlikely(a != b))
195 return 1; 195 return 1;
196 cs += sizeof(unsigned long); 196 cs += sizeof(unsigned long);
197 ct += sizeof(unsigned long); 197 ct += sizeof(unsigned long);
198 tcount -= sizeof(unsigned long); 198 tcount -= sizeof(unsigned long);
199 if (!tcount) 199 if (!tcount)
200 return 0; 200 return 0;
201 } 201 }
202 mask = ~(~0ul << tcount*8); 202 mask = ~(~0ul << tcount*8);
203 return unlikely(!!((a ^ b) & mask)); 203 return unlikely(!!((a ^ b) & mask));
204 } 204 }
205 205
206 #else 206 #else
207 207
208 static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount) 208 static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount)
209 { 209 {
210 do { 210 do {
211 if (*cs != *ct) 211 if (*cs != *ct)
212 return 1; 212 return 1;
213 cs++; 213 cs++;
214 ct++; 214 ct++;
215 tcount--; 215 tcount--;
216 } while (tcount); 216 } while (tcount);
217 return 0; 217 return 0;
218 } 218 }
219 219
220 #endif 220 #endif
221 221
222 static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *ct, unsigned tcount) 222 static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *ct, unsigned tcount)
223 { 223 {
224 const unsigned char *cs; 224 const unsigned char *cs;
225 /* 225 /*
226 * Be careful about RCU walk racing with rename: 226 * Be careful about RCU walk racing with rename:
227 * use ACCESS_ONCE to fetch the name pointer. 227 * use ACCESS_ONCE to fetch the name pointer.
228 * 228 *
229 * NOTE! Even if a rename will mean that the length 229 * NOTE! Even if a rename will mean that the length
230 * was not loaded atomically, we don't care. The 230 * was not loaded atomically, we don't care. The
231 * RCU walk will check the sequence count eventually, 231 * RCU walk will check the sequence count eventually,
232 * and catch it. And we won't overrun the buffer, 232 * and catch it. And we won't overrun the buffer,
233 * because we're reading the name pointer atomically, 233 * because we're reading the name pointer atomically,
234 * and a dentry name is guaranteed to be properly 234 * and a dentry name is guaranteed to be properly
235 * terminated with a NUL byte. 235 * terminated with a NUL byte.
236 * 236 *
237 * End result: even if 'len' is wrong, we'll exit 237 * End result: even if 'len' is wrong, we'll exit
238 * early because the data cannot match (there can 238 * early because the data cannot match (there can
239 * be no NUL in the ct/tcount data) 239 * be no NUL in the ct/tcount data)
240 */ 240 */
241 cs = ACCESS_ONCE(dentry->d_name.name); 241 cs = ACCESS_ONCE(dentry->d_name.name);
242 smp_read_barrier_depends(); 242 smp_read_barrier_depends();
243 return dentry_string_cmp(cs, ct, tcount); 243 return dentry_string_cmp(cs, ct, tcount);
244 } 244 }
245 245
246 static void __d_free(struct rcu_head *head) 246 static void __d_free(struct rcu_head *head)
247 { 247 {
248 struct dentry *dentry = container_of(head, struct dentry, d_u.d_rcu); 248 struct dentry *dentry = container_of(head, struct dentry, d_u.d_rcu);
249 249
250 WARN_ON(!hlist_unhashed(&dentry->d_alias)); 250 WARN_ON(!hlist_unhashed(&dentry->d_alias));
251 if (dname_external(dentry)) 251 if (dname_external(dentry))
252 kfree(dentry->d_name.name); 252 kfree(dentry->d_name.name);
253 kmem_cache_free(dentry_cache, dentry); 253 kmem_cache_free(dentry_cache, dentry);
254 } 254 }
255 255
256 /* 256 /*
257 * no locks, please. 257 * no locks, please.
258 */ 258 */
259 static void d_free(struct dentry *dentry) 259 static void d_free(struct dentry *dentry)
260 { 260 {
261 BUG_ON((int)dentry->d_lockref.count > 0); 261 BUG_ON((int)dentry->d_lockref.count > 0);
262 this_cpu_dec(nr_dentry); 262 this_cpu_dec(nr_dentry);
263 if (dentry->d_op && dentry->d_op->d_release) 263 if (dentry->d_op && dentry->d_op->d_release)
264 dentry->d_op->d_release(dentry); 264 dentry->d_op->d_release(dentry);
265 265
266 /* if dentry was never visible to RCU, immediate free is OK */ 266 /* if dentry was never visible to RCU, immediate free is OK */
267 if (!(dentry->d_flags & DCACHE_RCUACCESS)) 267 if (!(dentry->d_flags & DCACHE_RCUACCESS))
268 __d_free(&dentry->d_u.d_rcu); 268 __d_free(&dentry->d_u.d_rcu);
269 else 269 else
270 call_rcu(&dentry->d_u.d_rcu, __d_free); 270 call_rcu(&dentry->d_u.d_rcu, __d_free);
271 } 271 }
272 272
273 /** 273 /**
274 * dentry_rcuwalk_barrier - invalidate in-progress rcu-walk lookups 274 * dentry_rcuwalk_barrier - invalidate in-progress rcu-walk lookups
275 * @dentry: the target dentry 275 * @dentry: the target dentry
276 * After this call, in-progress rcu-walk path lookup will fail. This 276 * After this call, in-progress rcu-walk path lookup will fail. This
277 * should be called after unhashing, and after changing d_inode (if 277 * should be called after unhashing, and after changing d_inode (if
278 * the dentry has not already been unhashed). 278 * the dentry has not already been unhashed).
279 */ 279 */
280 static inline void dentry_rcuwalk_barrier(struct dentry *dentry) 280 static inline void dentry_rcuwalk_barrier(struct dentry *dentry)
281 { 281 {
282 assert_spin_locked(&dentry->d_lock); 282 assert_spin_locked(&dentry->d_lock);
283 /* Go through a barrier */ 283 /* Go through a barrier */
284 write_seqcount_barrier(&dentry->d_seq); 284 write_seqcount_barrier(&dentry->d_seq);
285 } 285 }
286 286
287 /* 287 /*
288 * Release the dentry's inode, using the filesystem 288 * Release the dentry's inode, using the filesystem
289 * d_iput() operation if defined. Dentry has no refcount 289 * d_iput() operation if defined. Dentry has no refcount
290 * and is unhashed. 290 * and is unhashed.
291 */ 291 */
292 static void dentry_iput(struct dentry * dentry) 292 static void dentry_iput(struct dentry * dentry)
293 __releases(dentry->d_lock) 293 __releases(dentry->d_lock)
294 __releases(dentry->d_inode->i_lock) 294 __releases(dentry->d_inode->i_lock)
295 { 295 {
296 struct inode *inode = dentry->d_inode; 296 struct inode *inode = dentry->d_inode;
297 if (inode) { 297 if (inode) {
298 dentry->d_inode = NULL; 298 dentry->d_inode = NULL;
299 hlist_del_init(&dentry->d_alias); 299 hlist_del_init(&dentry->d_alias);
300 spin_unlock(&dentry->d_lock); 300 spin_unlock(&dentry->d_lock);
301 spin_unlock(&inode->i_lock); 301 spin_unlock(&inode->i_lock);
302 if (!inode->i_nlink) 302 if (!inode->i_nlink)
303 fsnotify_inoderemove(inode); 303 fsnotify_inoderemove(inode);
304 if (dentry->d_op && dentry->d_op->d_iput) 304 if (dentry->d_op && dentry->d_op->d_iput)
305 dentry->d_op->d_iput(dentry, inode); 305 dentry->d_op->d_iput(dentry, inode);
306 else 306 else
307 iput(inode); 307 iput(inode);
308 } else { 308 } else {
309 spin_unlock(&dentry->d_lock); 309 spin_unlock(&dentry->d_lock);
310 } 310 }
311 } 311 }
312 312
313 /* 313 /*
314 * Release the dentry's inode, using the filesystem 314 * Release the dentry's inode, using the filesystem
315 * d_iput() operation if defined. dentry remains in-use. 315 * d_iput() operation if defined. dentry remains in-use.
316 */ 316 */
317 static void dentry_unlink_inode(struct dentry * dentry) 317 static void dentry_unlink_inode(struct dentry * dentry)
318 __releases(dentry->d_lock) 318 __releases(dentry->d_lock)
319 __releases(dentry->d_inode->i_lock) 319 __releases(dentry->d_inode->i_lock)
320 { 320 {
321 struct inode *inode = dentry->d_inode; 321 struct inode *inode = dentry->d_inode;
322 dentry->d_inode = NULL; 322 dentry->d_inode = NULL;
323 hlist_del_init(&dentry->d_alias); 323 hlist_del_init(&dentry->d_alias);
324 dentry_rcuwalk_barrier(dentry); 324 dentry_rcuwalk_barrier(dentry);
325 spin_unlock(&dentry->d_lock); 325 spin_unlock(&dentry->d_lock);
326 spin_unlock(&inode->i_lock); 326 spin_unlock(&inode->i_lock);
327 if (!inode->i_nlink) 327 if (!inode->i_nlink)
328 fsnotify_inoderemove(inode); 328 fsnotify_inoderemove(inode);
329 if (dentry->d_op && dentry->d_op->d_iput) 329 if (dentry->d_op && dentry->d_op->d_iput)
330 dentry->d_op->d_iput(dentry, inode); 330 dentry->d_op->d_iput(dentry, inode);
331 else 331 else
332 iput(inode); 332 iput(inode);
333 } 333 }
334 334
335 /* 335 /*
336 * dentry_lru_(add|del|prune|move_tail) must be called with d_lock held. 336 * dentry_lru_(add|del|prune|move_tail) must be called with d_lock held.
337 */ 337 */
338 static void dentry_lru_add(struct dentry *dentry) 338 static void dentry_lru_add(struct dentry *dentry)
339 { 339 {
340 if (unlikely(!(dentry->d_flags & DCACHE_LRU_LIST))) { 340 if (unlikely(!(dentry->d_flags & DCACHE_LRU_LIST))) {
341 spin_lock(&dcache_lru_lock); 341 spin_lock(&dcache_lru_lock);
342 dentry->d_flags |= DCACHE_LRU_LIST; 342 dentry->d_flags |= DCACHE_LRU_LIST;
343 list_add(&dentry->d_lru, &dentry->d_sb->s_dentry_lru); 343 list_add(&dentry->d_lru, &dentry->d_sb->s_dentry_lru);
344 dentry->d_sb->s_nr_dentry_unused++; 344 dentry->d_sb->s_nr_dentry_unused++;
345 dentry_stat.nr_unused++; 345 dentry_stat.nr_unused++;
346 spin_unlock(&dcache_lru_lock); 346 spin_unlock(&dcache_lru_lock);
347 } 347 }
348 } 348 }
349 349
350 static void __dentry_lru_del(struct dentry *dentry) 350 static void __dentry_lru_del(struct dentry *dentry)
351 { 351 {
352 list_del_init(&dentry->d_lru); 352 list_del_init(&dentry->d_lru);
353 dentry->d_flags &= ~(DCACHE_SHRINK_LIST | DCACHE_LRU_LIST); 353 dentry->d_flags &= ~(DCACHE_SHRINK_LIST | DCACHE_LRU_LIST);
354 dentry->d_sb->s_nr_dentry_unused--; 354 dentry->d_sb->s_nr_dentry_unused--;
355 dentry_stat.nr_unused--; 355 dentry_stat.nr_unused--;
356 } 356 }
357 357
358 /* 358 /*
359 * Remove a dentry with references from the LRU. 359 * Remove a dentry with references from the LRU.
360 */ 360 */
361 static void dentry_lru_del(struct dentry *dentry) 361 static void dentry_lru_del(struct dentry *dentry)
362 { 362 {
363 if (!list_empty(&dentry->d_lru)) { 363 if (!list_empty(&dentry->d_lru)) {
364 spin_lock(&dcache_lru_lock); 364 spin_lock(&dcache_lru_lock);
365 __dentry_lru_del(dentry); 365 __dentry_lru_del(dentry);
366 spin_unlock(&dcache_lru_lock); 366 spin_unlock(&dcache_lru_lock);
367 } 367 }
368 } 368 }
369 369
370 static void dentry_lru_move_list(struct dentry *dentry, struct list_head *list) 370 static void dentry_lru_move_list(struct dentry *dentry, struct list_head *list)
371 { 371 {
372 spin_lock(&dcache_lru_lock); 372 spin_lock(&dcache_lru_lock);
373 if (list_empty(&dentry->d_lru)) { 373 if (list_empty(&dentry->d_lru)) {
374 dentry->d_flags |= DCACHE_LRU_LIST; 374 dentry->d_flags |= DCACHE_LRU_LIST;
375 list_add_tail(&dentry->d_lru, list); 375 list_add_tail(&dentry->d_lru, list);
376 dentry->d_sb->s_nr_dentry_unused++; 376 dentry->d_sb->s_nr_dentry_unused++;
377 dentry_stat.nr_unused++; 377 dentry_stat.nr_unused++;
378 } else { 378 } else {
379 list_move_tail(&dentry->d_lru, list); 379 list_move_tail(&dentry->d_lru, list);
380 } 380 }
381 spin_unlock(&dcache_lru_lock); 381 spin_unlock(&dcache_lru_lock);
382 } 382 }
383 383
384 /** 384 /**
385 * d_kill - kill dentry and return parent 385 * d_kill - kill dentry and return parent
386 * @dentry: dentry to kill 386 * @dentry: dentry to kill
387 * @parent: parent dentry 387 * @parent: parent dentry
388 * 388 *
389 * The dentry must already be unhashed and removed from the LRU. 389 * The dentry must already be unhashed and removed from the LRU.
390 * 390 *
391 * If this is the root of the dentry tree, return NULL. 391 * If this is the root of the dentry tree, return NULL.
392 * 392 *
393 * dentry->d_lock and parent->d_lock must be held by caller, and are dropped by 393 * dentry->d_lock and parent->d_lock must be held by caller, and are dropped by
394 * d_kill. 394 * d_kill.
395 */ 395 */
396 static struct dentry *d_kill(struct dentry *dentry, struct dentry *parent) 396 static struct dentry *d_kill(struct dentry *dentry, struct dentry *parent)
397 __releases(dentry->d_lock) 397 __releases(dentry->d_lock)
398 __releases(parent->d_lock) 398 __releases(parent->d_lock)
399 __releases(dentry->d_inode->i_lock) 399 __releases(dentry->d_inode->i_lock)
400 { 400 {
401 list_del(&dentry->d_u.d_child); 401 list_del(&dentry->d_u.d_child);
402 /* 402 /*
403 * Inform try_to_ascend() that we are no longer attached to the 403 * Inform try_to_ascend() that we are no longer attached to the
404 * dentry tree 404 * dentry tree
405 */ 405 */
406 dentry->d_flags |= DCACHE_DENTRY_KILLED; 406 dentry->d_flags |= DCACHE_DENTRY_KILLED;
407 if (parent) 407 if (parent)
408 spin_unlock(&parent->d_lock); 408 spin_unlock(&parent->d_lock);
409 dentry_iput(dentry); 409 dentry_iput(dentry);
410 /* 410 /*
411 * dentry_iput drops the locks, at which point nobody (except 411 * dentry_iput drops the locks, at which point nobody (except
412 * transient RCU lookups) can reach this dentry. 412 * transient RCU lookups) can reach this dentry.
413 */ 413 */
414 d_free(dentry); 414 d_free(dentry);
415 return parent; 415 return parent;
416 } 416 }
417 417
418 /* 418 /*
419 * Unhash a dentry without inserting an RCU walk barrier or checking that 419 * Unhash a dentry without inserting an RCU walk barrier or checking that
420 * dentry->d_lock is locked. The caller must take care of that, if 420 * dentry->d_lock is locked. The caller must take care of that, if
421 * appropriate. 421 * appropriate.
422 */ 422 */
423 static void __d_shrink(struct dentry *dentry) 423 static void __d_shrink(struct dentry *dentry)
424 { 424 {
425 if (!d_unhashed(dentry)) { 425 if (!d_unhashed(dentry)) {
426 struct hlist_bl_head *b; 426 struct hlist_bl_head *b;
427 if (unlikely(dentry->d_flags & DCACHE_DISCONNECTED)) 427 if (unlikely(dentry->d_flags & DCACHE_DISCONNECTED))
428 b = &dentry->d_sb->s_anon; 428 b = &dentry->d_sb->s_anon;
429 else 429 else
430 b = d_hash(dentry->d_parent, dentry->d_name.hash); 430 b = d_hash(dentry->d_parent, dentry->d_name.hash);
431 431
432 hlist_bl_lock(b); 432 hlist_bl_lock(b);
433 __hlist_bl_del(&dentry->d_hash); 433 __hlist_bl_del(&dentry->d_hash);
434 dentry->d_hash.pprev = NULL; 434 dentry->d_hash.pprev = NULL;
435 hlist_bl_unlock(b); 435 hlist_bl_unlock(b);
436 } 436 }
437 } 437 }
438 438
439 /** 439 /**
440 * d_drop - drop a dentry 440 * d_drop - drop a dentry
441 * @dentry: dentry to drop 441 * @dentry: dentry to drop
442 * 442 *
443 * d_drop() unhashes the entry from the parent dentry hashes, so that it won't 443 * d_drop() unhashes the entry from the parent dentry hashes, so that it won't
444 * be found through a VFS lookup any more. Note that this is different from 444 * be found through a VFS lookup any more. Note that this is different from
445 * deleting the dentry - d_delete will try to mark the dentry negative if 445 * deleting the dentry - d_delete will try to mark the dentry negative if
446 * possible, giving a successful _negative_ lookup, while d_drop will 446 * possible, giving a successful _negative_ lookup, while d_drop will
447 * just make the cache lookup fail. 447 * just make the cache lookup fail.
448 * 448 *
449 * d_drop() is used mainly for stuff that wants to invalidate a dentry for some 449 * d_drop() is used mainly for stuff that wants to invalidate a dentry for some
450 * reason (NFS timeouts or autofs deletes). 450 * reason (NFS timeouts or autofs deletes).
451 * 451 *
452 * __d_drop requires dentry->d_lock. 452 * __d_drop requires dentry->d_lock.
453 */ 453 */
454 void __d_drop(struct dentry *dentry) 454 void __d_drop(struct dentry *dentry)
455 { 455 {
456 if (!d_unhashed(dentry)) { 456 if (!d_unhashed(dentry)) {
457 __d_shrink(dentry); 457 __d_shrink(dentry);
458 dentry_rcuwalk_barrier(dentry); 458 dentry_rcuwalk_barrier(dentry);
459 } 459 }
460 } 460 }
461 EXPORT_SYMBOL(__d_drop); 461 EXPORT_SYMBOL(__d_drop);
462 462
463 void d_drop(struct dentry *dentry) 463 void d_drop(struct dentry *dentry)
464 { 464 {
465 spin_lock(&dentry->d_lock); 465 spin_lock(&dentry->d_lock);
466 __d_drop(dentry); 466 __d_drop(dentry);
467 spin_unlock(&dentry->d_lock); 467 spin_unlock(&dentry->d_lock);
468 } 468 }
469 EXPORT_SYMBOL(d_drop); 469 EXPORT_SYMBOL(d_drop);
470 470
471 /* 471 /*
472 * Finish off a dentry we've decided to kill. 472 * Finish off a dentry we've decided to kill.
473 * dentry->d_lock must be held, returns with it unlocked. 473 * dentry->d_lock must be held, returns with it unlocked.
474 * If ref is non-zero, then decrement the refcount too. 474 * If ref is non-zero, then decrement the refcount too.
475 * Returns dentry requiring refcount drop, or NULL if we're done. 475 * Returns dentry requiring refcount drop, or NULL if we're done.
476 */ 476 */
477 static inline struct dentry *dentry_kill(struct dentry *dentry) 477 static inline struct dentry *dentry_kill(struct dentry *dentry)
478 __releases(dentry->d_lock) 478 __releases(dentry->d_lock)
479 { 479 {
480 struct inode *inode; 480 struct inode *inode;
481 struct dentry *parent; 481 struct dentry *parent;
482 482
483 inode = dentry->d_inode; 483 inode = dentry->d_inode;
484 if (inode && !spin_trylock(&inode->i_lock)) { 484 if (inode && !spin_trylock(&inode->i_lock)) {
485 relock: 485 relock:
486 spin_unlock(&dentry->d_lock); 486 spin_unlock(&dentry->d_lock);
487 cpu_relax(); 487 cpu_relax();
488 return dentry; /* try again with same dentry */ 488 return dentry; /* try again with same dentry */
489 } 489 }
490 if (IS_ROOT(dentry)) 490 if (IS_ROOT(dentry))
491 parent = NULL; 491 parent = NULL;
492 else 492 else
493 parent = dentry->d_parent; 493 parent = dentry->d_parent;
494 if (parent && !spin_trylock(&parent->d_lock)) { 494 if (parent && !spin_trylock(&parent->d_lock)) {
495 if (inode) 495 if (inode)
496 spin_unlock(&inode->i_lock); 496 spin_unlock(&inode->i_lock);
497 goto relock; 497 goto relock;
498 } 498 }
499 499
500 /* 500 /*
501 * The dentry is now unrecoverably dead to the world. 501 * The dentry is now unrecoverably dead to the world.
502 */ 502 */
503 lockref_mark_dead(&dentry->d_lockref); 503 lockref_mark_dead(&dentry->d_lockref);
504 504
505 /* 505 /*
506 * inform the fs via d_prune that this dentry is about to be 506 * inform the fs via d_prune that this dentry is about to be
507 * unhashed and destroyed. 507 * unhashed and destroyed.
508 */ 508 */
509 if ((dentry->d_flags & DCACHE_OP_PRUNE) && !d_unhashed(dentry)) 509 if ((dentry->d_flags & DCACHE_OP_PRUNE) && !d_unhashed(dentry))
510 dentry->d_op->d_prune(dentry); 510 dentry->d_op->d_prune(dentry);
511 511
512 dentry_lru_del(dentry); 512 dentry_lru_del(dentry);
513 /* if it was on the hash then remove it */ 513 /* if it was on the hash then remove it */
514 __d_drop(dentry); 514 __d_drop(dentry);
515 return d_kill(dentry, parent); 515 return d_kill(dentry, parent);
516 } 516 }
517 517
518 /* 518 /*
519 * This is dput 519 * This is dput
520 * 520 *
521 * This is complicated by the fact that we do not want to put 521 * This is complicated by the fact that we do not want to put
522 * dentries that are no longer on any hash chain on the unused 522 * dentries that are no longer on any hash chain on the unused
523 * list: we'd much rather just get rid of them immediately. 523 * list: we'd much rather just get rid of them immediately.
524 * 524 *
525 * However, that implies that we have to traverse the dentry 525 * However, that implies that we have to traverse the dentry
526 * tree upwards to the parents which might _also_ now be 526 * tree upwards to the parents which might _also_ now be
527 * scheduled for deletion (it may have been only waiting for 527 * scheduled for deletion (it may have been only waiting for
528 * its last child to go away). 528 * its last child to go away).
529 * 529 *
530 * This tail recursion is done by hand as we don't want to depend 530 * This tail recursion is done by hand as we don't want to depend
531 * on the compiler to always get this right (gcc generally doesn't). 531 * on the compiler to always get this right (gcc generally doesn't).
532 * Real recursion would eat up our stack space. 532 * Real recursion would eat up our stack space.
533 */ 533 */
534 534
535 /* 535 /*
536 * dput - release a dentry 536 * dput - release a dentry
537 * @dentry: dentry to release 537 * @dentry: dentry to release
538 * 538 *
539 * Release a dentry. This will drop the usage count and if appropriate 539 * Release a dentry. This will drop the usage count and if appropriate
540 * call the dentry unlink method as well as removing it from the queues and 540 * call the dentry unlink method as well as removing it from the queues and
541 * releasing its resources. If the parent dentries were scheduled for release 541 * releasing its resources. If the parent dentries were scheduled for release
542 * they too may now get deleted. 542 * they too may now get deleted.
543 */ 543 */
544 void dput(struct dentry *dentry) 544 void dput(struct dentry *dentry)
545 { 545 {
546 if (unlikely(!dentry)) 546 if (unlikely(!dentry))
547 return; 547 return;
548 548
549 repeat: 549 repeat:
550 if (lockref_put_or_lock(&dentry->d_lockref)) 550 if (lockref_put_or_lock(&dentry->d_lockref))
551 return; 551 return;
552 552
553 /* Unreachable? Get rid of it */ 553 /* Unreachable? Get rid of it */
554 if (unlikely(d_unhashed(dentry))) 554 if (unlikely(d_unhashed(dentry)))
555 goto kill_it; 555 goto kill_it;
556 556
557 if (unlikely(dentry->d_flags & DCACHE_OP_DELETE)) { 557 if (unlikely(dentry->d_flags & DCACHE_OP_DELETE)) {
558 if (dentry->d_op->d_delete(dentry)) 558 if (dentry->d_op->d_delete(dentry))
559 goto kill_it; 559 goto kill_it;
560 } 560 }
561 561
562 dentry->d_flags |= DCACHE_REFERENCED; 562 dentry->d_flags |= DCACHE_REFERENCED;
563 dentry_lru_add(dentry); 563 dentry_lru_add(dentry);
564 564
565 dentry->d_lockref.count--; 565 dentry->d_lockref.count--;
566 spin_unlock(&dentry->d_lock); 566 spin_unlock(&dentry->d_lock);
567 return; 567 return;
568 568
569 kill_it: 569 kill_it:
570 dentry = dentry_kill(dentry); 570 dentry = dentry_kill(dentry);
571 if (dentry) 571 if (dentry)
572 goto repeat; 572 goto repeat;
573 } 573 }
574 EXPORT_SYMBOL(dput); 574 EXPORT_SYMBOL(dput);
575 575
576 /** 576 /**
577 * d_invalidate - invalidate a dentry 577 * d_invalidate - invalidate a dentry
578 * @dentry: dentry to invalidate 578 * @dentry: dentry to invalidate
579 * 579 *
580 * Try to invalidate the dentry if it turns out to be 580 * Try to invalidate the dentry if it turns out to be
581 * possible. If there are other dentries that can be 581 * possible. If there are other dentries that can be
582 * reached through this one we can't delete it and we 582 * reached through this one we can't delete it and we
583 * return -EBUSY. On success we return 0. 583 * return -EBUSY. On success we return 0.
584 * 584 *
585 * no dcache lock. 585 * no dcache lock.
586 */ 586 */
587 587
588 int d_invalidate(struct dentry * dentry) 588 int d_invalidate(struct dentry * dentry)
589 { 589 {
590 /* 590 /*
591 * If it's already been dropped, return OK. 591 * If it's already been dropped, return OK.
592 */ 592 */
593 spin_lock(&dentry->d_lock); 593 spin_lock(&dentry->d_lock);
594 if (d_unhashed(dentry)) { 594 if (d_unhashed(dentry)) {
595 spin_unlock(&dentry->d_lock); 595 spin_unlock(&dentry->d_lock);
596 return 0; 596 return 0;
597 } 597 }
598 /* 598 /*
599 * Check whether to do a partial shrink_dcache 599 * Check whether to do a partial shrink_dcache
600 * to get rid of unused child entries. 600 * to get rid of unused child entries.
601 */ 601 */
602 if (!list_empty(&dentry->d_subdirs)) { 602 if (!list_empty(&dentry->d_subdirs)) {
603 spin_unlock(&dentry->d_lock); 603 spin_unlock(&dentry->d_lock);
604 shrink_dcache_parent(dentry); 604 shrink_dcache_parent(dentry);
605 spin_lock(&dentry->d_lock); 605 spin_lock(&dentry->d_lock);
606 } 606 }
607 607
608 /* 608 /*
609 * Somebody else still using it? 609 * Somebody else still using it?
610 * 610 *
611 * If it's a directory, we can't drop it 611 * If it's a directory, we can't drop it
612 * for fear of somebody re-populating it 612 * for fear of somebody re-populating it
613 * with children (even though dropping it 613 * with children (even though dropping it
614 * would make it unreachable from the root, 614 * would make it unreachable from the root,
615 * we might still populate it if it was a 615 * we might still populate it if it was a
616 * working directory or similar). 616 * working directory or similar).
617 * We also need to leave mountpoints alone, 617 * We also need to leave mountpoints alone,
618 * directory or not. 618 * directory or not.
619 */ 619 */
620 if (dentry->d_lockref.count > 1 && dentry->d_inode) { 620 if (dentry->d_lockref.count > 1 && dentry->d_inode) {
621 if (S_ISDIR(dentry->d_inode->i_mode) || d_mountpoint(dentry)) { 621 if (S_ISDIR(dentry->d_inode->i_mode) || d_mountpoint(dentry)) {
622 spin_unlock(&dentry->d_lock); 622 spin_unlock(&dentry->d_lock);
623 return -EBUSY; 623 return -EBUSY;
624 } 624 }
625 } 625 }
626 626
627 __d_drop(dentry); 627 __d_drop(dentry);
628 spin_unlock(&dentry->d_lock); 628 spin_unlock(&dentry->d_lock);
629 return 0; 629 return 0;
630 } 630 }
631 EXPORT_SYMBOL(d_invalidate); 631 EXPORT_SYMBOL(d_invalidate);
632 632
633 /* This must be called with d_lock held */ 633 /* This must be called with d_lock held */
634 static inline void __dget_dlock(struct dentry *dentry) 634 static inline void __dget_dlock(struct dentry *dentry)
635 { 635 {
636 dentry->d_lockref.count++; 636 dentry->d_lockref.count++;
637 } 637 }
638 638
639 static inline void __dget(struct dentry *dentry) 639 static inline void __dget(struct dentry *dentry)
640 { 640 {
641 lockref_get(&dentry->d_lockref); 641 lockref_get(&dentry->d_lockref);
642 } 642 }
643 643
644 struct dentry *dget_parent(struct dentry *dentry) 644 struct dentry *dget_parent(struct dentry *dentry)
645 { 645 {
646 int gotref; 646 int gotref;
647 struct dentry *ret; 647 struct dentry *ret;
648 648
649 /* 649 /*
650 * Do optimistic parent lookup without any 650 * Do optimistic parent lookup without any
651 * locking. 651 * locking.
652 */ 652 */
653 rcu_read_lock(); 653 rcu_read_lock();
654 ret = ACCESS_ONCE(dentry->d_parent); 654 ret = ACCESS_ONCE(dentry->d_parent);
655 gotref = lockref_get_not_zero(&ret->d_lockref); 655 gotref = lockref_get_not_zero(&ret->d_lockref);
656 rcu_read_unlock(); 656 rcu_read_unlock();
657 if (likely(gotref)) { 657 if (likely(gotref)) {
658 if (likely(ret == ACCESS_ONCE(dentry->d_parent))) 658 if (likely(ret == ACCESS_ONCE(dentry->d_parent)))
659 return ret; 659 return ret;
660 dput(ret); 660 dput(ret);
661 } 661 }
662 662
663 repeat: 663 repeat:
664 /* 664 /*
665 * Don't need rcu_dereference because we re-check it was correct under 665 * Don't need rcu_dereference because we re-check it was correct under
666 * the lock. 666 * the lock.
667 */ 667 */
668 rcu_read_lock(); 668 rcu_read_lock();
669 ret = dentry->d_parent; 669 ret = dentry->d_parent;
670 spin_lock(&ret->d_lock); 670 spin_lock(&ret->d_lock);
671 if (unlikely(ret != dentry->d_parent)) { 671 if (unlikely(ret != dentry->d_parent)) {
672 spin_unlock(&ret->d_lock); 672 spin_unlock(&ret->d_lock);
673 rcu_read_unlock(); 673 rcu_read_unlock();
674 goto repeat; 674 goto repeat;
675 } 675 }
676 rcu_read_unlock(); 676 rcu_read_unlock();
677 BUG_ON(!ret->d_lockref.count); 677 BUG_ON(!ret->d_lockref.count);
678 ret->d_lockref.count++; 678 ret->d_lockref.count++;
679 spin_unlock(&ret->d_lock); 679 spin_unlock(&ret->d_lock);
680 return ret; 680 return ret;
681 } 681 }
682 EXPORT_SYMBOL(dget_parent); 682 EXPORT_SYMBOL(dget_parent);
683 683
684 /** 684 /**
685 * d_find_alias - grab a hashed alias of inode 685 * d_find_alias - grab a hashed alias of inode
686 * @inode: inode in question 686 * @inode: inode in question
687 * @want_discon: flag, used by d_splice_alias, to request 687 * @want_discon: flag, used by d_splice_alias, to request
688 * that only a DISCONNECTED alias be returned. 688 * that only a DISCONNECTED alias be returned.
689 * 689 *
690 * If inode has a hashed alias, or is a directory and has any alias, 690 * If inode has a hashed alias, or is a directory and has any alias,
691 * acquire the reference to alias and return it. Otherwise return NULL. 691 * acquire the reference to alias and return it. Otherwise return NULL.
692 * Notice that if inode is a directory there can be only one alias and 692 * Notice that if inode is a directory there can be only one alias and
693 * it can be unhashed only if it has no children, or if it is the root 693 * it can be unhashed only if it has no children, or if it is the root
694 * of a filesystem. 694 * of a filesystem.
695 * 695 *
696 * If the inode has an IS_ROOT, DCACHE_DISCONNECTED alias, then prefer 696 * If the inode has an IS_ROOT, DCACHE_DISCONNECTED alias, then prefer
697 * any other hashed alias over that one unless @want_discon is set, 697 * any other hashed alias over that one unless @want_discon is set,
698 * in which case only return an IS_ROOT, DCACHE_DISCONNECTED alias. 698 * in which case only return an IS_ROOT, DCACHE_DISCONNECTED alias.
699 */ 699 */
700 static struct dentry *__d_find_alias(struct inode *inode, int want_discon) 700 static struct dentry *__d_find_alias(struct inode *inode, int want_discon)
701 { 701 {
702 struct dentry *alias, *discon_alias; 702 struct dentry *alias, *discon_alias;
703 703
704 again: 704 again:
705 discon_alias = NULL; 705 discon_alias = NULL;
706 hlist_for_each_entry(alias, &inode->i_dentry, d_alias) { 706 hlist_for_each_entry(alias, &inode->i_dentry, d_alias) {
707 spin_lock(&alias->d_lock); 707 spin_lock(&alias->d_lock);
708 if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) { 708 if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) {
709 if (IS_ROOT(alias) && 709 if (IS_ROOT(alias) &&
710 (alias->d_flags & DCACHE_DISCONNECTED)) { 710 (alias->d_flags & DCACHE_DISCONNECTED)) {
711 discon_alias = alias; 711 discon_alias = alias;
712 } else if (!want_discon) { 712 } else if (!want_discon) {
713 __dget_dlock(alias); 713 __dget_dlock(alias);
714 spin_unlock(&alias->d_lock); 714 spin_unlock(&alias->d_lock);
715 return alias; 715 return alias;
716 } 716 }
717 } 717 }
718 spin_unlock(&alias->d_lock); 718 spin_unlock(&alias->d_lock);
719 } 719 }
720 if (discon_alias) { 720 if (discon_alias) {
721 alias = discon_alias; 721 alias = discon_alias;
722 spin_lock(&alias->d_lock); 722 spin_lock(&alias->d_lock);
723 if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) { 723 if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) {
724 if (IS_ROOT(alias) && 724 if (IS_ROOT(alias) &&
725 (alias->d_flags & DCACHE_DISCONNECTED)) { 725 (alias->d_flags & DCACHE_DISCONNECTED)) {
726 __dget_dlock(alias); 726 __dget_dlock(alias);
727 spin_unlock(&alias->d_lock); 727 spin_unlock(&alias->d_lock);
728 return alias; 728 return alias;
729 } 729 }
730 } 730 }
731 spin_unlock(&alias->d_lock); 731 spin_unlock(&alias->d_lock);
732 goto again; 732 goto again;
733 } 733 }
734 return NULL; 734 return NULL;
735 } 735 }
736 736
737 struct dentry *d_find_alias(struct inode *inode) 737 struct dentry *d_find_alias(struct inode *inode)
738 { 738 {
739 struct dentry *de = NULL; 739 struct dentry *de = NULL;
740 740
741 if (!hlist_empty(&inode->i_dentry)) { 741 if (!hlist_empty(&inode->i_dentry)) {
742 spin_lock(&inode->i_lock); 742 spin_lock(&inode->i_lock);
743 de = __d_find_alias(inode, 0); 743 de = __d_find_alias(inode, 0);
744 spin_unlock(&inode->i_lock); 744 spin_unlock(&inode->i_lock);
745 } 745 }
746 return de; 746 return de;
747 } 747 }
748 EXPORT_SYMBOL(d_find_alias); 748 EXPORT_SYMBOL(d_find_alias);
749 749
750 /* 750 /*
751 * Try to kill dentries associated with this inode. 751 * Try to kill dentries associated with this inode.
752 * WARNING: you must own a reference to inode. 752 * WARNING: you must own a reference to inode.
753 */ 753 */
754 void d_prune_aliases(struct inode *inode) 754 void d_prune_aliases(struct inode *inode)
755 { 755 {
756 struct dentry *dentry; 756 struct dentry *dentry;
757 restart: 757 restart:
758 spin_lock(&inode->i_lock); 758 spin_lock(&inode->i_lock);
759 hlist_for_each_entry(dentry, &inode->i_dentry, d_alias) { 759 hlist_for_each_entry(dentry, &inode->i_dentry, d_alias) {
760 spin_lock(&dentry->d_lock); 760 spin_lock(&dentry->d_lock);
761 if (!dentry->d_lockref.count) { 761 if (!dentry->d_lockref.count) {
762 /* 762 /*
763 * inform the fs via d_prune that this dentry 763 * inform the fs via d_prune that this dentry
764 * is about to be unhashed and destroyed. 764 * is about to be unhashed and destroyed.
765 */ 765 */
766 if ((dentry->d_flags & DCACHE_OP_PRUNE) && 766 if ((dentry->d_flags & DCACHE_OP_PRUNE) &&
767 !d_unhashed(dentry)) 767 !d_unhashed(dentry))
768 dentry->d_op->d_prune(dentry); 768 dentry->d_op->d_prune(dentry);
769 769
770 __dget_dlock(dentry); 770 __dget_dlock(dentry);
771 __d_drop(dentry); 771 __d_drop(dentry);
772 spin_unlock(&dentry->d_lock); 772 spin_unlock(&dentry->d_lock);
773 spin_unlock(&inode->i_lock); 773 spin_unlock(&inode->i_lock);
774 dput(dentry); 774 dput(dentry);
775 goto restart; 775 goto restart;
776 } 776 }
777 spin_unlock(&dentry->d_lock); 777 spin_unlock(&dentry->d_lock);
778 } 778 }
779 spin_unlock(&inode->i_lock); 779 spin_unlock(&inode->i_lock);
780 } 780 }
781 EXPORT_SYMBOL(d_prune_aliases); 781 EXPORT_SYMBOL(d_prune_aliases);
782 782
783 /* 783 /*
784 * Try to throw away a dentry - free the inode, dput the parent. 784 * Try to throw away a dentry - free the inode, dput the parent.
785 * Requires dentry->d_lock is held, and dentry->d_count == 0. 785 * Requires dentry->d_lock is held, and dentry->d_count == 0.
786 * Releases dentry->d_lock. 786 * Releases dentry->d_lock.
787 * 787 *
788 * This may fail if locks cannot be acquired no problem, just try again. 788 * This may fail if locks cannot be acquired no problem, just try again.
789 */ 789 */
790 static void try_prune_one_dentry(struct dentry *dentry) 790 static void try_prune_one_dentry(struct dentry *dentry)
791 __releases(dentry->d_lock) 791 __releases(dentry->d_lock)
792 { 792 {
793 struct dentry *parent; 793 struct dentry *parent;
794 794
795 parent = dentry_kill(dentry); 795 parent = dentry_kill(dentry);
796 /* 796 /*
797 * If dentry_kill returns NULL, we have nothing more to do. 797 * If dentry_kill returns NULL, we have nothing more to do.
798 * if it returns the same dentry, trylocks failed. In either 798 * if it returns the same dentry, trylocks failed. In either
799 * case, just loop again. 799 * case, just loop again.
800 * 800 *
801 * Otherwise, we need to prune ancestors too. This is necessary 801 * Otherwise, we need to prune ancestors too. This is necessary
802 * to prevent quadratic behavior of shrink_dcache_parent(), but 802 * to prevent quadratic behavior of shrink_dcache_parent(), but
803 * is also expected to be beneficial in reducing dentry cache 803 * is also expected to be beneficial in reducing dentry cache
804 * fragmentation. 804 * fragmentation.
805 */ 805 */
806 if (!parent) 806 if (!parent)
807 return; 807 return;
808 if (parent == dentry) 808 if (parent == dentry)
809 return; 809 return;
810 810
811 /* Prune ancestors. */ 811 /* Prune ancestors. */
812 dentry = parent; 812 dentry = parent;
813 while (dentry) { 813 while (dentry) {
814 if (lockref_put_or_lock(&dentry->d_lockref)) 814 if (lockref_put_or_lock(&dentry->d_lockref))
815 return; 815 return;
816 dentry = dentry_kill(dentry); 816 dentry = dentry_kill(dentry);
817 } 817 }
818 } 818 }
819 819
820 static void shrink_dentry_list(struct list_head *list) 820 static void shrink_dentry_list(struct list_head *list)
821 { 821 {
822 struct dentry *dentry; 822 struct dentry *dentry;
823 823
824 rcu_read_lock(); 824 rcu_read_lock();
825 for (;;) { 825 for (;;) {
826 dentry = list_entry_rcu(list->prev, struct dentry, d_lru); 826 dentry = list_entry_rcu(list->prev, struct dentry, d_lru);
827 if (&dentry->d_lru == list) 827 if (&dentry->d_lru == list)
828 break; /* empty */ 828 break; /* empty */
829 spin_lock(&dentry->d_lock); 829 spin_lock(&dentry->d_lock);
830 if (dentry != list_entry(list->prev, struct dentry, d_lru)) { 830 if (dentry != list_entry(list->prev, struct dentry, d_lru)) {
831 spin_unlock(&dentry->d_lock); 831 spin_unlock(&dentry->d_lock);
832 continue; 832 continue;
833 } 833 }
834 834
835 /* 835 /*
836 * We found an inuse dentry which was not removed from 836 * We found an inuse dentry which was not removed from
837 * the LRU because of laziness during lookup. Do not free 837 * the LRU because of laziness during lookup. Do not free
838 * it - just keep it off the LRU list. 838 * it - just keep it off the LRU list.
839 */ 839 */
840 if (dentry->d_lockref.count) { 840 if (dentry->d_lockref.count) {
841 dentry_lru_del(dentry); 841 dentry_lru_del(dentry);
842 spin_unlock(&dentry->d_lock); 842 spin_unlock(&dentry->d_lock);
843 continue; 843 continue;
844 } 844 }
845 845
846 rcu_read_unlock(); 846 rcu_read_unlock();
847 847
848 try_prune_one_dentry(dentry); 848 try_prune_one_dentry(dentry);
849 849
850 rcu_read_lock(); 850 rcu_read_lock();
851 } 851 }
852 rcu_read_unlock(); 852 rcu_read_unlock();
853 } 853 }
854 854
855 /** 855 /**
856 * prune_dcache_sb - shrink the dcache 856 * prune_dcache_sb - shrink the dcache
857 * @sb: superblock 857 * @sb: superblock
858 * @count: number of entries to try to free 858 * @count: number of entries to try to free
859 * 859 *
860 * Attempt to shrink the superblock dcache LRU by @count entries. This is 860 * Attempt to shrink the superblock dcache LRU by @count entries. This is
861 * done when we need more memory an called from the superblock shrinker 861 * done when we need more memory an called from the superblock shrinker
862 * function. 862 * function.
863 * 863 *
864 * This function may fail to free any resources if all the dentries are in 864 * This function may fail to free any resources if all the dentries are in
865 * use. 865 * use.
866 */ 866 */
867 void prune_dcache_sb(struct super_block *sb, int count) 867 void prune_dcache_sb(struct super_block *sb, int count)
868 { 868 {
869 struct dentry *dentry; 869 struct dentry *dentry;
870 LIST_HEAD(referenced); 870 LIST_HEAD(referenced);
871 LIST_HEAD(tmp); 871 LIST_HEAD(tmp);
872 872
873 relock: 873 relock:
874 spin_lock(&dcache_lru_lock); 874 spin_lock(&dcache_lru_lock);
875 while (!list_empty(&sb->s_dentry_lru)) { 875 while (!list_empty(&sb->s_dentry_lru)) {
876 dentry = list_entry(sb->s_dentry_lru.prev, 876 dentry = list_entry(sb->s_dentry_lru.prev,
877 struct dentry, d_lru); 877 struct dentry, d_lru);
878 BUG_ON(dentry->d_sb != sb); 878 BUG_ON(dentry->d_sb != sb);
879 879
880 if (!spin_trylock(&dentry->d_lock)) { 880 if (!spin_trylock(&dentry->d_lock)) {
881 spin_unlock(&dcache_lru_lock); 881 spin_unlock(&dcache_lru_lock);
882 cpu_relax(); 882 cpu_relax();
883 goto relock; 883 goto relock;
884 } 884 }
885 885
886 if (dentry->d_flags & DCACHE_REFERENCED) { 886 if (dentry->d_flags & DCACHE_REFERENCED) {
887 dentry->d_flags &= ~DCACHE_REFERENCED; 887 dentry->d_flags &= ~DCACHE_REFERENCED;
888 list_move(&dentry->d_lru, &referenced); 888 list_move(&dentry->d_lru, &referenced);
889 spin_unlock(&dentry->d_lock); 889 spin_unlock(&dentry->d_lock);
890 } else { 890 } else {
891 list_move_tail(&dentry->d_lru, &tmp); 891 list_move_tail(&dentry->d_lru, &tmp);
892 dentry->d_flags |= DCACHE_SHRINK_LIST; 892 dentry->d_flags |= DCACHE_SHRINK_LIST;
893 spin_unlock(&dentry->d_lock); 893 spin_unlock(&dentry->d_lock);
894 if (!--count) 894 if (!--count)
895 break; 895 break;
896 } 896 }
897 cond_resched_lock(&dcache_lru_lock); 897 cond_resched_lock(&dcache_lru_lock);
898 } 898 }
899 if (!list_empty(&referenced)) 899 if (!list_empty(&referenced))
900 list_splice(&referenced, &sb->s_dentry_lru); 900 list_splice(&referenced, &sb->s_dentry_lru);
901 spin_unlock(&dcache_lru_lock); 901 spin_unlock(&dcache_lru_lock);
902 902
903 shrink_dentry_list(&tmp); 903 shrink_dentry_list(&tmp);
904 } 904 }
905 905
906 /** 906 /**
907 * shrink_dcache_sb - shrink dcache for a superblock 907 * shrink_dcache_sb - shrink dcache for a superblock
908 * @sb: superblock 908 * @sb: superblock
909 * 909 *
910 * Shrink the dcache for the specified super block. This is used to free 910 * Shrink the dcache for the specified super block. This is used to free
911 * the dcache before unmounting a file system. 911 * the dcache before unmounting a file system.
912 */ 912 */
913 void shrink_dcache_sb(struct super_block *sb) 913 void shrink_dcache_sb(struct super_block *sb)
914 { 914 {
915 LIST_HEAD(tmp); 915 LIST_HEAD(tmp);
916 916
917 spin_lock(&dcache_lru_lock); 917 spin_lock(&dcache_lru_lock);
918 while (!list_empty(&sb->s_dentry_lru)) { 918 while (!list_empty(&sb->s_dentry_lru)) {
919 list_splice_init(&sb->s_dentry_lru, &tmp); 919 list_splice_init(&sb->s_dentry_lru, &tmp);
920 spin_unlock(&dcache_lru_lock); 920 spin_unlock(&dcache_lru_lock);
921 shrink_dentry_list(&tmp); 921 shrink_dentry_list(&tmp);
922 spin_lock(&dcache_lru_lock); 922 spin_lock(&dcache_lru_lock);
923 } 923 }
924 spin_unlock(&dcache_lru_lock); 924 spin_unlock(&dcache_lru_lock);
925 } 925 }
926 EXPORT_SYMBOL(shrink_dcache_sb); 926 EXPORT_SYMBOL(shrink_dcache_sb);
927 927
928 /* 928 /*
929 * destroy a single subtree of dentries for unmount 929 * destroy a single subtree of dentries for unmount
930 * - see the comments on shrink_dcache_for_umount() for a description of the 930 * - see the comments on shrink_dcache_for_umount() for a description of the
931 * locking 931 * locking
932 */ 932 */
933 static void shrink_dcache_for_umount_subtree(struct dentry *dentry) 933 static void shrink_dcache_for_umount_subtree(struct dentry *dentry)
934 { 934 {
935 struct dentry *parent; 935 struct dentry *parent;
936 936
937 BUG_ON(!IS_ROOT(dentry)); 937 BUG_ON(!IS_ROOT(dentry));
938 938
939 for (;;) { 939 for (;;) {
940 /* descend to the first leaf in the current subtree */ 940 /* descend to the first leaf in the current subtree */
941 while (!list_empty(&dentry->d_subdirs)) 941 while (!list_empty(&dentry->d_subdirs))
942 dentry = list_entry(dentry->d_subdirs.next, 942 dentry = list_entry(dentry->d_subdirs.next,
943 struct dentry, d_u.d_child); 943 struct dentry, d_u.d_child);
944 944
945 /* consume the dentries from this leaf up through its parents 945 /* consume the dentries from this leaf up through its parents
946 * until we find one with children or run out altogether */ 946 * until we find one with children or run out altogether */
947 do { 947 do {
948 struct inode *inode; 948 struct inode *inode;
949 949
950 /* 950 /*
951 * inform the fs that this dentry is about to be 951 * inform the fs that this dentry is about to be
952 * unhashed and destroyed. 952 * unhashed and destroyed.
953 */ 953 */
954 if ((dentry->d_flags & DCACHE_OP_PRUNE) && 954 if ((dentry->d_flags & DCACHE_OP_PRUNE) &&
955 !d_unhashed(dentry)) 955 !d_unhashed(dentry))
956 dentry->d_op->d_prune(dentry); 956 dentry->d_op->d_prune(dentry);
957 957
958 dentry_lru_del(dentry); 958 dentry_lru_del(dentry);
959 __d_shrink(dentry); 959 __d_shrink(dentry);
960 960
961 if (dentry->d_lockref.count != 0) { 961 if (dentry->d_lockref.count != 0) {
962 printk(KERN_ERR 962 printk(KERN_ERR
963 "BUG: Dentry %p{i=%lx,n=%s}" 963 "BUG: Dentry %p{i=%lx,n=%s}"
964 " still in use (%d)" 964 " still in use (%d)"
965 " [unmount of %s %s]\n", 965 " [unmount of %s %s]\n",
966 dentry, 966 dentry,
967 dentry->d_inode ? 967 dentry->d_inode ?
968 dentry->d_inode->i_ino : 0UL, 968 dentry->d_inode->i_ino : 0UL,
969 dentry->d_name.name, 969 dentry->d_name.name,
970 dentry->d_lockref.count, 970 dentry->d_lockref.count,
971 dentry->d_sb->s_type->name, 971 dentry->d_sb->s_type->name,
972 dentry->d_sb->s_id); 972 dentry->d_sb->s_id);
973 BUG(); 973 BUG();
974 } 974 }
975 975
976 if (IS_ROOT(dentry)) { 976 if (IS_ROOT(dentry)) {
977 parent = NULL; 977 parent = NULL;
978 list_del(&dentry->d_u.d_child); 978 list_del(&dentry->d_u.d_child);
979 } else { 979 } else {
980 parent = dentry->d_parent; 980 parent = dentry->d_parent;
981 parent->d_lockref.count--; 981 parent->d_lockref.count--;
982 list_del(&dentry->d_u.d_child); 982 list_del(&dentry->d_u.d_child);
983 } 983 }
984 984
985 inode = dentry->d_inode; 985 inode = dentry->d_inode;
986 if (inode) { 986 if (inode) {
987 dentry->d_inode = NULL; 987 dentry->d_inode = NULL;
988 hlist_del_init(&dentry->d_alias); 988 hlist_del_init(&dentry->d_alias);
989 if (dentry->d_op && dentry->d_op->d_iput) 989 if (dentry->d_op && dentry->d_op->d_iput)
990 dentry->d_op->d_iput(dentry, inode); 990 dentry->d_op->d_iput(dentry, inode);
991 else 991 else
992 iput(inode); 992 iput(inode);
993 } 993 }
994 994
995 d_free(dentry); 995 d_free(dentry);
996 996
997 /* finished when we fall off the top of the tree, 997 /* finished when we fall off the top of the tree,
998 * otherwise we ascend to the parent and move to the 998 * otherwise we ascend to the parent and move to the
999 * next sibling if there is one */ 999 * next sibling if there is one */
1000 if (!parent) 1000 if (!parent)
1001 return; 1001 return;
1002 dentry = parent; 1002 dentry = parent;
1003 } while (list_empty(&dentry->d_subdirs)); 1003 } while (list_empty(&dentry->d_subdirs));
1004 1004
1005 dentry = list_entry(dentry->d_subdirs.next, 1005 dentry = list_entry(dentry->d_subdirs.next,
1006 struct dentry, d_u.d_child); 1006 struct dentry, d_u.d_child);
1007 } 1007 }
1008 } 1008 }
1009 1009
1010 /* 1010 /*
1011 * destroy the dentries attached to a superblock on unmounting 1011 * destroy the dentries attached to a superblock on unmounting
1012 * - we don't need to use dentry->d_lock because: 1012 * - we don't need to use dentry->d_lock because:
1013 * - the superblock is detached from all mountings and open files, so the 1013 * - the superblock is detached from all mountings and open files, so the
1014 * dentry trees will not be rearranged by the VFS 1014 * dentry trees will not be rearranged by the VFS
1015 * - s_umount is write-locked, so the memory pressure shrinker will ignore 1015 * - s_umount is write-locked, so the memory pressure shrinker will ignore
1016 * any dentries belonging to this superblock that it comes across 1016 * any dentries belonging to this superblock that it comes across
1017 * - the filesystem itself is no longer permitted to rearrange the dentries 1017 * - the filesystem itself is no longer permitted to rearrange the dentries
1018 * in this superblock 1018 * in this superblock
1019 */ 1019 */
1020 void shrink_dcache_for_umount(struct super_block *sb) 1020 void shrink_dcache_for_umount(struct super_block *sb)
1021 { 1021 {
1022 struct dentry *dentry; 1022 struct dentry *dentry;
1023 1023
1024 if (down_read_trylock(&sb->s_umount)) 1024 if (down_read_trylock(&sb->s_umount))
1025 BUG(); 1025 BUG();
1026 1026
1027 dentry = sb->s_root; 1027 dentry = sb->s_root;
1028 sb->s_root = NULL; 1028 sb->s_root = NULL;
1029 dentry->d_lockref.count--; 1029 dentry->d_lockref.count--;
1030 shrink_dcache_for_umount_subtree(dentry); 1030 shrink_dcache_for_umount_subtree(dentry);
1031 1031
1032 while (!hlist_bl_empty(&sb->s_anon)) { 1032 while (!hlist_bl_empty(&sb->s_anon)) {
1033 dentry = hlist_bl_entry(hlist_bl_first(&sb->s_anon), struct dentry, d_hash); 1033 dentry = hlist_bl_entry(hlist_bl_first(&sb->s_anon), struct dentry, d_hash);
1034 shrink_dcache_for_umount_subtree(dentry); 1034 shrink_dcache_for_umount_subtree(dentry);
1035 } 1035 }
1036 } 1036 }
1037 1037
1038 /* 1038 /*
1039 * This tries to ascend one level of parenthood, but 1039 * This tries to ascend one level of parenthood, but
1040 * we can race with renaming, so we need to re-check 1040 * we can race with renaming, so we need to re-check
1041 * the parenthood after dropping the lock and check 1041 * the parenthood after dropping the lock and check
1042 * that the sequence number still matches. 1042 * that the sequence number still matches.
1043 */ 1043 */
1044 static struct dentry *try_to_ascend(struct dentry *old, unsigned seq) 1044 static struct dentry *try_to_ascend(struct dentry *old, unsigned seq)
1045 { 1045 {
1046 struct dentry *new = old->d_parent; 1046 struct dentry *new = old->d_parent;
1047 1047
1048 rcu_read_lock(); 1048 rcu_read_lock();
1049 spin_unlock(&old->d_lock); 1049 spin_unlock(&old->d_lock);
1050 spin_lock(&new->d_lock); 1050 spin_lock(&new->d_lock);
1051 1051
1052 /* 1052 /*
1053 * might go back up the wrong parent if we have had a rename 1053 * might go back up the wrong parent if we have had a rename
1054 * or deletion 1054 * or deletion
1055 */ 1055 */
1056 if (new != old->d_parent || 1056 if (new != old->d_parent ||
1057 (old->d_flags & DCACHE_DENTRY_KILLED) || 1057 (old->d_flags & DCACHE_DENTRY_KILLED) ||
1058 need_seqretry(&rename_lock, seq)) { 1058 need_seqretry(&rename_lock, seq)) {
1059 spin_unlock(&new->d_lock); 1059 spin_unlock(&new->d_lock);
1060 new = NULL; 1060 new = NULL;
1061 } 1061 }
1062 rcu_read_unlock(); 1062 rcu_read_unlock();
1063 return new; 1063 return new;
1064 } 1064 }
1065 1065
1066 /** 1066 /**
1067 * enum d_walk_ret - action to talke during tree walk 1067 * enum d_walk_ret - action to talke during tree walk
1068 * @D_WALK_CONTINUE: contrinue walk 1068 * @D_WALK_CONTINUE: contrinue walk
1069 * @D_WALK_QUIT: quit walk 1069 * @D_WALK_QUIT: quit walk
1070 * @D_WALK_NORETRY: quit when retry is needed 1070 * @D_WALK_NORETRY: quit when retry is needed
1071 * @D_WALK_SKIP: skip this dentry and its children 1071 * @D_WALK_SKIP: skip this dentry and its children
1072 */ 1072 */
1073 enum d_walk_ret { 1073 enum d_walk_ret {
1074 D_WALK_CONTINUE, 1074 D_WALK_CONTINUE,
1075 D_WALK_QUIT, 1075 D_WALK_QUIT,
1076 D_WALK_NORETRY, 1076 D_WALK_NORETRY,
1077 D_WALK_SKIP, 1077 D_WALK_SKIP,
1078 }; 1078 };
1079 1079
1080 /** 1080 /**
1081 * d_walk - walk the dentry tree 1081 * d_walk - walk the dentry tree
1082 * @parent: start of walk 1082 * @parent: start of walk
1083 * @data: data passed to @enter() and @finish() 1083 * @data: data passed to @enter() and @finish()
1084 * @enter: callback when first entering the dentry 1084 * @enter: callback when first entering the dentry
1085 * @finish: callback when successfully finished the walk 1085 * @finish: callback when successfully finished the walk
1086 * 1086 *
1087 * The @enter() and @finish() callbacks are called with d_lock held. 1087 * The @enter() and @finish() callbacks are called with d_lock held.
1088 */ 1088 */
1089 static void d_walk(struct dentry *parent, void *data, 1089 static void d_walk(struct dentry *parent, void *data,
1090 enum d_walk_ret (*enter)(void *, struct dentry *), 1090 enum d_walk_ret (*enter)(void *, struct dentry *),
1091 void (*finish)(void *)) 1091 void (*finish)(void *))
1092 { 1092 {
1093 struct dentry *this_parent; 1093 struct dentry *this_parent;
1094 struct list_head *next; 1094 struct list_head *next;
1095 unsigned seq = 0; 1095 unsigned seq = 0;
1096 enum d_walk_ret ret; 1096 enum d_walk_ret ret;
1097 bool retry = true; 1097 bool retry = true;
1098 1098
1099 again: 1099 again:
1100 read_seqbegin_or_lock(&rename_lock, &seq); 1100 read_seqbegin_or_lock(&rename_lock, &seq);
1101 this_parent = parent; 1101 this_parent = parent;
1102 spin_lock(&this_parent->d_lock); 1102 spin_lock(&this_parent->d_lock);
1103 1103
1104 ret = enter(data, this_parent); 1104 ret = enter(data, this_parent);
1105 switch (ret) { 1105 switch (ret) {
1106 case D_WALK_CONTINUE: 1106 case D_WALK_CONTINUE:
1107 break; 1107 break;
1108 case D_WALK_QUIT: 1108 case D_WALK_QUIT:
1109 case D_WALK_SKIP: 1109 case D_WALK_SKIP:
1110 goto out_unlock; 1110 goto out_unlock;
1111 case D_WALK_NORETRY: 1111 case D_WALK_NORETRY:
1112 retry = false; 1112 retry = false;
1113 break; 1113 break;
1114 } 1114 }
1115 repeat: 1115 repeat:
1116 next = this_parent->d_subdirs.next; 1116 next = this_parent->d_subdirs.next;
1117 resume: 1117 resume:
1118 while (next != &this_parent->d_subdirs) { 1118 while (next != &this_parent->d_subdirs) {
1119 struct list_head *tmp = next; 1119 struct list_head *tmp = next;
1120 struct dentry *dentry = list_entry(tmp, struct dentry, d_u.d_child); 1120 struct dentry *dentry = list_entry(tmp, struct dentry, d_u.d_child);
1121 next = tmp->next; 1121 next = tmp->next;
1122 1122
1123 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); 1123 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
1124 1124
1125 ret = enter(data, dentry); 1125 ret = enter(data, dentry);
1126 switch (ret) { 1126 switch (ret) {
1127 case D_WALK_CONTINUE: 1127 case D_WALK_CONTINUE:
1128 break; 1128 break;
1129 case D_WALK_QUIT: 1129 case D_WALK_QUIT:
1130 spin_unlock(&dentry->d_lock); 1130 spin_unlock(&dentry->d_lock);
1131 goto out_unlock; 1131 goto out_unlock;
1132 case D_WALK_NORETRY: 1132 case D_WALK_NORETRY:
1133 retry = false; 1133 retry = false;
1134 break; 1134 break;
1135 case D_WALK_SKIP: 1135 case D_WALK_SKIP:
1136 spin_unlock(&dentry->d_lock); 1136 spin_unlock(&dentry->d_lock);
1137 continue; 1137 continue;
1138 } 1138 }
1139 1139
1140 if (!list_empty(&dentry->d_subdirs)) { 1140 if (!list_empty(&dentry->d_subdirs)) {
1141 spin_unlock(&this_parent->d_lock); 1141 spin_unlock(&this_parent->d_lock);
1142 spin_release(&dentry->d_lock.dep_map, 1, _RET_IP_); 1142 spin_release(&dentry->d_lock.dep_map, 1, _RET_IP_);
1143 this_parent = dentry; 1143 this_parent = dentry;
1144 spin_acquire(&this_parent->d_lock.dep_map, 0, 1, _RET_IP_); 1144 spin_acquire(&this_parent->d_lock.dep_map, 0, 1, _RET_IP_);
1145 goto repeat; 1145 goto repeat;
1146 } 1146 }
1147 spin_unlock(&dentry->d_lock); 1147 spin_unlock(&dentry->d_lock);
1148 } 1148 }
1149 /* 1149 /*
1150 * All done at this level ... ascend and resume the search. 1150 * All done at this level ... ascend and resume the search.
1151 */ 1151 */
1152 if (this_parent != parent) { 1152 if (this_parent != parent) {
1153 struct dentry *child = this_parent; 1153 struct dentry *child = this_parent;
1154 this_parent = try_to_ascend(this_parent, seq); 1154 this_parent = try_to_ascend(this_parent, seq);
1155 if (!this_parent) 1155 if (!this_parent)
1156 goto rename_retry; 1156 goto rename_retry;
1157 next = child->d_u.d_child.next; 1157 next = child->d_u.d_child.next;
1158 goto resume; 1158 goto resume;
1159 } 1159 }
1160 if (need_seqretry(&rename_lock, seq)) { 1160 if (need_seqretry(&rename_lock, seq)) {
1161 spin_unlock(&this_parent->d_lock); 1161 spin_unlock(&this_parent->d_lock);
1162 goto rename_retry; 1162 goto rename_retry;
1163 } 1163 }
1164 if (finish) 1164 if (finish)
1165 finish(data); 1165 finish(data);
1166 1166
1167 out_unlock: 1167 out_unlock:
1168 spin_unlock(&this_parent->d_lock); 1168 spin_unlock(&this_parent->d_lock);
1169 done_seqretry(&rename_lock, seq); 1169 done_seqretry(&rename_lock, seq);
1170 return; 1170 return;
1171 1171
1172 rename_retry: 1172 rename_retry:
1173 if (!retry) 1173 if (!retry)
1174 return; 1174 return;
1175 seq = 1; 1175 seq = 1;
1176 goto again; 1176 goto again;
1177 } 1177 }
1178 1178
1179 /* 1179 /*
1180 * Search for at least 1 mount point in the dentry's subdirs. 1180 * Search for at least 1 mount point in the dentry's subdirs.
1181 * We descend to the next level whenever the d_subdirs 1181 * We descend to the next level whenever the d_subdirs
1182 * list is non-empty and continue searching. 1182 * list is non-empty and continue searching.
1183 */ 1183 */
1184 1184
1185 /** 1185 /**
1186 * have_submounts - check for mounts over a dentry 1186 * have_submounts - check for mounts over a dentry
1187 * @parent: dentry to check. 1187 * @parent: dentry to check.
1188 * 1188 *
1189 * Return true if the parent or its subdirectories contain 1189 * Return true if the parent or its subdirectories contain
1190 * a mount point 1190 * a mount point
1191 */ 1191 */
1192 1192
1193 static enum d_walk_ret check_mount(void *data, struct dentry *dentry) 1193 static enum d_walk_ret check_mount(void *data, struct dentry *dentry)
1194 { 1194 {
1195 int *ret = data; 1195 int *ret = data;
1196 if (d_mountpoint(dentry)) { 1196 if (d_mountpoint(dentry)) {
1197 *ret = 1; 1197 *ret = 1;
1198 return D_WALK_QUIT; 1198 return D_WALK_QUIT;
1199 } 1199 }
1200 return D_WALK_CONTINUE; 1200 return D_WALK_CONTINUE;
1201 } 1201 }
1202 1202
1203 int have_submounts(struct dentry *parent) 1203 int have_submounts(struct dentry *parent)
1204 { 1204 {
1205 int ret = 0; 1205 int ret = 0;
1206 1206
1207 d_walk(parent, &ret, check_mount, NULL); 1207 d_walk(parent, &ret, check_mount, NULL);
1208 1208
1209 return ret; 1209 return ret;
1210 } 1210 }
1211 EXPORT_SYMBOL(have_submounts); 1211 EXPORT_SYMBOL(have_submounts);
1212 1212
1213 /* 1213 /*
1214 * Called by mount code to set a mountpoint and check if the mountpoint is 1214 * Called by mount code to set a mountpoint and check if the mountpoint is
1215 * reachable (e.g. NFS can unhash a directory dentry and then the complete 1215 * reachable (e.g. NFS can unhash a directory dentry and then the complete
1216 * subtree can become unreachable). 1216 * subtree can become unreachable).
1217 * 1217 *
1218 * Only one of check_submounts_and_drop() and d_set_mounted() must succeed. For 1218 * Only one of check_submounts_and_drop() and d_set_mounted() must succeed. For
1219 * this reason take rename_lock and d_lock on dentry and ancestors. 1219 * this reason take rename_lock and d_lock on dentry and ancestors.
1220 */ 1220 */
1221 int d_set_mounted(struct dentry *dentry) 1221 int d_set_mounted(struct dentry *dentry)
1222 { 1222 {
1223 struct dentry *p; 1223 struct dentry *p;
1224 int ret = -ENOENT; 1224 int ret = -ENOENT;
1225 write_seqlock(&rename_lock); 1225 write_seqlock(&rename_lock);
1226 for (p = dentry->d_parent; !IS_ROOT(p); p = p->d_parent) { 1226 for (p = dentry->d_parent; !IS_ROOT(p); p = p->d_parent) {
1227 /* Need exclusion wrt. check_submounts_and_drop() */ 1227 /* Need exclusion wrt. check_submounts_and_drop() */
1228 spin_lock(&p->d_lock); 1228 spin_lock(&p->d_lock);
1229 if (unlikely(d_unhashed(p))) { 1229 if (unlikely(d_unhashed(p))) {
1230 spin_unlock(&p->d_lock); 1230 spin_unlock(&p->d_lock);
1231 goto out; 1231 goto out;
1232 } 1232 }
1233 spin_unlock(&p->d_lock); 1233 spin_unlock(&p->d_lock);
1234 } 1234 }
1235 spin_lock(&dentry->d_lock); 1235 spin_lock(&dentry->d_lock);
1236 if (!d_unlinked(dentry)) { 1236 if (!d_unlinked(dentry)) {
1237 dentry->d_flags |= DCACHE_MOUNTED; 1237 dentry->d_flags |= DCACHE_MOUNTED;
1238 ret = 0; 1238 ret = 0;
1239 } 1239 }
1240 spin_unlock(&dentry->d_lock); 1240 spin_unlock(&dentry->d_lock);
1241 out: 1241 out:
1242 write_sequnlock(&rename_lock); 1242 write_sequnlock(&rename_lock);
1243 return ret; 1243 return ret;
1244 } 1244 }
1245 1245
1246 /* 1246 /*
1247 * Search the dentry child list of the specified parent, 1247 * Search the dentry child list of the specified parent,
1248 * and move any unused dentries to the end of the unused 1248 * and move any unused dentries to the end of the unused
1249 * list for prune_dcache(). We descend to the next level 1249 * list for prune_dcache(). We descend to the next level
1250 * whenever the d_subdirs list is non-empty and continue 1250 * whenever the d_subdirs list is non-empty and continue
1251 * searching. 1251 * searching.
1252 * 1252 *
1253 * It returns zero iff there are no unused children, 1253 * It returns zero iff there are no unused children,
1254 * otherwise it returns the number of children moved to 1254 * otherwise it returns the number of children moved to
1255 * the end of the unused list. This may not be the total 1255 * the end of the unused list. This may not be the total
1256 * number of unused children, because select_parent can 1256 * number of unused children, because select_parent can
1257 * drop the lock and return early due to latency 1257 * drop the lock and return early due to latency
1258 * constraints. 1258 * constraints.
1259 */ 1259 */
1260 1260
1261 struct select_data { 1261 struct select_data {
1262 struct dentry *start; 1262 struct dentry *start;
1263 struct list_head dispose; 1263 struct list_head dispose;
1264 int found; 1264 int found;
1265 }; 1265 };
1266 1266
1267 static enum d_walk_ret select_collect(void *_data, struct dentry *dentry) 1267 static enum d_walk_ret select_collect(void *_data, struct dentry *dentry)
1268 { 1268 {
1269 struct select_data *data = _data; 1269 struct select_data *data = _data;
1270 enum d_walk_ret ret = D_WALK_CONTINUE; 1270 enum d_walk_ret ret = D_WALK_CONTINUE;
1271 1271
1272 if (data->start == dentry) 1272 if (data->start == dentry)
1273 goto out; 1273 goto out;
1274 1274
1275 /* 1275 /*
1276 * move only zero ref count dentries to the dispose list. 1276 * move only zero ref count dentries to the dispose list.
1277 * 1277 *
1278 * Those which are presently on the shrink list, being processed 1278 * Those which are presently on the shrink list, being processed
1279 * by shrink_dentry_list(), shouldn't be moved. Otherwise the 1279 * by shrink_dentry_list(), shouldn't be moved. Otherwise the
1280 * loop in shrink_dcache_parent() might not make any progress 1280 * loop in shrink_dcache_parent() might not make any progress
1281 * and loop forever. 1281 * and loop forever.
1282 */ 1282 */
1283 if (dentry->d_lockref.count) { 1283 if (dentry->d_lockref.count) {
1284 dentry_lru_del(dentry); 1284 dentry_lru_del(dentry);
1285 } else if (!(dentry->d_flags & DCACHE_SHRINK_LIST)) { 1285 } else if (!(dentry->d_flags & DCACHE_SHRINK_LIST)) {
1286 dentry_lru_move_list(dentry, &data->dispose); 1286 dentry_lru_move_list(dentry, &data->dispose);
1287 dentry->d_flags |= DCACHE_SHRINK_LIST; 1287 dentry->d_flags |= DCACHE_SHRINK_LIST;
1288 data->found++; 1288 data->found++;
1289 ret = D_WALK_NORETRY; 1289 ret = D_WALK_NORETRY;
1290 } 1290 }
1291 /* 1291 /*
1292 * We can return to the caller if we have found some (this 1292 * We can return to the caller if we have found some (this
1293 * ensures forward progress). We'll be coming back to find 1293 * ensures forward progress). We'll be coming back to find
1294 * the rest. 1294 * the rest.
1295 */ 1295 */
1296 if (data->found && need_resched()) 1296 if (data->found && need_resched())
1297 ret = D_WALK_QUIT; 1297 ret = D_WALK_QUIT;
1298 out: 1298 out:
1299 return ret; 1299 return ret;
1300 } 1300 }
1301 1301
1302 /** 1302 /**
1303 * shrink_dcache_parent - prune dcache 1303 * shrink_dcache_parent - prune dcache
1304 * @parent: parent of entries to prune 1304 * @parent: parent of entries to prune
1305 * 1305 *
1306 * Prune the dcache to remove unused children of the parent dentry. 1306 * Prune the dcache to remove unused children of the parent dentry.
1307 */ 1307 */
1308 void shrink_dcache_parent(struct dentry *parent) 1308 void shrink_dcache_parent(struct dentry *parent)
1309 { 1309 {
1310 for (;;) { 1310 for (;;) {
1311 struct select_data data; 1311 struct select_data data;
1312 1312
1313 INIT_LIST_HEAD(&data.dispose); 1313 INIT_LIST_HEAD(&data.dispose);
1314 data.start = parent; 1314 data.start = parent;
1315 data.found = 0; 1315 data.found = 0;
1316 1316
1317 d_walk(parent, &data, select_collect, NULL); 1317 d_walk(parent, &data, select_collect, NULL);
1318 if (!data.found) 1318 if (!data.found)
1319 break; 1319 break;
1320 1320
1321 shrink_dentry_list(&data.dispose); 1321 shrink_dentry_list(&data.dispose);
1322 cond_resched(); 1322 cond_resched();
1323 } 1323 }
1324 } 1324 }
1325 EXPORT_SYMBOL(shrink_dcache_parent); 1325 EXPORT_SYMBOL(shrink_dcache_parent);
1326 1326
1327 static enum d_walk_ret check_and_collect(void *_data, struct dentry *dentry) 1327 static enum d_walk_ret check_and_collect(void *_data, struct dentry *dentry)
1328 { 1328 {
1329 struct select_data *data = _data; 1329 struct select_data *data = _data;
1330 1330
1331 if (d_mountpoint(dentry)) { 1331 if (d_mountpoint(dentry)) {
1332 data->found = -EBUSY; 1332 data->found = -EBUSY;
1333 return D_WALK_QUIT; 1333 return D_WALK_QUIT;
1334 } 1334 }
1335 1335
1336 return select_collect(_data, dentry); 1336 return select_collect(_data, dentry);
1337 } 1337 }
1338 1338
1339 static void check_and_drop(void *_data) 1339 static void check_and_drop(void *_data)
1340 { 1340 {
1341 struct select_data *data = _data; 1341 struct select_data *data = _data;
1342 1342
1343 if (d_mountpoint(data->start)) 1343 if (d_mountpoint(data->start))
1344 data->found = -EBUSY; 1344 data->found = -EBUSY;
1345 if (!data->found) 1345 if (!data->found)
1346 __d_drop(data->start); 1346 __d_drop(data->start);
1347 } 1347 }
1348 1348
1349 /** 1349 /**
1350 * check_submounts_and_drop - prune dcache, check for submounts and drop 1350 * check_submounts_and_drop - prune dcache, check for submounts and drop
1351 * 1351 *
1352 * All done as a single atomic operation relative to has_unlinked_ancestor(). 1352 * All done as a single atomic operation relative to has_unlinked_ancestor().
1353 * Returns 0 if successfully unhashed @parent. If there were submounts then 1353 * Returns 0 if successfully unhashed @parent. If there were submounts then
1354 * return -EBUSY. 1354 * return -EBUSY.
1355 * 1355 *
1356 * @dentry: dentry to prune and drop 1356 * @dentry: dentry to prune and drop
1357 */ 1357 */
1358 int check_submounts_and_drop(struct dentry *dentry) 1358 int check_submounts_and_drop(struct dentry *dentry)
1359 { 1359 {
1360 int ret = 0; 1360 int ret = 0;
1361 1361
1362 /* Negative dentries can be dropped without further checks */ 1362 /* Negative dentries can be dropped without further checks */
1363 if (!dentry->d_inode) { 1363 if (!dentry->d_inode) {
1364 d_drop(dentry); 1364 d_drop(dentry);
1365 goto out; 1365 goto out;
1366 } 1366 }
1367 1367
1368 for (;;) { 1368 for (;;) {
1369 struct select_data data; 1369 struct select_data data;
1370 1370
1371 INIT_LIST_HEAD(&data.dispose); 1371 INIT_LIST_HEAD(&data.dispose);
1372 data.start = dentry; 1372 data.start = dentry;
1373 data.found = 0; 1373 data.found = 0;
1374 1374
1375 d_walk(dentry, &data, check_and_collect, check_and_drop); 1375 d_walk(dentry, &data, check_and_collect, check_and_drop);
1376 ret = data.found; 1376 ret = data.found;
1377 1377
1378 if (!list_empty(&data.dispose)) 1378 if (!list_empty(&data.dispose))
1379 shrink_dentry_list(&data.dispose); 1379 shrink_dentry_list(&data.dispose);
1380 1380
1381 if (ret <= 0) 1381 if (ret <= 0)
1382 break; 1382 break;
1383 1383
1384 cond_resched(); 1384 cond_resched();
1385 } 1385 }
1386 1386
1387 out: 1387 out:
1388 return ret; 1388 return ret;
1389 } 1389 }
1390 EXPORT_SYMBOL(check_submounts_and_drop); 1390 EXPORT_SYMBOL(check_submounts_and_drop);
1391 1391
1392 /** 1392 /**
1393 * __d_alloc - allocate a dcache entry 1393 * __d_alloc - allocate a dcache entry
1394 * @sb: filesystem it will belong to 1394 * @sb: filesystem it will belong to
1395 * @name: qstr of the name 1395 * @name: qstr of the name
1396 * 1396 *
1397 * Allocates a dentry. It returns %NULL if there is insufficient memory 1397 * Allocates a dentry. It returns %NULL if there is insufficient memory
1398 * available. On a success the dentry is returned. The name passed in is 1398 * available. On a success the dentry is returned. The name passed in is
1399 * copied and the copy passed in may be reused after this call. 1399 * copied and the copy passed in may be reused after this call.
1400 */ 1400 */
1401 1401
1402 struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) 1402 struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name)
1403 { 1403 {
1404 struct dentry *dentry; 1404 struct dentry *dentry;
1405 char *dname; 1405 char *dname;
1406 1406
1407 dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL); 1407 dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
1408 if (!dentry) 1408 if (!dentry)
1409 return NULL; 1409 return NULL;
1410 1410
1411 /* 1411 /*
1412 * We guarantee that the inline name is always NUL-terminated. 1412 * We guarantee that the inline name is always NUL-terminated.
1413 * This way the memcpy() done by the name switching in rename 1413 * This way the memcpy() done by the name switching in rename
1414 * will still always have a NUL at the end, even if we might 1414 * will still always have a NUL at the end, even if we might
1415 * be overwriting an internal NUL character 1415 * be overwriting an internal NUL character
1416 */ 1416 */
1417 dentry->d_iname[DNAME_INLINE_LEN-1] = 0; 1417 dentry->d_iname[DNAME_INLINE_LEN-1] = 0;
1418 if (name->len > DNAME_INLINE_LEN-1) { 1418 if (name->len > DNAME_INLINE_LEN-1) {
1419 dname = kmalloc(name->len + 1, GFP_KERNEL); 1419 dname = kmalloc(name->len + 1, GFP_KERNEL);
1420 if (!dname) { 1420 if (!dname) {
1421 kmem_cache_free(dentry_cache, dentry); 1421 kmem_cache_free(dentry_cache, dentry);
1422 return NULL; 1422 return NULL;
1423 } 1423 }
1424 } else { 1424 } else {
1425 dname = dentry->d_iname; 1425 dname = dentry->d_iname;
1426 } 1426 }
1427 1427
1428 dentry->d_name.len = name->len; 1428 dentry->d_name.len = name->len;
1429 dentry->d_name.hash = name->hash; 1429 dentry->d_name.hash = name->hash;
1430 memcpy(dname, name->name, name->len); 1430 memcpy(dname, name->name, name->len);
1431 dname[name->len] = 0; 1431 dname[name->len] = 0;
1432 1432
1433 /* Make sure we always see the terminating NUL character */ 1433 /* Make sure we always see the terminating NUL character */
1434 smp_wmb(); 1434 smp_wmb();
1435 dentry->d_name.name = dname; 1435 dentry->d_name.name = dname;
1436 1436
1437 dentry->d_lockref.count = 1; 1437 dentry->d_lockref.count = 1;
1438 dentry->d_flags = 0; 1438 dentry->d_flags = 0;
1439 spin_lock_init(&dentry->d_lock); 1439 spin_lock_init(&dentry->d_lock);
1440 seqcount_init(&dentry->d_seq); 1440 seqcount_init(&dentry->d_seq);
1441 dentry->d_inode = NULL; 1441 dentry->d_inode = NULL;
1442 dentry->d_parent = dentry; 1442 dentry->d_parent = dentry;
1443 dentry->d_sb = sb; 1443 dentry->d_sb = sb;
1444 dentry->d_op = NULL; 1444 dentry->d_op = NULL;
1445 dentry->d_fsdata = NULL; 1445 dentry->d_fsdata = NULL;
1446 INIT_HLIST_BL_NODE(&dentry->d_hash); 1446 INIT_HLIST_BL_NODE(&dentry->d_hash);
1447 INIT_LIST_HEAD(&dentry->d_lru); 1447 INIT_LIST_HEAD(&dentry->d_lru);
1448 INIT_LIST_HEAD(&dentry->d_subdirs); 1448 INIT_LIST_HEAD(&dentry->d_subdirs);
1449 INIT_HLIST_NODE(&dentry->d_alias); 1449 INIT_HLIST_NODE(&dentry->d_alias);
1450 INIT_LIST_HEAD(&dentry->d_u.d_child); 1450 INIT_LIST_HEAD(&dentry->d_u.d_child);
1451 d_set_d_op(dentry, dentry->d_sb->s_d_op); 1451 d_set_d_op(dentry, dentry->d_sb->s_d_op);
1452 1452
1453 this_cpu_inc(nr_dentry); 1453 this_cpu_inc(nr_dentry);
1454 1454
1455 return dentry; 1455 return dentry;
1456 } 1456 }
1457 1457
1458 /** 1458 /**
1459 * d_alloc - allocate a dcache entry 1459 * d_alloc - allocate a dcache entry
1460 * @parent: parent of entry to allocate 1460 * @parent: parent of entry to allocate
1461 * @name: qstr of the name 1461 * @name: qstr of the name
1462 * 1462 *
1463 * Allocates a dentry. It returns %NULL if there is insufficient memory 1463 * Allocates a dentry. It returns %NULL if there is insufficient memory
1464 * available. On a success the dentry is returned. The name passed in is 1464 * available. On a success the dentry is returned. The name passed in is
1465 * copied and the copy passed in may be reused after this call. 1465 * copied and the copy passed in may be reused after this call.
1466 */ 1466 */
1467 struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) 1467 struct dentry *d_alloc(struct dentry * parent, const struct qstr *name)
1468 { 1468 {
1469 struct dentry *dentry = __d_alloc(parent->d_sb, name); 1469 struct dentry *dentry = __d_alloc(parent->d_sb, name);
1470 if (!dentry) 1470 if (!dentry)
1471 return NULL; 1471 return NULL;
1472 1472
1473 spin_lock(&parent->d_lock); 1473 spin_lock(&parent->d_lock);
1474 /* 1474 /*
1475 * don't need child lock because it is not subject 1475 * don't need child lock because it is not subject
1476 * to concurrency here 1476 * to concurrency here
1477 */ 1477 */
1478 __dget_dlock(parent); 1478 __dget_dlock(parent);
1479 dentry->d_parent = parent; 1479 dentry->d_parent = parent;
1480 list_add(&dentry->d_u.d_child, &parent->d_subdirs); 1480 list_add(&dentry->d_u.d_child, &parent->d_subdirs);
1481 spin_unlock(&parent->d_lock); 1481 spin_unlock(&parent->d_lock);
1482 1482
1483 return dentry; 1483 return dentry;
1484 } 1484 }
1485 EXPORT_SYMBOL(d_alloc); 1485 EXPORT_SYMBOL(d_alloc);
1486 1486
1487 struct dentry *d_alloc_pseudo(struct super_block *sb, const struct qstr *name) 1487 struct dentry *d_alloc_pseudo(struct super_block *sb, const struct qstr *name)
1488 { 1488 {
1489 struct dentry *dentry = __d_alloc(sb, name); 1489 struct dentry *dentry = __d_alloc(sb, name);
1490 if (dentry) 1490 if (dentry)
1491 dentry->d_flags |= DCACHE_DISCONNECTED; 1491 dentry->d_flags |= DCACHE_DISCONNECTED;
1492 return dentry; 1492 return dentry;
1493 } 1493 }
1494 EXPORT_SYMBOL(d_alloc_pseudo); 1494 EXPORT_SYMBOL(d_alloc_pseudo);
1495 1495
1496 struct dentry *d_alloc_name(struct dentry *parent, const char *name) 1496 struct dentry *d_alloc_name(struct dentry *parent, const char *name)
1497 { 1497 {
1498 struct qstr q; 1498 struct qstr q;
1499 1499
1500 q.name = name; 1500 q.name = name;
1501 q.len = strlen(name); 1501 q.len = strlen(name);
1502 q.hash = full_name_hash(q.name, q.len); 1502 q.hash = full_name_hash(q.name, q.len);
1503 return d_alloc(parent, &q); 1503 return d_alloc(parent, &q);
1504 } 1504 }
1505 EXPORT_SYMBOL(d_alloc_name); 1505 EXPORT_SYMBOL(d_alloc_name);
1506 1506
1507 void d_set_d_op(struct dentry *dentry, const struct dentry_operations *op) 1507 void d_set_d_op(struct dentry *dentry, const struct dentry_operations *op)
1508 { 1508 {
1509 WARN_ON_ONCE(dentry->d_op); 1509 WARN_ON_ONCE(dentry->d_op);
1510 WARN_ON_ONCE(dentry->d_flags & (DCACHE_OP_HASH | 1510 WARN_ON_ONCE(dentry->d_flags & (DCACHE_OP_HASH |
1511 DCACHE_OP_COMPARE | 1511 DCACHE_OP_COMPARE |
1512 DCACHE_OP_REVALIDATE | 1512 DCACHE_OP_REVALIDATE |
1513 DCACHE_OP_WEAK_REVALIDATE | 1513 DCACHE_OP_WEAK_REVALIDATE |
1514 DCACHE_OP_DELETE )); 1514 DCACHE_OP_DELETE ));
1515 dentry->d_op = op; 1515 dentry->d_op = op;
1516 if (!op) 1516 if (!op)
1517 return; 1517 return;
1518 if (op->d_hash) 1518 if (op->d_hash)
1519 dentry->d_flags |= DCACHE_OP_HASH; 1519 dentry->d_flags |= DCACHE_OP_HASH;
1520 if (op->d_compare) 1520 if (op->d_compare)
1521 dentry->d_flags |= DCACHE_OP_COMPARE; 1521 dentry->d_flags |= DCACHE_OP_COMPARE;
1522 if (op->d_revalidate) 1522 if (op->d_revalidate)
1523 dentry->d_flags |= DCACHE_OP_REVALIDATE; 1523 dentry->d_flags |= DCACHE_OP_REVALIDATE;
1524 if (op->d_weak_revalidate) 1524 if (op->d_weak_revalidate)
1525 dentry->d_flags |= DCACHE_OP_WEAK_REVALIDATE; 1525 dentry->d_flags |= DCACHE_OP_WEAK_REVALIDATE;
1526 if (op->d_delete) 1526 if (op->d_delete)
1527 dentry->d_flags |= DCACHE_OP_DELETE; 1527 dentry->d_flags |= DCACHE_OP_DELETE;
1528 if (op->d_prune) 1528 if (op->d_prune)
1529 dentry->d_flags |= DCACHE_OP_PRUNE; 1529 dentry->d_flags |= DCACHE_OP_PRUNE;
1530 1530
1531 } 1531 }
1532 EXPORT_SYMBOL(d_set_d_op); 1532 EXPORT_SYMBOL(d_set_d_op);
1533 1533
1534 static void __d_instantiate(struct dentry *dentry, struct inode *inode) 1534 static void __d_instantiate(struct dentry *dentry, struct inode *inode)
1535 { 1535 {
1536 spin_lock(&dentry->d_lock); 1536 spin_lock(&dentry->d_lock);
1537 if (inode) { 1537 if (inode) {
1538 if (unlikely(IS_AUTOMOUNT(inode))) 1538 if (unlikely(IS_AUTOMOUNT(inode)))
1539 dentry->d_flags |= DCACHE_NEED_AUTOMOUNT; 1539 dentry->d_flags |= DCACHE_NEED_AUTOMOUNT;
1540 hlist_add_head(&dentry->d_alias, &inode->i_dentry); 1540 hlist_add_head(&dentry->d_alias, &inode->i_dentry);
1541 } 1541 }
1542 dentry->d_inode = inode; 1542 dentry->d_inode = inode;
1543 dentry_rcuwalk_barrier(dentry); 1543 dentry_rcuwalk_barrier(dentry);
1544 spin_unlock(&dentry->d_lock); 1544 spin_unlock(&dentry->d_lock);
1545 fsnotify_d_instantiate(dentry, inode); 1545 fsnotify_d_instantiate(dentry, inode);
1546 } 1546 }
1547 1547
1548 /** 1548 /**
1549 * d_instantiate - fill in inode information for a dentry 1549 * d_instantiate - fill in inode information for a dentry
1550 * @entry: dentry to complete 1550 * @entry: dentry to complete
1551 * @inode: inode to attach to this dentry 1551 * @inode: inode to attach to this dentry
1552 * 1552 *
1553 * Fill in inode information in the entry. 1553 * Fill in inode information in the entry.
1554 * 1554 *
1555 * This turns negative dentries into productive full members 1555 * This turns negative dentries into productive full members
1556 * of society. 1556 * of society.
1557 * 1557 *
1558 * NOTE! This assumes that the inode count has been incremented 1558 * NOTE! This assumes that the inode count has been incremented
1559 * (or otherwise set) by the caller to indicate that it is now 1559 * (or otherwise set) by the caller to indicate that it is now
1560 * in use by the dcache. 1560 * in use by the dcache.
1561 */ 1561 */
1562 1562
1563 void d_instantiate(struct dentry *entry, struct inode * inode) 1563 void d_instantiate(struct dentry *entry, struct inode * inode)
1564 { 1564 {
1565 BUG_ON(!hlist_unhashed(&entry->d_alias)); 1565 BUG_ON(!hlist_unhashed(&entry->d_alias));
1566 if (inode) 1566 if (inode)
1567 spin_lock(&inode->i_lock); 1567 spin_lock(&inode->i_lock);
1568 __d_instantiate(entry, inode); 1568 __d_instantiate(entry, inode);
1569 if (inode) 1569 if (inode)
1570 spin_unlock(&inode->i_lock); 1570 spin_unlock(&inode->i_lock);
1571 security_d_instantiate(entry, inode); 1571 security_d_instantiate(entry, inode);
1572 } 1572 }
1573 EXPORT_SYMBOL(d_instantiate); 1573 EXPORT_SYMBOL(d_instantiate);
1574 1574
1575 /** 1575 /**
1576 * d_instantiate_unique - instantiate a non-aliased dentry 1576 * d_instantiate_unique - instantiate a non-aliased dentry
1577 * @entry: dentry to instantiate 1577 * @entry: dentry to instantiate
1578 * @inode: inode to attach to this dentry 1578 * @inode: inode to attach to this dentry
1579 * 1579 *
1580 * Fill in inode information in the entry. On success, it returns NULL. 1580 * Fill in inode information in the entry. On success, it returns NULL.
1581 * If an unhashed alias of "entry" already exists, then we return the 1581 * If an unhashed alias of "entry" already exists, then we return the
1582 * aliased dentry instead and drop one reference to inode. 1582 * aliased dentry instead and drop one reference to inode.
1583 * 1583 *
1584 * Note that in order to avoid conflicts with rename() etc, the caller 1584 * Note that in order to avoid conflicts with rename() etc, the caller
1585 * had better be holding the parent directory semaphore. 1585 * had better be holding the parent directory semaphore.
1586 * 1586 *
1587 * This also assumes that the inode count has been incremented 1587 * This also assumes that the inode count has been incremented
1588 * (or otherwise set) by the caller to indicate that it is now 1588 * (or otherwise set) by the caller to indicate that it is now
1589 * in use by the dcache. 1589 * in use by the dcache.
1590 */ 1590 */
1591 static struct dentry *__d_instantiate_unique(struct dentry *entry, 1591 static struct dentry *__d_instantiate_unique(struct dentry *entry,
1592 struct inode *inode) 1592 struct inode *inode)
1593 { 1593 {
1594 struct dentry *alias; 1594 struct dentry *alias;
1595 int len = entry->d_name.len; 1595 int len = entry->d_name.len;
1596 const char *name = entry->d_name.name; 1596 const char *name = entry->d_name.name;
1597 unsigned int hash = entry->d_name.hash; 1597 unsigned int hash = entry->d_name.hash;
1598 1598
1599 if (!inode) { 1599 if (!inode) {
1600 __d_instantiate(entry, NULL); 1600 __d_instantiate(entry, NULL);
1601 return NULL; 1601 return NULL;
1602 } 1602 }
1603 1603
1604 hlist_for_each_entry(alias, &inode->i_dentry, d_alias) { 1604 hlist_for_each_entry(alias, &inode->i_dentry, d_alias) {
1605 /* 1605 /*
1606 * Don't need alias->d_lock here, because aliases with 1606 * Don't need alias->d_lock here, because aliases with
1607 * d_parent == entry->d_parent are not subject to name or 1607 * d_parent == entry->d_parent are not subject to name or
1608 * parent changes, because the parent inode i_mutex is held. 1608 * parent changes, because the parent inode i_mutex is held.
1609 */ 1609 */
1610 if (alias->d_name.hash != hash) 1610 if (alias->d_name.hash != hash)
1611 continue; 1611 continue;
1612 if (alias->d_parent != entry->d_parent) 1612 if (alias->d_parent != entry->d_parent)
1613 continue; 1613 continue;
1614 if (alias->d_name.len != len) 1614 if (alias->d_name.len != len)
1615 continue; 1615 continue;
1616 if (dentry_cmp(alias, name, len)) 1616 if (dentry_cmp(alias, name, len))
1617 continue; 1617 continue;
1618 __dget(alias); 1618 __dget(alias);
1619 return alias; 1619 return alias;
1620 } 1620 }
1621 1621
1622 __d_instantiate(entry, inode); 1622 __d_instantiate(entry, inode);
1623 return NULL; 1623 return NULL;
1624 } 1624 }
1625 1625
1626 struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode) 1626 struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode)
1627 { 1627 {
1628 struct dentry *result; 1628 struct dentry *result;
1629 1629
1630 BUG_ON(!hlist_unhashed(&entry->d_alias)); 1630 BUG_ON(!hlist_unhashed(&entry->d_alias));
1631 1631
1632 if (inode) 1632 if (inode)
1633 spin_lock(&inode->i_lock); 1633 spin_lock(&inode->i_lock);
1634 result = __d_instantiate_unique(entry, inode); 1634 result = __d_instantiate_unique(entry, inode);
1635 if (inode) 1635 if (inode)
1636 spin_unlock(&inode->i_lock); 1636 spin_unlock(&inode->i_lock);
1637 1637
1638 if (!result) { 1638 if (!result) {
1639 security_d_instantiate(entry, inode); 1639 security_d_instantiate(entry, inode);
1640 return NULL; 1640 return NULL;
1641 } 1641 }
1642 1642
1643 BUG_ON(!d_unhashed(result)); 1643 BUG_ON(!d_unhashed(result));
1644 iput(inode); 1644 iput(inode);
1645 return result; 1645 return result;
1646 } 1646 }
1647 1647
1648 EXPORT_SYMBOL(d_instantiate_unique); 1648 EXPORT_SYMBOL(d_instantiate_unique);
1649 1649
1650 struct dentry *d_make_root(struct inode *root_inode) 1650 struct dentry *d_make_root(struct inode *root_inode)
1651 { 1651 {
1652 struct dentry *res = NULL; 1652 struct dentry *res = NULL;
1653 1653
1654 if (root_inode) { 1654 if (root_inode) {
1655 static const struct qstr name = QSTR_INIT("/", 1); 1655 static const struct qstr name = QSTR_INIT("/", 1);
1656 1656
1657 res = __d_alloc(root_inode->i_sb, &name); 1657 res = __d_alloc(root_inode->i_sb, &name);
1658 if (res) 1658 if (res)
1659 d_instantiate(res, root_inode); 1659 d_instantiate(res, root_inode);
1660 else 1660 else
1661 iput(root_inode); 1661 iput(root_inode);
1662 } 1662 }
1663 return res; 1663 return res;
1664 } 1664 }
1665 EXPORT_SYMBOL(d_make_root); 1665 EXPORT_SYMBOL(d_make_root);
1666 1666
1667 static struct dentry * __d_find_any_alias(struct inode *inode) 1667 static struct dentry * __d_find_any_alias(struct inode *inode)
1668 { 1668 {
1669 struct dentry *alias; 1669 struct dentry *alias;
1670 1670
1671 if (hlist_empty(&inode->i_dentry)) 1671 if (hlist_empty(&inode->i_dentry))
1672 return NULL; 1672 return NULL;
1673 alias = hlist_entry(inode->i_dentry.first, struct dentry, d_alias); 1673 alias = hlist_entry(inode->i_dentry.first, struct dentry, d_alias);
1674 __dget(alias); 1674 __dget(alias);
1675 return alias; 1675 return alias;
1676 } 1676 }
1677 1677
1678 /** 1678 /**
1679 * d_find_any_alias - find any alias for a given inode 1679 * d_find_any_alias - find any alias for a given inode
1680 * @inode: inode to find an alias for 1680 * @inode: inode to find an alias for
1681 * 1681 *
1682 * If any aliases exist for the given inode, take and return a 1682 * If any aliases exist for the given inode, take and return a
1683 * reference for one of them. If no aliases exist, return %NULL. 1683 * reference for one of them. If no aliases exist, return %NULL.
1684 */ 1684 */
1685 struct dentry *d_find_any_alias(struct inode *inode) 1685 struct dentry *d_find_any_alias(struct inode *inode)
1686 { 1686 {
1687 struct dentry *de; 1687 struct dentry *de;
1688 1688
1689 spin_lock(&inode->i_lock); 1689 spin_lock(&inode->i_lock);
1690 de = __d_find_any_alias(inode); 1690 de = __d_find_any_alias(inode);
1691 spin_unlock(&inode->i_lock); 1691 spin_unlock(&inode->i_lock);
1692 return de; 1692 return de;
1693 } 1693 }
1694 EXPORT_SYMBOL(d_find_any_alias); 1694 EXPORT_SYMBOL(d_find_any_alias);
1695 1695
1696 /** 1696 /**
1697 * d_obtain_alias - find or allocate a dentry for a given inode 1697 * d_obtain_alias - find or allocate a dentry for a given inode
1698 * @inode: inode to allocate the dentry for 1698 * @inode: inode to allocate the dentry for
1699 * 1699 *
1700 * Obtain a dentry for an inode resulting from NFS filehandle conversion or 1700 * Obtain a dentry for an inode resulting from NFS filehandle conversion or
1701 * similar open by handle operations. The returned dentry may be anonymous, 1701 * similar open by handle operations. The returned dentry may be anonymous,
1702 * or may have a full name (if the inode was already in the cache). 1702 * or may have a full name (if the inode was already in the cache).
1703 * 1703 *
1704 * When called on a directory inode, we must ensure that the inode only ever 1704 * When called on a directory inode, we must ensure that the inode only ever
1705 * has one dentry. If a dentry is found, that is returned instead of 1705 * has one dentry. If a dentry is found, that is returned instead of
1706 * allocating a new one. 1706 * allocating a new one.
1707 * 1707 *
1708 * On successful return, the reference to the inode has been transferred 1708 * On successful return, the reference to the inode has been transferred
1709 * to the dentry. In case of an error the reference on the inode is released. 1709 * to the dentry. In case of an error the reference on the inode is released.
1710 * To make it easier to use in export operations a %NULL or IS_ERR inode may 1710 * To make it easier to use in export operations a %NULL or IS_ERR inode may
1711 * be passed in and will be the error will be propagate to the return value, 1711 * be passed in and will be the error will be propagate to the return value,
1712 * with a %NULL @inode replaced by ERR_PTR(-ESTALE). 1712 * with a %NULL @inode replaced by ERR_PTR(-ESTALE).
1713 */ 1713 */
1714 struct dentry *d_obtain_alias(struct inode *inode) 1714 struct dentry *d_obtain_alias(struct inode *inode)
1715 { 1715 {
1716 static const struct qstr anonstring = QSTR_INIT("/", 1); 1716 static const struct qstr anonstring = QSTR_INIT("/", 1);
1717 struct dentry *tmp; 1717 struct dentry *tmp;
1718 struct dentry *res; 1718 struct dentry *res;
1719 1719
1720 if (!inode) 1720 if (!inode)
1721 return ERR_PTR(-ESTALE); 1721 return ERR_PTR(-ESTALE);
1722 if (IS_ERR(inode)) 1722 if (IS_ERR(inode))
1723 return ERR_CAST(inode); 1723 return ERR_CAST(inode);
1724 1724
1725 res = d_find_any_alias(inode); 1725 res = d_find_any_alias(inode);
1726 if (res) 1726 if (res)
1727 goto out_iput; 1727 goto out_iput;
1728 1728
1729 tmp = __d_alloc(inode->i_sb, &anonstring); 1729 tmp = __d_alloc(inode->i_sb, &anonstring);
1730 if (!tmp) { 1730 if (!tmp) {
1731 res = ERR_PTR(-ENOMEM); 1731 res = ERR_PTR(-ENOMEM);
1732 goto out_iput; 1732 goto out_iput;
1733 } 1733 }
1734 1734
1735 spin_lock(&inode->i_lock); 1735 spin_lock(&inode->i_lock);
1736 res = __d_find_any_alias(inode); 1736 res = __d_find_any_alias(inode);
1737 if (res) { 1737 if (res) {
1738 spin_unlock(&inode->i_lock); 1738 spin_unlock(&inode->i_lock);
1739 dput(tmp); 1739 dput(tmp);
1740 goto out_iput; 1740 goto out_iput;
1741 } 1741 }
1742 1742
1743 /* attach a disconnected dentry */ 1743 /* attach a disconnected dentry */
1744 spin_lock(&tmp->d_lock); 1744 spin_lock(&tmp->d_lock);
1745 tmp->d_inode = inode; 1745 tmp->d_inode = inode;
1746 tmp->d_flags |= DCACHE_DISCONNECTED; 1746 tmp->d_flags |= DCACHE_DISCONNECTED;
1747 hlist_add_head(&tmp->d_alias, &inode->i_dentry); 1747 hlist_add_head(&tmp->d_alias, &inode->i_dentry);
1748 hlist_bl_lock(&tmp->d_sb->s_anon); 1748 hlist_bl_lock(&tmp->d_sb->s_anon);
1749 hlist_bl_add_head(&tmp->d_hash, &tmp->d_sb->s_anon); 1749 hlist_bl_add_head(&tmp->d_hash, &tmp->d_sb->s_anon);
1750 hlist_bl_unlock(&tmp->d_sb->s_anon); 1750 hlist_bl_unlock(&tmp->d_sb->s_anon);
1751 spin_unlock(&tmp->d_lock); 1751 spin_unlock(&tmp->d_lock);
1752 spin_unlock(&inode->i_lock); 1752 spin_unlock(&inode->i_lock);
1753 security_d_instantiate(tmp, inode); 1753 security_d_instantiate(tmp, inode);
1754 1754
1755 return tmp; 1755 return tmp;
1756 1756
1757 out_iput: 1757 out_iput:
1758 if (res && !IS_ERR(res)) 1758 if (res && !IS_ERR(res))
1759 security_d_instantiate(res, inode); 1759 security_d_instantiate(res, inode);
1760 iput(inode); 1760 iput(inode);
1761 return res; 1761 return res;
1762 } 1762 }
1763 EXPORT_SYMBOL(d_obtain_alias); 1763 EXPORT_SYMBOL(d_obtain_alias);
1764 1764
1765 /** 1765 /**
1766 * d_splice_alias - splice a disconnected dentry into the tree if one exists 1766 * d_splice_alias - splice a disconnected dentry into the tree if one exists
1767 * @inode: the inode which may have a disconnected dentry 1767 * @inode: the inode which may have a disconnected dentry
1768 * @dentry: a negative dentry which we want to point to the inode. 1768 * @dentry: a negative dentry which we want to point to the inode.
1769 * 1769 *
1770 * If inode is a directory and has a 'disconnected' dentry (i.e. IS_ROOT and 1770 * If inode is a directory and has a 'disconnected' dentry (i.e. IS_ROOT and
1771 * DCACHE_DISCONNECTED), then d_move that in place of the given dentry 1771 * DCACHE_DISCONNECTED), then d_move that in place of the given dentry
1772 * and return it, else simply d_add the inode to the dentry and return NULL. 1772 * and return it, else simply d_add the inode to the dentry and return NULL.
1773 * 1773 *
1774 * This is needed in the lookup routine of any filesystem that is exportable 1774 * This is needed in the lookup routine of any filesystem that is exportable
1775 * (via knfsd) so that we can build dcache paths to directories effectively. 1775 * (via knfsd) so that we can build dcache paths to directories effectively.
1776 * 1776 *
1777 * If a dentry was found and moved, then it is returned. Otherwise NULL 1777 * If a dentry was found and moved, then it is returned. Otherwise NULL
1778 * is returned. This matches the expected return value of ->lookup. 1778 * is returned. This matches the expected return value of ->lookup.
1779 * 1779 *
1780 * Cluster filesystems may call this function with a negative, hashed dentry. 1780 * Cluster filesystems may call this function with a negative, hashed dentry.
1781 * In that case, we know that the inode will be a regular file, and also this 1781 * In that case, we know that the inode will be a regular file, and also this
1782 * will only occur during atomic_open. So we need to check for the dentry 1782 * will only occur during atomic_open. So we need to check for the dentry
1783 * being already hashed only in the final case. 1783 * being already hashed only in the final case.
1784 */ 1784 */
1785 struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry) 1785 struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry)
1786 { 1786 {
1787 struct dentry *new = NULL; 1787 struct dentry *new = NULL;
1788 1788
1789 if (IS_ERR(inode)) 1789 if (IS_ERR(inode))
1790 return ERR_CAST(inode); 1790 return ERR_CAST(inode);
1791 1791
1792 if (inode && S_ISDIR(inode->i_mode)) { 1792 if (inode && S_ISDIR(inode->i_mode)) {
1793 spin_lock(&inode->i_lock); 1793 spin_lock(&inode->i_lock);
1794 new = __d_find_alias(inode, 1); 1794 new = __d_find_alias(inode, 1);
1795 if (new) { 1795 if (new) {
1796 BUG_ON(!(new->d_flags & DCACHE_DISCONNECTED)); 1796 BUG_ON(!(new->d_flags & DCACHE_DISCONNECTED));
1797 spin_unlock(&inode->i_lock); 1797 spin_unlock(&inode->i_lock);
1798 security_d_instantiate(new, inode); 1798 security_d_instantiate(new, inode);
1799 d_move(new, dentry); 1799 d_move(new, dentry);
1800 iput(inode); 1800 iput(inode);
1801 } else { 1801 } else {
1802 /* already taking inode->i_lock, so d_add() by hand */ 1802 /* already taking inode->i_lock, so d_add() by hand */
1803 __d_instantiate(dentry, inode); 1803 __d_instantiate(dentry, inode);
1804 spin_unlock(&inode->i_lock); 1804 spin_unlock(&inode->i_lock);
1805 security_d_instantiate(dentry, inode); 1805 security_d_instantiate(dentry, inode);
1806 d_rehash(dentry); 1806 d_rehash(dentry);
1807 } 1807 }
1808 } else { 1808 } else {
1809 d_instantiate(dentry, inode); 1809 d_instantiate(dentry, inode);
1810 if (d_unhashed(dentry)) 1810 if (d_unhashed(dentry))
1811 d_rehash(dentry); 1811 d_rehash(dentry);
1812 } 1812 }
1813 return new; 1813 return new;
1814 } 1814 }
1815 EXPORT_SYMBOL(d_splice_alias); 1815 EXPORT_SYMBOL(d_splice_alias);
1816 1816
1817 /** 1817 /**
1818 * d_add_ci - lookup or allocate new dentry with case-exact name 1818 * d_add_ci - lookup or allocate new dentry with case-exact name
1819 * @inode: the inode case-insensitive lookup has found 1819 * @inode: the inode case-insensitive lookup has found
1820 * @dentry: the negative dentry that was passed to the parent's lookup func 1820 * @dentry: the negative dentry that was passed to the parent's lookup func
1821 * @name: the case-exact name to be associated with the returned dentry 1821 * @name: the case-exact name to be associated with the returned dentry
1822 * 1822 *
1823 * This is to avoid filling the dcache with case-insensitive names to the 1823 * This is to avoid filling the dcache with case-insensitive names to the
1824 * same inode, only the actual correct case is stored in the dcache for 1824 * same inode, only the actual correct case is stored in the dcache for
1825 * case-insensitive filesystems. 1825 * case-insensitive filesystems.
1826 * 1826 *
1827 * For a case-insensitive lookup match and if the the case-exact dentry 1827 * For a case-insensitive lookup match and if the the case-exact dentry
1828 * already exists in in the dcache, use it and return it. 1828 * already exists in in the dcache, use it and return it.
1829 * 1829 *
1830 * If no entry exists with the exact case name, allocate new dentry with 1830 * If no entry exists with the exact case name, allocate new dentry with
1831 * the exact case, and return the spliced entry. 1831 * the exact case, and return the spliced entry.
1832 */ 1832 */
1833 struct dentry *d_add_ci(struct dentry *dentry, struct inode *inode, 1833 struct dentry *d_add_ci(struct dentry *dentry, struct inode *inode,
1834 struct qstr *name) 1834 struct qstr *name)
1835 { 1835 {
1836 struct dentry *found; 1836 struct dentry *found;
1837 struct dentry *new; 1837 struct dentry *new;
1838 1838
1839 /* 1839 /*
1840 * First check if a dentry matching the name already exists, 1840 * First check if a dentry matching the name already exists,
1841 * if not go ahead and create it now. 1841 * if not go ahead and create it now.
1842 */ 1842 */
1843 found = d_hash_and_lookup(dentry->d_parent, name); 1843 found = d_hash_and_lookup(dentry->d_parent, name);
1844 if (unlikely(IS_ERR(found))) 1844 if (unlikely(IS_ERR(found)))
1845 goto err_out; 1845 goto err_out;
1846 if (!found) { 1846 if (!found) {
1847 new = d_alloc(dentry->d_parent, name); 1847 new = d_alloc(dentry->d_parent, name);
1848 if (!new) { 1848 if (!new) {
1849 found = ERR_PTR(-ENOMEM); 1849 found = ERR_PTR(-ENOMEM);
1850 goto err_out; 1850 goto err_out;
1851 } 1851 }
1852 1852
1853 found = d_splice_alias(inode, new); 1853 found = d_splice_alias(inode, new);
1854 if (found) { 1854 if (found) {
1855 dput(new); 1855 dput(new);
1856 return found; 1856 return found;
1857 } 1857 }
1858 return new; 1858 return new;
1859 } 1859 }
1860 1860
1861 /* 1861 /*
1862 * If a matching dentry exists, and it's not negative use it. 1862 * If a matching dentry exists, and it's not negative use it.
1863 * 1863 *
1864 * Decrement the reference count to balance the iget() done 1864 * Decrement the reference count to balance the iget() done
1865 * earlier on. 1865 * earlier on.
1866 */ 1866 */
1867 if (found->d_inode) { 1867 if (found->d_inode) {
1868 if (unlikely(found->d_inode != inode)) { 1868 if (unlikely(found->d_inode != inode)) {
1869 /* This can't happen because bad inodes are unhashed. */ 1869 /* This can't happen because bad inodes are unhashed. */
1870 BUG_ON(!is_bad_inode(inode)); 1870 BUG_ON(!is_bad_inode(inode));
1871 BUG_ON(!is_bad_inode(found->d_inode)); 1871 BUG_ON(!is_bad_inode(found->d_inode));
1872 } 1872 }
1873 iput(inode); 1873 iput(inode);
1874 return found; 1874 return found;
1875 } 1875 }
1876 1876
1877 /* 1877 /*
1878 * Negative dentry: instantiate it unless the inode is a directory and 1878 * Negative dentry: instantiate it unless the inode is a directory and
1879 * already has a dentry. 1879 * already has a dentry.
1880 */ 1880 */
1881 new = d_splice_alias(inode, found); 1881 new = d_splice_alias(inode, found);
1882 if (new) { 1882 if (new) {
1883 dput(found); 1883 dput(found);
1884 found = new; 1884 found = new;
1885 } 1885 }
1886 return found; 1886 return found;
1887 1887
1888 err_out: 1888 err_out:
1889 iput(inode); 1889 iput(inode);
1890 return found; 1890 return found;
1891 } 1891 }
1892 EXPORT_SYMBOL(d_add_ci); 1892 EXPORT_SYMBOL(d_add_ci);
1893 1893
1894 /* 1894 /*
1895 * Do the slow-case of the dentry name compare. 1895 * Do the slow-case of the dentry name compare.
1896 * 1896 *
1897 * Unlike the dentry_cmp() function, we need to atomically 1897 * Unlike the dentry_cmp() function, we need to atomically
1898 * load the name and length information, so that the 1898 * load the name and length information, so that the
1899 * filesystem can rely on them, and can use the 'name' and 1899 * filesystem can rely on them, and can use the 'name' and
1900 * 'len' information without worrying about walking off the 1900 * 'len' information without worrying about walking off the
1901 * end of memory etc. 1901 * end of memory etc.
1902 * 1902 *
1903 * Thus the read_seqcount_retry() and the "duplicate" info 1903 * Thus the read_seqcount_retry() and the "duplicate" info
1904 * in arguments (the low-level filesystem should not look 1904 * in arguments (the low-level filesystem should not look
1905 * at the dentry inode or name contents directly, since 1905 * at the dentry inode or name contents directly, since
1906 * rename can change them while we're in RCU mode). 1906 * rename can change them while we're in RCU mode).
1907 */ 1907 */
1908 enum slow_d_compare { 1908 enum slow_d_compare {
1909 D_COMP_OK, 1909 D_COMP_OK,
1910 D_COMP_NOMATCH, 1910 D_COMP_NOMATCH,
1911 D_COMP_SEQRETRY, 1911 D_COMP_SEQRETRY,
1912 }; 1912 };
1913 1913
1914 static noinline enum slow_d_compare slow_dentry_cmp( 1914 static noinline enum slow_d_compare slow_dentry_cmp(
1915 const struct dentry *parent, 1915 const struct dentry *parent,
1916 struct dentry *dentry, 1916 struct dentry *dentry,
1917 unsigned int seq, 1917 unsigned int seq,
1918 const struct qstr *name) 1918 const struct qstr *name)
1919 { 1919 {
1920 int tlen = dentry->d_name.len; 1920 int tlen = dentry->d_name.len;
1921 const char *tname = dentry->d_name.name; 1921 const char *tname = dentry->d_name.name;
1922 1922
1923 if (read_seqcount_retry(&dentry->d_seq, seq)) { 1923 if (read_seqcount_retry(&dentry->d_seq, seq)) {
1924 cpu_relax(); 1924 cpu_relax();
1925 return D_COMP_SEQRETRY; 1925 return D_COMP_SEQRETRY;
1926 } 1926 }
1927 if (parent->d_op->d_compare(parent, dentry, tlen, tname, name)) 1927 if (parent->d_op->d_compare(parent, dentry, tlen, tname, name))
1928 return D_COMP_NOMATCH; 1928 return D_COMP_NOMATCH;
1929 return D_COMP_OK; 1929 return D_COMP_OK;
1930 } 1930 }
1931 1931
1932 /** 1932 /**
1933 * __d_lookup_rcu - search for a dentry (racy, store-free) 1933 * __d_lookup_rcu - search for a dentry (racy, store-free)
1934 * @parent: parent dentry 1934 * @parent: parent dentry
1935 * @name: qstr of name we wish to find 1935 * @name: qstr of name we wish to find
1936 * @seqp: returns d_seq value at the point where the dentry was found 1936 * @seqp: returns d_seq value at the point where the dentry was found
1937 * Returns: dentry, or NULL 1937 * Returns: dentry, or NULL
1938 * 1938 *
1939 * __d_lookup_rcu is the dcache lookup function for rcu-walk name 1939 * __d_lookup_rcu is the dcache lookup function for rcu-walk name
1940 * resolution (store-free path walking) design described in 1940 * resolution (store-free path walking) design described in
1941 * Documentation/filesystems/path-lookup.txt. 1941 * Documentation/filesystems/path-lookup.txt.
1942 * 1942 *
1943 * This is not to be used outside core vfs. 1943 * This is not to be used outside core vfs.
1944 * 1944 *
1945 * __d_lookup_rcu must only be used in rcu-walk mode, ie. with vfsmount lock 1945 * __d_lookup_rcu must only be used in rcu-walk mode, ie. with vfsmount lock
1946 * held, and rcu_read_lock held. The returned dentry must not be stored into 1946 * held, and rcu_read_lock held. The returned dentry must not be stored into
1947 * without taking d_lock and checking d_seq sequence count against @seq 1947 * without taking d_lock and checking d_seq sequence count against @seq
1948 * returned here. 1948 * returned here.
1949 * 1949 *
1950 * A refcount may be taken on the found dentry with the d_rcu_to_refcount 1950 * A refcount may be taken on the found dentry with the d_rcu_to_refcount
1951 * function. 1951 * function.
1952 * 1952 *
1953 * Alternatively, __d_lookup_rcu may be called again to look up the child of 1953 * Alternatively, __d_lookup_rcu may be called again to look up the child of
1954 * the returned dentry, so long as its parent's seqlock is checked after the 1954 * the returned dentry, so long as its parent's seqlock is checked after the
1955 * child is looked up. Thus, an interlocking stepping of sequence lock checks 1955 * child is looked up. Thus, an interlocking stepping of sequence lock checks
1956 * is formed, giving integrity down the path walk. 1956 * is formed, giving integrity down the path walk.
1957 * 1957 *
1958 * NOTE! The caller *has* to check the resulting dentry against the sequence 1958 * NOTE! The caller *has* to check the resulting dentry against the sequence
1959 * number we've returned before using any of the resulting dentry state! 1959 * number we've returned before using any of the resulting dentry state!
1960 */ 1960 */
1961 struct dentry *__d_lookup_rcu(const struct dentry *parent, 1961 struct dentry *__d_lookup_rcu(const struct dentry *parent,
1962 const struct qstr *name, 1962 const struct qstr *name,
1963 unsigned *seqp) 1963 unsigned *seqp)
1964 { 1964 {
1965 u64 hashlen = name->hash_len; 1965 u64 hashlen = name->hash_len;
1966 const unsigned char *str = name->name; 1966 const unsigned char *str = name->name;
1967 struct hlist_bl_head *b = d_hash(parent, hashlen_hash(hashlen)); 1967 struct hlist_bl_head *b = d_hash(parent, hashlen_hash(hashlen));
1968 struct hlist_bl_node *node; 1968 struct hlist_bl_node *node;
1969 struct dentry *dentry; 1969 struct dentry *dentry;
1970 1970
1971 /* 1971 /*
1972 * Note: There is significant duplication with __d_lookup_rcu which is 1972 * Note: There is significant duplication with __d_lookup_rcu which is
1973 * required to prevent single threaded performance regressions 1973 * required to prevent single threaded performance regressions
1974 * especially on architectures where smp_rmb (in seqcounts) are costly. 1974 * especially on architectures where smp_rmb (in seqcounts) are costly.
1975 * Keep the two functions in sync. 1975 * Keep the two functions in sync.
1976 */ 1976 */
1977 1977
1978 /* 1978 /*
1979 * The hash list is protected using RCU. 1979 * The hash list is protected using RCU.
1980 * 1980 *
1981 * Carefully use d_seq when comparing a candidate dentry, to avoid 1981 * Carefully use d_seq when comparing a candidate dentry, to avoid
1982 * races with d_move(). 1982 * races with d_move().
1983 * 1983 *
1984 * It is possible that concurrent renames can mess up our list 1984 * It is possible that concurrent renames can mess up our list
1985 * walk here and result in missing our dentry, resulting in the 1985 * walk here and result in missing our dentry, resulting in the
1986 * false-negative result. d_lookup() protects against concurrent 1986 * false-negative result. d_lookup() protects against concurrent
1987 * renames using rename_lock seqlock. 1987 * renames using rename_lock seqlock.
1988 * 1988 *
1989 * See Documentation/filesystems/path-lookup.txt for more details. 1989 * See Documentation/filesystems/path-lookup.txt for more details.
1990 */ 1990 */
1991 hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) { 1991 hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) {
1992 unsigned seq; 1992 unsigned seq;
1993 1993
1994 seqretry: 1994 seqretry:
1995 /* 1995 /*
1996 * The dentry sequence count protects us from concurrent 1996 * The dentry sequence count protects us from concurrent
1997 * renames, and thus protects parent and name fields. 1997 * renames, and thus protects parent and name fields.
1998 * 1998 *
1999 * The caller must perform a seqcount check in order 1999 * The caller must perform a seqcount check in order
2000 * to do anything useful with the returned dentry. 2000 * to do anything useful with the returned dentry.
2001 * 2001 *
2002 * NOTE! We do a "raw" seqcount_begin here. That means that 2002 * NOTE! We do a "raw" seqcount_begin here. That means that
2003 * we don't wait for the sequence count to stabilize if it 2003 * we don't wait for the sequence count to stabilize if it
2004 * is in the middle of a sequence change. If we do the slow 2004 * is in the middle of a sequence change. If we do the slow
2005 * dentry compare, we will do seqretries until it is stable, 2005 * dentry compare, we will do seqretries until it is stable,
2006 * and if we end up with a successful lookup, we actually 2006 * and if we end up with a successful lookup, we actually
2007 * want to exit RCU lookup anyway. 2007 * want to exit RCU lookup anyway.
2008 */ 2008 */
2009 seq = raw_seqcount_begin(&dentry->d_seq); 2009 seq = raw_seqcount_begin(&dentry->d_seq);
2010 if (dentry->d_parent != parent) 2010 if (dentry->d_parent != parent)
2011 continue; 2011 continue;
2012 if (d_unhashed(dentry)) 2012 if (d_unhashed(dentry))
2013 continue; 2013 continue;
2014 2014
2015 if (unlikely(parent->d_flags & DCACHE_OP_COMPARE)) { 2015 if (unlikely(parent->d_flags & DCACHE_OP_COMPARE)) {
2016 if (dentry->d_name.hash != hashlen_hash(hashlen)) 2016 if (dentry->d_name.hash != hashlen_hash(hashlen))
2017 continue; 2017 continue;
2018 *seqp = seq; 2018 *seqp = seq;
2019 switch (slow_dentry_cmp(parent, dentry, seq, name)) { 2019 switch (slow_dentry_cmp(parent, dentry, seq, name)) {
2020 case D_COMP_OK: 2020 case D_COMP_OK:
2021 return dentry; 2021 return dentry;
2022 case D_COMP_NOMATCH: 2022 case D_COMP_NOMATCH:
2023 continue; 2023 continue;
2024 default: 2024 default:
2025 goto seqretry; 2025 goto seqretry;
2026 } 2026 }
2027 } 2027 }
2028 2028
2029 if (dentry->d_name.hash_len != hashlen) 2029 if (dentry->d_name.hash_len != hashlen)
2030 continue; 2030 continue;
2031 *seqp = seq; 2031 *seqp = seq;
2032 if (!dentry_cmp(dentry, str, hashlen_len(hashlen))) 2032 if (!dentry_cmp(dentry, str, hashlen_len(hashlen)))
2033 return dentry; 2033 return dentry;
2034 } 2034 }
2035 return NULL; 2035 return NULL;
2036 } 2036 }
2037 2037
2038 /** 2038 /**
2039 * d_lookup - search for a dentry 2039 * d_lookup - search for a dentry
2040 * @parent: parent dentry 2040 * @parent: parent dentry
2041 * @name: qstr of name we wish to find 2041 * @name: qstr of name we wish to find
2042 * Returns: dentry, or NULL 2042 * Returns: dentry, or NULL
2043 * 2043 *
2044 * d_lookup searches the children of the parent dentry for the name in 2044 * d_lookup searches the children of the parent dentry for the name in
2045 * question. If the dentry is found its reference count is incremented and the 2045 * question. If the dentry is found its reference count is incremented and the
2046 * dentry is returned. The caller must use dput to free the entry when it has 2046 * dentry is returned. The caller must use dput to free the entry when it has
2047 * finished using it. %NULL is returned if the dentry does not exist. 2047 * finished using it. %NULL is returned if the dentry does not exist.
2048 */ 2048 */
2049 struct dentry *d_lookup(const struct dentry *parent, const struct qstr *name) 2049 struct dentry *d_lookup(const struct dentry *parent, const struct qstr *name)
2050 { 2050 {
2051 struct dentry *dentry; 2051 struct dentry *dentry;
2052 unsigned seq; 2052 unsigned seq;
2053 2053
2054 do { 2054 do {
2055 seq = read_seqbegin(&rename_lock); 2055 seq = read_seqbegin(&rename_lock);
2056 dentry = __d_lookup(parent, name); 2056 dentry = __d_lookup(parent, name);
2057 if (dentry) 2057 if (dentry)
2058 break; 2058 break;
2059 } while (read_seqretry(&rename_lock, seq)); 2059 } while (read_seqretry(&rename_lock, seq));
2060 return dentry; 2060 return dentry;
2061 } 2061 }
2062 EXPORT_SYMBOL(d_lookup); 2062 EXPORT_SYMBOL(d_lookup);
2063 2063
2064 /** 2064 /**
2065 * __d_lookup - search for a dentry (racy) 2065 * __d_lookup - search for a dentry (racy)
2066 * @parent: parent dentry 2066 * @parent: parent dentry
2067 * @name: qstr of name we wish to find 2067 * @name: qstr of name we wish to find
2068 * Returns: dentry, or NULL 2068 * Returns: dentry, or NULL
2069 * 2069 *
2070 * __d_lookup is like d_lookup, however it may (rarely) return a 2070 * __d_lookup is like d_lookup, however it may (rarely) return a
2071 * false-negative result due to unrelated rename activity. 2071 * false-negative result due to unrelated rename activity.
2072 * 2072 *
2073 * __d_lookup is slightly faster by avoiding rename_lock read seqlock, 2073 * __d_lookup is slightly faster by avoiding rename_lock read seqlock,
2074 * however it must be used carefully, eg. with a following d_lookup in 2074 * however it must be used carefully, eg. with a following d_lookup in
2075 * the case of failure. 2075 * the case of failure.
2076 * 2076 *
2077 * __d_lookup callers must be commented. 2077 * __d_lookup callers must be commented.
2078 */ 2078 */
2079 struct dentry *__d_lookup(const struct dentry *parent, const struct qstr *name) 2079 struct dentry *__d_lookup(const struct dentry *parent, const struct qstr *name)
2080 { 2080 {
2081 unsigned int len = name->len; 2081 unsigned int len = name->len;
2082 unsigned int hash = name->hash; 2082 unsigned int hash = name->hash;
2083 const unsigned char *str = name->name; 2083 const unsigned char *str = name->name;
2084 struct hlist_bl_head *b = d_hash(parent, hash); 2084 struct hlist_bl_head *b = d_hash(parent, hash);
2085 struct hlist_bl_node *node; 2085 struct hlist_bl_node *node;
2086 struct dentry *found = NULL; 2086 struct dentry *found = NULL;
2087 struct dentry *dentry; 2087 struct dentry *dentry;
2088 2088
2089 /* 2089 /*
2090 * Note: There is significant duplication with __d_lookup_rcu which is 2090 * Note: There is significant duplication with __d_lookup_rcu which is
2091 * required to prevent single threaded performance regressions 2091 * required to prevent single threaded performance regressions
2092 * especially on architectures where smp_rmb (in seqcounts) are costly. 2092 * especially on architectures where smp_rmb (in seqcounts) are costly.
2093 * Keep the two functions in sync. 2093 * Keep the two functions in sync.
2094 */ 2094 */
2095 2095
2096 /* 2096 /*
2097 * The hash list is protected using RCU. 2097 * The hash list is protected using RCU.
2098 * 2098 *
2099 * Take d_lock when comparing a candidate dentry, to avoid races 2099 * Take d_lock when comparing a candidate dentry, to avoid races
2100 * with d_move(). 2100 * with d_move().
2101 * 2101 *
2102 * It is possible that concurrent renames can mess up our list 2102 * It is possible that concurrent renames can mess up our list
2103 * walk here and result in missing our dentry, resulting in the 2103 * walk here and result in missing our dentry, resulting in the
2104 * false-negative result. d_lookup() protects against concurrent 2104 * false-negative result. d_lookup() protects against concurrent
2105 * renames using rename_lock seqlock. 2105 * renames using rename_lock seqlock.
2106 * 2106 *
2107 * See Documentation/filesystems/path-lookup.txt for more details. 2107 * See Documentation/filesystems/path-lookup.txt for more details.
2108 */ 2108 */
2109 rcu_read_lock(); 2109 rcu_read_lock();
2110 2110
2111 hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) { 2111 hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) {
2112 2112
2113 if (dentry->d_name.hash != hash) 2113 if (dentry->d_name.hash != hash)
2114 continue; 2114 continue;
2115 2115
2116 spin_lock(&dentry->d_lock); 2116 spin_lock(&dentry->d_lock);
2117 if (dentry->d_parent != parent) 2117 if (dentry->d_parent != parent)
2118 goto next; 2118 goto next;
2119 if (d_unhashed(dentry)) 2119 if (d_unhashed(dentry))
2120 goto next; 2120 goto next;
2121 2121
2122 /* 2122 /*
2123 * It is safe to compare names since d_move() cannot 2123 * It is safe to compare names since d_move() cannot
2124 * change the qstr (protected by d_lock). 2124 * change the qstr (protected by d_lock).
2125 */ 2125 */
2126 if (parent->d_flags & DCACHE_OP_COMPARE) { 2126 if (parent->d_flags & DCACHE_OP_COMPARE) {
2127 int tlen = dentry->d_name.len; 2127 int tlen = dentry->d_name.len;
2128 const char *tname = dentry->d_name.name; 2128 const char *tname = dentry->d_name.name;
2129 if (parent->d_op->d_compare(parent, dentry, tlen, tname, name)) 2129 if (parent->d_op->d_compare(parent, dentry, tlen, tname, name))
2130 goto next; 2130 goto next;
2131 } else { 2131 } else {
2132 if (dentry->d_name.len != len) 2132 if (dentry->d_name.len != len)
2133 goto next; 2133 goto next;
2134 if (dentry_cmp(dentry, str, len)) 2134 if (dentry_cmp(dentry, str, len))
2135 goto next; 2135 goto next;
2136 } 2136 }
2137 2137
2138 dentry->d_lockref.count++; 2138 dentry->d_lockref.count++;
2139 found = dentry; 2139 found = dentry;
2140 spin_unlock(&dentry->d_lock); 2140 spin_unlock(&dentry->d_lock);
2141 break; 2141 break;
2142 next: 2142 next:
2143 spin_unlock(&dentry->d_lock); 2143 spin_unlock(&dentry->d_lock);
2144 } 2144 }
2145 rcu_read_unlock(); 2145 rcu_read_unlock();
2146 2146
2147 return found; 2147 return found;
2148 } 2148 }
2149 2149
2150 /** 2150 /**
2151 * d_hash_and_lookup - hash the qstr then search for a dentry 2151 * d_hash_and_lookup - hash the qstr then search for a dentry
2152 * @dir: Directory to search in 2152 * @dir: Directory to search in
2153 * @name: qstr of name we wish to find 2153 * @name: qstr of name we wish to find
2154 * 2154 *
2155 * On lookup failure NULL is returned; on bad name - ERR_PTR(-error) 2155 * On lookup failure NULL is returned; on bad name - ERR_PTR(-error)
2156 */ 2156 */
2157 struct dentry *d_hash_and_lookup(struct dentry *dir, struct qstr *name) 2157 struct dentry *d_hash_and_lookup(struct dentry *dir, struct qstr *name)
2158 { 2158 {
2159 /* 2159 /*
2160 * Check for a fs-specific hash function. Note that we must 2160 * Check for a fs-specific hash function. Note that we must
2161 * calculate the standard hash first, as the d_op->d_hash() 2161 * calculate the standard hash first, as the d_op->d_hash()
2162 * routine may choose to leave the hash value unchanged. 2162 * routine may choose to leave the hash value unchanged.
2163 */ 2163 */
2164 name->hash = full_name_hash(name->name, name->len); 2164 name->hash = full_name_hash(name->name, name->len);
2165 if (dir->d_flags & DCACHE_OP_HASH) { 2165 if (dir->d_flags & DCACHE_OP_HASH) {
2166 int err = dir->d_op->d_hash(dir, name); 2166 int err = dir->d_op->d_hash(dir, name);
2167 if (unlikely(err < 0)) 2167 if (unlikely(err < 0))
2168 return ERR_PTR(err); 2168 return ERR_PTR(err);
2169 } 2169 }
2170 return d_lookup(dir, name); 2170 return d_lookup(dir, name);
2171 } 2171 }
2172 EXPORT_SYMBOL(d_hash_and_lookup); 2172 EXPORT_SYMBOL(d_hash_and_lookup);
2173 2173
2174 /** 2174 /**
2175 * d_validate - verify dentry provided from insecure source (deprecated) 2175 * d_validate - verify dentry provided from insecure source (deprecated)
2176 * @dentry: The dentry alleged to be valid child of @dparent 2176 * @dentry: The dentry alleged to be valid child of @dparent
2177 * @dparent: The parent dentry (known to be valid) 2177 * @dparent: The parent dentry (known to be valid)
2178 * 2178 *
2179 * An insecure source has sent us a dentry, here we verify it and dget() it. 2179 * An insecure source has sent us a dentry, here we verify it and dget() it.
2180 * This is used by ncpfs in its readdir implementation. 2180 * This is used by ncpfs in its readdir implementation.
2181 * Zero is returned in the dentry is invalid. 2181 * Zero is returned in the dentry is invalid.
2182 * 2182 *
2183 * This function is slow for big directories, and deprecated, do not use it. 2183 * This function is slow for big directories, and deprecated, do not use it.
2184 */ 2184 */
2185 int d_validate(struct dentry *dentry, struct dentry *dparent) 2185 int d_validate(struct dentry *dentry, struct dentry *dparent)
2186 { 2186 {
2187 struct dentry *child; 2187 struct dentry *child;
2188 2188
2189 spin_lock(&dparent->d_lock); 2189 spin_lock(&dparent->d_lock);
2190 list_for_each_entry(child, &dparent->d_subdirs, d_u.d_child) { 2190 list_for_each_entry(child, &dparent->d_subdirs, d_u.d_child) {
2191 if (dentry == child) { 2191 if (dentry == child) {
2192 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); 2192 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
2193 __dget_dlock(dentry); 2193 __dget_dlock(dentry);
2194 spin_unlock(&dentry->d_lock); 2194 spin_unlock(&dentry->d_lock);
2195 spin_unlock(&dparent->d_lock); 2195 spin_unlock(&dparent->d_lock);
2196 return 1; 2196 return 1;
2197 } 2197 }
2198 } 2198 }
2199 spin_unlock(&dparent->d_lock); 2199 spin_unlock(&dparent->d_lock);
2200 2200
2201 return 0; 2201 return 0;
2202 } 2202 }
2203 EXPORT_SYMBOL(d_validate); 2203 EXPORT_SYMBOL(d_validate);
2204 2204
2205 /* 2205 /*
2206 * When a file is deleted, we have two options: 2206 * When a file is deleted, we have two options:
2207 * - turn this dentry into a negative dentry 2207 * - turn this dentry into a negative dentry
2208 * - unhash this dentry and free it. 2208 * - unhash this dentry and free it.
2209 * 2209 *
2210 * Usually, we want to just turn this into 2210 * Usually, we want to just turn this into
2211 * a negative dentry, but if anybody else is 2211 * a negative dentry, but if anybody else is
2212 * currently using the dentry or the inode 2212 * currently using the dentry or the inode
2213 * we can't do that and we fall back on removing 2213 * we can't do that and we fall back on removing
2214 * it from the hash queues and waiting for 2214 * it from the hash queues and waiting for
2215 * it to be deleted later when it has no users 2215 * it to be deleted later when it has no users
2216 */ 2216 */
2217 2217
2218 /** 2218 /**
2219 * d_delete - delete a dentry 2219 * d_delete - delete a dentry
2220 * @dentry: The dentry to delete 2220 * @dentry: The dentry to delete
2221 * 2221 *
2222 * Turn the dentry into a negative dentry if possible, otherwise 2222 * Turn the dentry into a negative dentry if possible, otherwise
2223 * remove it from the hash queues so it can be deleted later 2223 * remove it from the hash queues so it can be deleted later
2224 */ 2224 */
2225 2225
2226 void d_delete(struct dentry * dentry) 2226 void d_delete(struct dentry * dentry)
2227 { 2227 {
2228 struct inode *inode; 2228 struct inode *inode;
2229 int isdir = 0; 2229 int isdir = 0;
2230 /* 2230 /*
2231 * Are we the only user? 2231 * Are we the only user?
2232 */ 2232 */
2233 again: 2233 again:
2234 spin_lock(&dentry->d_lock); 2234 spin_lock(&dentry->d_lock);
2235 inode = dentry->d_inode; 2235 inode = dentry->d_inode;
2236 isdir = S_ISDIR(inode->i_mode); 2236 isdir = S_ISDIR(inode->i_mode);
2237 if (dentry->d_lockref.count == 1) { 2237 if (dentry->d_lockref.count == 1) {
2238 if (!spin_trylock(&inode->i_lock)) { 2238 if (!spin_trylock(&inode->i_lock)) {
2239 spin_unlock(&dentry->d_lock); 2239 spin_unlock(&dentry->d_lock);
2240 cpu_relax(); 2240 cpu_relax();
2241 goto again; 2241 goto again;
2242 } 2242 }
2243 dentry->d_flags &= ~DCACHE_CANT_MOUNT; 2243 dentry->d_flags &= ~DCACHE_CANT_MOUNT;
2244 dentry_unlink_inode(dentry); 2244 dentry_unlink_inode(dentry);
2245 fsnotify_nameremove(dentry, isdir); 2245 fsnotify_nameremove(dentry, isdir);
2246 return; 2246 return;
2247 } 2247 }
2248 2248
2249 if (!d_unhashed(dentry)) 2249 if (!d_unhashed(dentry))
2250 __d_drop(dentry); 2250 __d_drop(dentry);
2251 2251
2252 spin_unlock(&dentry->d_lock); 2252 spin_unlock(&dentry->d_lock);
2253 2253
2254 fsnotify_nameremove(dentry, isdir); 2254 fsnotify_nameremove(dentry, isdir);
2255 } 2255 }
2256 EXPORT_SYMBOL(d_delete); 2256 EXPORT_SYMBOL(d_delete);
2257 2257
2258 static void __d_rehash(struct dentry * entry, struct hlist_bl_head *b) 2258 static void __d_rehash(struct dentry * entry, struct hlist_bl_head *b)
2259 { 2259 {
2260 BUG_ON(!d_unhashed(entry)); 2260 BUG_ON(!d_unhashed(entry));
2261 hlist_bl_lock(b); 2261 hlist_bl_lock(b);
2262 entry->d_flags |= DCACHE_RCUACCESS; 2262 entry->d_flags |= DCACHE_RCUACCESS;
2263 hlist_bl_add_head_rcu(&entry->d_hash, b); 2263 hlist_bl_add_head_rcu(&entry->d_hash, b);
2264 hlist_bl_unlock(b); 2264 hlist_bl_unlock(b);
2265 } 2265 }
2266 2266
2267 static void _d_rehash(struct dentry * entry) 2267 static void _d_rehash(struct dentry * entry)
2268 { 2268 {
2269 __d_rehash(entry, d_hash(entry->d_parent, entry->d_name.hash)); 2269 __d_rehash(entry, d_hash(entry->d_parent, entry->d_name.hash));
2270 } 2270 }
2271 2271
2272 /** 2272 /**
2273 * d_rehash - add an entry back to the hash 2273 * d_rehash - add an entry back to the hash
2274 * @entry: dentry to add to the hash 2274 * @entry: dentry to add to the hash
2275 * 2275 *
2276 * Adds a dentry to the hash according to its name. 2276 * Adds a dentry to the hash according to its name.
2277 */ 2277 */
2278 2278
2279 void d_rehash(struct dentry * entry) 2279 void d_rehash(struct dentry * entry)
2280 { 2280 {
2281 spin_lock(&entry->d_lock); 2281 spin_lock(&entry->d_lock);
2282 _d_rehash(entry); 2282 _d_rehash(entry);
2283 spin_unlock(&entry->d_lock); 2283 spin_unlock(&entry->d_lock);
2284 } 2284 }
2285 EXPORT_SYMBOL(d_rehash); 2285 EXPORT_SYMBOL(d_rehash);
2286 2286
2287 /** 2287 /**
2288 * dentry_update_name_case - update case insensitive dentry with a new name 2288 * dentry_update_name_case - update case insensitive dentry with a new name
2289 * @dentry: dentry to be updated 2289 * @dentry: dentry to be updated
2290 * @name: new name 2290 * @name: new name
2291 * 2291 *
2292 * Update a case insensitive dentry with new case of name. 2292 * Update a case insensitive dentry with new case of name.
2293 * 2293 *
2294 * dentry must have been returned by d_lookup with name @name. Old and new 2294 * dentry must have been returned by d_lookup with name @name. Old and new
2295 * name lengths must match (ie. no d_compare which allows mismatched name 2295 * name lengths must match (ie. no d_compare which allows mismatched name
2296 * lengths). 2296 * lengths).
2297 * 2297 *
2298 * Parent inode i_mutex must be held over d_lookup and into this call (to 2298 * Parent inode i_mutex must be held over d_lookup and into this call (to
2299 * keep renames and concurrent inserts, and readdir(2) away). 2299 * keep renames and concurrent inserts, and readdir(2) away).
2300 */ 2300 */
2301 void dentry_update_name_case(struct dentry *dentry, struct qstr *name) 2301 void dentry_update_name_case(struct dentry *dentry, struct qstr *name)
2302 { 2302 {
2303 BUG_ON(!mutex_is_locked(&dentry->d_parent->d_inode->i_mutex)); 2303 BUG_ON(!mutex_is_locked(&dentry->d_parent->d_inode->i_mutex));
2304 BUG_ON(dentry->d_name.len != name->len); /* d_lookup gives this */ 2304 BUG_ON(dentry->d_name.len != name->len); /* d_lookup gives this */
2305 2305
2306 spin_lock(&dentry->d_lock); 2306 spin_lock(&dentry->d_lock);
2307 write_seqcount_begin(&dentry->d_seq); 2307 write_seqcount_begin(&dentry->d_seq);
2308 memcpy((unsigned char *)dentry->d_name.name, name->name, name->len); 2308 memcpy((unsigned char *)dentry->d_name.name, name->name, name->len);
2309 write_seqcount_end(&dentry->d_seq); 2309 write_seqcount_end(&dentry->d_seq);
2310 spin_unlock(&dentry->d_lock); 2310 spin_unlock(&dentry->d_lock);
2311 } 2311 }
2312 EXPORT_SYMBOL(dentry_update_name_case); 2312 EXPORT_SYMBOL(dentry_update_name_case);
2313 2313
2314 static void switch_names(struct dentry *dentry, struct dentry *target) 2314 static void switch_names(struct dentry *dentry, struct dentry *target)
2315 { 2315 {
2316 if (dname_external(target)) { 2316 if (dname_external(target)) {
2317 if (dname_external(dentry)) { 2317 if (dname_external(dentry)) {
2318 /* 2318 /*
2319 * Both external: swap the pointers 2319 * Both external: swap the pointers
2320 */ 2320 */
2321 swap(target->d_name.name, dentry->d_name.name); 2321 swap(target->d_name.name, dentry->d_name.name);
2322 } else { 2322 } else {
2323 /* 2323 /*
2324 * dentry:internal, target:external. Steal target's 2324 * dentry:internal, target:external. Steal target's
2325 * storage and make target internal. 2325 * storage and make target internal.
2326 */ 2326 */
2327 memcpy(target->d_iname, dentry->d_name.name, 2327 memcpy(target->d_iname, dentry->d_name.name,
2328 dentry->d_name.len + 1); 2328 dentry->d_name.len + 1);
2329 dentry->d_name.name = target->d_name.name; 2329 dentry->d_name.name = target->d_name.name;
2330 target->d_name.name = target->d_iname; 2330 target->d_name.name = target->d_iname;
2331 } 2331 }
2332 } else { 2332 } else {
2333 if (dname_external(dentry)) { 2333 if (dname_external(dentry)) {
2334 /* 2334 /*
2335 * dentry:external, target:internal. Give dentry's 2335 * dentry:external, target:internal. Give dentry's
2336 * storage to target and make dentry internal 2336 * storage to target and make dentry internal
2337 */ 2337 */
2338 memcpy(dentry->d_iname, target->d_name.name, 2338 memcpy(dentry->d_iname, target->d_name.name,
2339 target->d_name.len + 1); 2339 target->d_name.len + 1);
2340 target->d_name.name = dentry->d_name.name; 2340 target->d_name.name = dentry->d_name.name;
2341 dentry->d_name.name = dentry->d_iname; 2341 dentry->d_name.name = dentry->d_iname;
2342 } else { 2342 } else {
2343 /* 2343 /*
2344 * Both are internal. Just copy target to dentry 2344 * Both are internal. Just copy target to dentry
2345 */ 2345 */
2346 memcpy(dentry->d_iname, target->d_name.name, 2346 memcpy(dentry->d_iname, target->d_name.name,
2347 target->d_name.len + 1); 2347 target->d_name.len + 1);
2348 dentry->d_name.len = target->d_name.len; 2348 dentry->d_name.len = target->d_name.len;
2349 return; 2349 return;
2350 } 2350 }
2351 } 2351 }
2352 swap(dentry->d_name.len, target->d_name.len); 2352 swap(dentry->d_name.len, target->d_name.len);
2353 } 2353 }
2354 2354
2355 static void dentry_lock_for_move(struct dentry *dentry, struct dentry *target) 2355 static void dentry_lock_for_move(struct dentry *dentry, struct dentry *target)
2356 { 2356 {
2357 /* 2357 /*
2358 * XXXX: do we really need to take target->d_lock? 2358 * XXXX: do we really need to take target->d_lock?
2359 */ 2359 */
2360 if (IS_ROOT(dentry) || dentry->d_parent == target->d_parent) 2360 if (IS_ROOT(dentry) || dentry->d_parent == target->d_parent)
2361 spin_lock(&target->d_parent->d_lock); 2361 spin_lock(&target->d_parent->d_lock);
2362 else { 2362 else {
2363 if (d_ancestor(dentry->d_parent, target->d_parent)) { 2363 if (d_ancestor(dentry->d_parent, target->d_parent)) {
2364 spin_lock(&dentry->d_parent->d_lock); 2364 spin_lock(&dentry->d_parent->d_lock);
2365 spin_lock_nested(&target->d_parent->d_lock, 2365 spin_lock_nested(&target->d_parent->d_lock,
2366 DENTRY_D_LOCK_NESTED); 2366 DENTRY_D_LOCK_NESTED);
2367 } else { 2367 } else {
2368 spin_lock(&target->d_parent->d_lock); 2368 spin_lock(&target->d_parent->d_lock);
2369 spin_lock_nested(&dentry->d_parent->d_lock, 2369 spin_lock_nested(&dentry->d_parent->d_lock,
2370 DENTRY_D_LOCK_NESTED); 2370 DENTRY_D_LOCK_NESTED);
2371 } 2371 }
2372 } 2372 }
2373 if (target < dentry) { 2373 if (target < dentry) {
2374 spin_lock_nested(&target->d_lock, 2); 2374 spin_lock_nested(&target->d_lock, 2);
2375 spin_lock_nested(&dentry->d_lock, 3); 2375 spin_lock_nested(&dentry->d_lock, 3);
2376 } else { 2376 } else {
2377 spin_lock_nested(&dentry->d_lock, 2); 2377 spin_lock_nested(&dentry->d_lock, 2);
2378 spin_lock_nested(&target->d_lock, 3); 2378 spin_lock_nested(&target->d_lock, 3);
2379 } 2379 }
2380 } 2380 }
2381 2381
2382 static void dentry_unlock_parents_for_move(struct dentry *dentry, 2382 static void dentry_unlock_parents_for_move(struct dentry *dentry,
2383 struct dentry *target) 2383 struct dentry *target)
2384 { 2384 {
2385 if (target->d_parent != dentry->d_parent) 2385 if (target->d_parent != dentry->d_parent)
2386 spin_unlock(&dentry->d_parent->d_lock); 2386 spin_unlock(&dentry->d_parent->d_lock);
2387 if (target->d_parent != target) 2387 if (target->d_parent != target)
2388 spin_unlock(&target->d_parent->d_lock); 2388 spin_unlock(&target->d_parent->d_lock);
2389 } 2389 }
2390 2390
2391 /* 2391 /*
2392 * When switching names, the actual string doesn't strictly have to 2392 * When switching names, the actual string doesn't strictly have to
2393 * be preserved in the target - because we're dropping the target 2393 * be preserved in the target - because we're dropping the target
2394 * anyway. As such, we can just do a simple memcpy() to copy over 2394 * anyway. As such, we can just do a simple memcpy() to copy over
2395 * the new name before we switch. 2395 * the new name before we switch.
2396 * 2396 *
2397 * Note that we have to be a lot more careful about getting the hash 2397 * Note that we have to be a lot more careful about getting the hash
2398 * switched - we have to switch the hash value properly even if it 2398 * switched - we have to switch the hash value properly even if it
2399 * then no longer matches the actual (corrupted) string of the target. 2399 * then no longer matches the actual (corrupted) string of the target.
2400 * The hash value has to match the hash queue that the dentry is on.. 2400 * The hash value has to match the hash queue that the dentry is on..
2401 */ 2401 */
2402 /* 2402 /*
2403 * __d_move - move a dentry 2403 * __d_move - move a dentry
2404 * @dentry: entry to move 2404 * @dentry: entry to move
2405 * @target: new dentry 2405 * @target: new dentry
2406 * 2406 *
2407 * Update the dcache to reflect the move of a file name. Negative 2407 * Update the dcache to reflect the move of a file name. Negative
2408 * dcache entries should not be moved in this way. Caller must hold 2408 * dcache entries should not be moved in this way. Caller must hold
2409 * rename_lock, the i_mutex of the source and target directories, 2409 * rename_lock, the i_mutex of the source and target directories,
2410 * and the sb->s_vfs_rename_mutex if they differ. See lock_rename(). 2410 * and the sb->s_vfs_rename_mutex if they differ. See lock_rename().
2411 */ 2411 */
2412 static void __d_move(struct dentry * dentry, struct dentry * target) 2412 static void __d_move(struct dentry * dentry, struct dentry * target)
2413 { 2413 {
2414 if (!dentry->d_inode) 2414 if (!dentry->d_inode)
2415 printk(KERN_WARNING "VFS: moving negative dcache entry\n"); 2415 printk(KERN_WARNING "VFS: moving negative dcache entry\n");
2416 2416
2417 BUG_ON(d_ancestor(dentry, target)); 2417 BUG_ON(d_ancestor(dentry, target));
2418 BUG_ON(d_ancestor(target, dentry)); 2418 BUG_ON(d_ancestor(target, dentry));
2419 2419
2420 dentry_lock_for_move(dentry, target); 2420 dentry_lock_for_move(dentry, target);
2421 2421
2422 write_seqcount_begin(&dentry->d_seq); 2422 write_seqcount_begin(&dentry->d_seq);
2423 write_seqcount_begin(&target->d_seq); 2423 write_seqcount_begin(&target->d_seq);
2424 2424
2425 /* __d_drop does write_seqcount_barrier, but they're OK to nest. */ 2425 /* __d_drop does write_seqcount_barrier, but they're OK to nest. */
2426 2426
2427 /* 2427 /*
2428 * Move the dentry to the target hash queue. Don't bother checking 2428 * Move the dentry to the target hash queue. Don't bother checking
2429 * for the same hash queue because of how unlikely it is. 2429 * for the same hash queue because of how unlikely it is.
2430 */ 2430 */
2431 __d_drop(dentry); 2431 __d_drop(dentry);
2432 __d_rehash(dentry, d_hash(target->d_parent, target->d_name.hash)); 2432 __d_rehash(dentry, d_hash(target->d_parent, target->d_name.hash));
2433 2433
2434 /* Unhash the target: dput() will then get rid of it */ 2434 /* Unhash the target: dput() will then get rid of it */
2435 __d_drop(target); 2435 __d_drop(target);
2436 2436
2437 list_del(&dentry->d_u.d_child); 2437 list_del(&dentry->d_u.d_child);
2438 list_del(&target->d_u.d_child); 2438 list_del(&target->d_u.d_child);
2439 2439
2440 /* Switch the names.. */ 2440 /* Switch the names.. */
2441 switch_names(dentry, target); 2441 switch_names(dentry, target);
2442 swap(dentry->d_name.hash, target->d_name.hash); 2442 swap(dentry->d_name.hash, target->d_name.hash);
2443 2443
2444 /* ... and switch the parents */ 2444 /* ... and switch the parents */
2445 if (IS_ROOT(dentry)) { 2445 if (IS_ROOT(dentry)) {
2446 dentry->d_parent = target->d_parent; 2446 dentry->d_parent = target->d_parent;
2447 target->d_parent = target; 2447 target->d_parent = target;
2448 INIT_LIST_HEAD(&target->d_u.d_child); 2448 INIT_LIST_HEAD(&target->d_u.d_child);
2449 } else { 2449 } else {
2450 swap(dentry->d_parent, target->d_parent); 2450 swap(dentry->d_parent, target->d_parent);
2451 2451
2452 /* And add them back to the (new) parent lists */ 2452 /* And add them back to the (new) parent lists */
2453 list_add(&target->d_u.d_child, &target->d_parent->d_subdirs); 2453 list_add(&target->d_u.d_child, &target->d_parent->d_subdirs);
2454 } 2454 }
2455 2455
2456 list_add(&dentry->d_u.d_child, &dentry->d_parent->d_subdirs); 2456 list_add(&dentry->d_u.d_child, &dentry->d_parent->d_subdirs);
2457 2457
2458 write_seqcount_end(&target->d_seq); 2458 write_seqcount_end(&target->d_seq);
2459 write_seqcount_end(&dentry->d_seq); 2459 write_seqcount_end(&dentry->d_seq);
2460 2460
2461 dentry_unlock_parents_for_move(dentry, target); 2461 dentry_unlock_parents_for_move(dentry, target);
2462 spin_unlock(&target->d_lock); 2462 spin_unlock(&target->d_lock);
2463 fsnotify_d_move(dentry); 2463 fsnotify_d_move(dentry);
2464 spin_unlock(&dentry->d_lock); 2464 spin_unlock(&dentry->d_lock);
2465 } 2465 }
2466 2466
2467 /* 2467 /*
2468 * d_move - move a dentry 2468 * d_move - move a dentry
2469 * @dentry: entry to move 2469 * @dentry: entry to move
2470 * @target: new dentry 2470 * @target: new dentry
2471 * 2471 *
2472 * Update the dcache to reflect the move of a file name. Negative 2472 * Update the dcache to reflect the move of a file name. Negative
2473 * dcache entries should not be moved in this way. See the locking 2473 * dcache entries should not be moved in this way. See the locking
2474 * requirements for __d_move. 2474 * requirements for __d_move.
2475 */ 2475 */
2476 void d_move(struct dentry *dentry, struct dentry *target) 2476 void d_move(struct dentry *dentry, struct dentry *target)
2477 { 2477 {
2478 write_seqlock(&rename_lock); 2478 write_seqlock(&rename_lock);
2479 __d_move(dentry, target); 2479 __d_move(dentry, target);
2480 write_sequnlock(&rename_lock); 2480 write_sequnlock(&rename_lock);
2481 } 2481 }
2482 EXPORT_SYMBOL(d_move); 2482 EXPORT_SYMBOL(d_move);
2483 2483
2484 /** 2484 /**
2485 * d_ancestor - search for an ancestor 2485 * d_ancestor - search for an ancestor
2486 * @p1: ancestor dentry 2486 * @p1: ancestor dentry
2487 * @p2: child dentry 2487 * @p2: child dentry
2488 * 2488 *
2489 * Returns the ancestor dentry of p2 which is a child of p1, if p1 is 2489 * Returns the ancestor dentry of p2 which is a child of p1, if p1 is
2490 * an ancestor of p2, else NULL. 2490 * an ancestor of p2, else NULL.
2491 */ 2491 */
2492 struct dentry *d_ancestor(struct dentry *p1, struct dentry *p2) 2492 struct dentry *d_ancestor(struct dentry *p1, struct dentry *p2)
2493 { 2493 {
2494 struct dentry *p; 2494 struct dentry *p;
2495 2495
2496 for (p = p2; !IS_ROOT(p); p = p->d_parent) { 2496 for (p = p2; !IS_ROOT(p); p = p->d_parent) {
2497 if (p->d_parent == p1) 2497 if (p->d_parent == p1)
2498 return p; 2498 return p;
2499 } 2499 }
2500 return NULL; 2500 return NULL;
2501 } 2501 }
2502 2502
2503 /* 2503 /*
2504 * This helper attempts to cope with remotely renamed directories 2504 * This helper attempts to cope with remotely renamed directories
2505 * 2505 *
2506 * It assumes that the caller is already holding 2506 * It assumes that the caller is already holding
2507 * dentry->d_parent->d_inode->i_mutex, inode->i_lock and rename_lock 2507 * dentry->d_parent->d_inode->i_mutex, inode->i_lock and rename_lock
2508 * 2508 *
2509 * Note: If ever the locking in lock_rename() changes, then please 2509 * Note: If ever the locking in lock_rename() changes, then please
2510 * remember to update this too... 2510 * remember to update this too...
2511 */ 2511 */
2512 static struct dentry *__d_unalias(struct inode *inode, 2512 static struct dentry *__d_unalias(struct inode *inode,
2513 struct dentry *dentry, struct dentry *alias) 2513 struct dentry *dentry, struct dentry *alias)
2514 { 2514 {
2515 struct mutex *m1 = NULL, *m2 = NULL; 2515 struct mutex *m1 = NULL, *m2 = NULL;
2516 struct dentry *ret = ERR_PTR(-EBUSY); 2516 struct dentry *ret = ERR_PTR(-EBUSY);
2517 2517
2518 /* If alias and dentry share a parent, then no extra locks required */ 2518 /* If alias and dentry share a parent, then no extra locks required */
2519 if (alias->d_parent == dentry->d_parent) 2519 if (alias->d_parent == dentry->d_parent)
2520 goto out_unalias; 2520 goto out_unalias;
2521 2521
2522 /* See lock_rename() */ 2522 /* See lock_rename() */
2523 if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex)) 2523 if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex))
2524 goto out_err; 2524 goto out_err;
2525 m1 = &dentry->d_sb->s_vfs_rename_mutex; 2525 m1 = &dentry->d_sb->s_vfs_rename_mutex;
2526 if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex)) 2526 if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex))
2527 goto out_err; 2527 goto out_err;
2528 m2 = &alias->d_parent->d_inode->i_mutex; 2528 m2 = &alias->d_parent->d_inode->i_mutex;
2529 out_unalias: 2529 out_unalias:
2530 if (likely(!d_mountpoint(alias))) { 2530 if (likely(!d_mountpoint(alias))) {
2531 __d_move(alias, dentry); 2531 __d_move(alias, dentry);
2532 ret = alias; 2532 ret = alias;
2533 } 2533 }
2534 out_err: 2534 out_err:
2535 spin_unlock(&inode->i_lock); 2535 spin_unlock(&inode->i_lock);
2536 if (m2) 2536 if (m2)
2537 mutex_unlock(m2); 2537 mutex_unlock(m2);
2538 if (m1) 2538 if (m1)
2539 mutex_unlock(m1); 2539 mutex_unlock(m1);
2540 return ret; 2540 return ret;
2541 } 2541 }
2542 2542
2543 /* 2543 /*
2544 * Prepare an anonymous dentry for life in the superblock's dentry tree as a 2544 * Prepare an anonymous dentry for life in the superblock's dentry tree as a
2545 * named dentry in place of the dentry to be replaced. 2545 * named dentry in place of the dentry to be replaced.
2546 * returns with anon->d_lock held! 2546 * returns with anon->d_lock held!
2547 */ 2547 */
2548 static void __d_materialise_dentry(struct dentry *dentry, struct dentry *anon) 2548 static void __d_materialise_dentry(struct dentry *dentry, struct dentry *anon)
2549 { 2549 {
2550 struct dentry *dparent; 2550 struct dentry *dparent;
2551 2551
2552 dentry_lock_for_move(anon, dentry); 2552 dentry_lock_for_move(anon, dentry);
2553 2553
2554 write_seqcount_begin(&dentry->d_seq); 2554 write_seqcount_begin(&dentry->d_seq);
2555 write_seqcount_begin(&anon->d_seq); 2555 write_seqcount_begin(&anon->d_seq);
2556 2556
2557 dparent = dentry->d_parent; 2557 dparent = dentry->d_parent;
2558 2558
2559 switch_names(dentry, anon); 2559 switch_names(dentry, anon);
2560 swap(dentry->d_name.hash, anon->d_name.hash); 2560 swap(dentry->d_name.hash, anon->d_name.hash);
2561 2561
2562 dentry->d_parent = dentry; 2562 dentry->d_parent = dentry;
2563 list_del_init(&dentry->d_u.d_child); 2563 list_del_init(&dentry->d_u.d_child);
2564 anon->d_parent = dparent; 2564 anon->d_parent = dparent;
2565 list_move(&anon->d_u.d_child, &dparent->d_subdirs); 2565 list_move(&anon->d_u.d_child, &dparent->d_subdirs);
2566 2566
2567 write_seqcount_end(&dentry->d_seq); 2567 write_seqcount_end(&dentry->d_seq);
2568 write_seqcount_end(&anon->d_seq); 2568 write_seqcount_end(&anon->d_seq);
2569 2569
2570 dentry_unlock_parents_for_move(anon, dentry); 2570 dentry_unlock_parents_for_move(anon, dentry);
2571 spin_unlock(&dentry->d_lock); 2571 spin_unlock(&dentry->d_lock);
2572 2572
2573 /* anon->d_lock still locked, returns locked */ 2573 /* anon->d_lock still locked, returns locked */
2574 anon->d_flags &= ~DCACHE_DISCONNECTED; 2574 anon->d_flags &= ~DCACHE_DISCONNECTED;
2575 } 2575 }
2576 2576
2577 /** 2577 /**
2578 * d_materialise_unique - introduce an inode into the tree 2578 * d_materialise_unique - introduce an inode into the tree
2579 * @dentry: candidate dentry 2579 * @dentry: candidate dentry
2580 * @inode: inode to bind to the dentry, to which aliases may be attached 2580 * @inode: inode to bind to the dentry, to which aliases may be attached
2581 * 2581 *
2582 * Introduces an dentry into the tree, substituting an extant disconnected 2582 * Introduces an dentry into the tree, substituting an extant disconnected
2583 * root directory alias in its place if there is one. Caller must hold the 2583 * root directory alias in its place if there is one. Caller must hold the
2584 * i_mutex of the parent directory. 2584 * i_mutex of the parent directory.
2585 */ 2585 */
2586 struct dentry *d_materialise_unique(struct dentry *dentry, struct inode *inode) 2586 struct dentry *d_materialise_unique(struct dentry *dentry, struct inode *inode)
2587 { 2587 {
2588 struct dentry *actual; 2588 struct dentry *actual;
2589 2589
2590 BUG_ON(!d_unhashed(dentry)); 2590 BUG_ON(!d_unhashed(dentry));
2591 2591
2592 if (!inode) { 2592 if (!inode) {
2593 actual = dentry; 2593 actual = dentry;
2594 __d_instantiate(dentry, NULL); 2594 __d_instantiate(dentry, NULL);
2595 d_rehash(actual); 2595 d_rehash(actual);
2596 goto out_nolock; 2596 goto out_nolock;
2597 } 2597 }
2598 2598
2599 spin_lock(&inode->i_lock); 2599 spin_lock(&inode->i_lock);
2600 2600
2601 if (S_ISDIR(inode->i_mode)) { 2601 if (S_ISDIR(inode->i_mode)) {
2602 struct dentry *alias; 2602 struct dentry *alias;
2603 2603
2604 /* Does an aliased dentry already exist? */ 2604 /* Does an aliased dentry already exist? */
2605 alias = __d_find_alias(inode, 0); 2605 alias = __d_find_alias(inode, 0);
2606 if (alias) { 2606 if (alias) {
2607 actual = alias; 2607 actual = alias;
2608 write_seqlock(&rename_lock); 2608 write_seqlock(&rename_lock);
2609 2609
2610 if (d_ancestor(alias, dentry)) { 2610 if (d_ancestor(alias, dentry)) {
2611 /* Check for loops */ 2611 /* Check for loops */
2612 actual = ERR_PTR(-ELOOP); 2612 actual = ERR_PTR(-ELOOP);
2613 spin_unlock(&inode->i_lock); 2613 spin_unlock(&inode->i_lock);
2614 } else if (IS_ROOT(alias)) { 2614 } else if (IS_ROOT(alias)) {
2615 /* Is this an anonymous mountpoint that we 2615 /* Is this an anonymous mountpoint that we
2616 * could splice into our tree? */ 2616 * could splice into our tree? */
2617 __d_materialise_dentry(dentry, alias); 2617 __d_materialise_dentry(dentry, alias);
2618 write_sequnlock(&rename_lock); 2618 write_sequnlock(&rename_lock);
2619 __d_drop(alias); 2619 __d_drop(alias);
2620 goto found; 2620 goto found;
2621 } else { 2621 } else {
2622 /* Nope, but we must(!) avoid directory 2622 /* Nope, but we must(!) avoid directory
2623 * aliasing. This drops inode->i_lock */ 2623 * aliasing. This drops inode->i_lock */
2624 actual = __d_unalias(inode, dentry, alias); 2624 actual = __d_unalias(inode, dentry, alias);
2625 } 2625 }
2626 write_sequnlock(&rename_lock); 2626 write_sequnlock(&rename_lock);
2627 if (IS_ERR(actual)) { 2627 if (IS_ERR(actual)) {
2628 if (PTR_ERR(actual) == -ELOOP) 2628 if (PTR_ERR(actual) == -ELOOP)
2629 pr_warn_ratelimited( 2629 pr_warn_ratelimited(
2630 "VFS: Lookup of '%s' in %s %s" 2630 "VFS: Lookup of '%s' in %s %s"
2631 " would have caused loop\n", 2631 " would have caused loop\n",
2632 dentry->d_name.name, 2632 dentry->d_name.name,
2633 inode->i_sb->s_type->name, 2633 inode->i_sb->s_type->name,
2634 inode->i_sb->s_id); 2634 inode->i_sb->s_id);
2635 dput(alias); 2635 dput(alias);
2636 } 2636 }
2637 goto out_nolock; 2637 goto out_nolock;
2638 } 2638 }
2639 } 2639 }
2640 2640
2641 /* Add a unique reference */ 2641 /* Add a unique reference */
2642 actual = __d_instantiate_unique(dentry, inode); 2642 actual = __d_instantiate_unique(dentry, inode);
2643 if (!actual) 2643 if (!actual)
2644 actual = dentry; 2644 actual = dentry;
2645 else 2645 else
2646 BUG_ON(!d_unhashed(actual)); 2646 BUG_ON(!d_unhashed(actual));
2647 2647
2648 spin_lock(&actual->d_lock); 2648 spin_lock(&actual->d_lock);
2649 found: 2649 found:
2650 _d_rehash(actual); 2650 _d_rehash(actual);
2651 spin_unlock(&actual->d_lock); 2651 spin_unlock(&actual->d_lock);
2652 spin_unlock(&inode->i_lock); 2652 spin_unlock(&inode->i_lock);
2653 out_nolock: 2653 out_nolock:
2654 if (actual == dentry) { 2654 if (actual == dentry) {
2655 security_d_instantiate(dentry, inode); 2655 security_d_instantiate(dentry, inode);
2656 return NULL; 2656 return NULL;
2657 } 2657 }
2658 2658
2659 iput(inode); 2659 iput(inode);
2660 return actual; 2660 return actual;
2661 } 2661 }
2662 EXPORT_SYMBOL_GPL(d_materialise_unique); 2662 EXPORT_SYMBOL_GPL(d_materialise_unique);
2663 2663
2664 static int prepend(char **buffer, int *buflen, const char *str, int namelen) 2664 static int prepend(char **buffer, int *buflen, const char *str, int namelen)
2665 { 2665 {
2666 *buflen -= namelen; 2666 *buflen -= namelen;
2667 if (*buflen < 0) 2667 if (*buflen < 0)
2668 return -ENAMETOOLONG; 2668 return -ENAMETOOLONG;
2669 *buffer -= namelen; 2669 *buffer -= namelen;
2670 memcpy(*buffer, str, namelen); 2670 memcpy(*buffer, str, namelen);
2671 return 0; 2671 return 0;
2672 } 2672 }
2673 2673
2674 /** 2674 /**
2675 * prepend_name - prepend a pathname in front of current buffer pointer 2675 * prepend_name - prepend a pathname in front of current buffer pointer
2676 * @buffer: buffer pointer 2676 * @buffer: buffer pointer
2677 * @buflen: allocated length of the buffer 2677 * @buflen: allocated length of the buffer
2678 * @name: name string and length qstr structure 2678 * @name: name string and length qstr structure
2679 * 2679 *
2680 * With RCU path tracing, it may race with d_move(). Use ACCESS_ONCE() to 2680 * With RCU path tracing, it may race with d_move(). Use ACCESS_ONCE() to
2681 * make sure that either the old or the new name pointer and length are 2681 * make sure that either the old or the new name pointer and length are
2682 * fetched. However, there may be mismatch between length and pointer. 2682 * fetched. However, there may be mismatch between length and pointer.
2683 * The length cannot be trusted, we need to copy it byte-by-byte until 2683 * The length cannot be trusted, we need to copy it byte-by-byte until
2684 * the length is reached or a null byte is found. It also prepends "/" at 2684 * the length is reached or a null byte is found. It also prepends "/" at
2685 * the beginning of the name. The sequence number check at the caller will 2685 * the beginning of the name. The sequence number check at the caller will
2686 * retry it again when a d_move() does happen. So any garbage in the buffer 2686 * retry it again when a d_move() does happen. So any garbage in the buffer
2687 * due to mismatched pointer and length will be discarded. 2687 * due to mismatched pointer and length will be discarded.
2688 */ 2688 */
2689 static int prepend_name(char **buffer, int *buflen, struct qstr *name) 2689 static int prepend_name(char **buffer, int *buflen, struct qstr *name)
2690 { 2690 {
2691 const char *dname = ACCESS_ONCE(name->name); 2691 const char *dname = ACCESS_ONCE(name->name);
2692 u32 dlen = ACCESS_ONCE(name->len); 2692 u32 dlen = ACCESS_ONCE(name->len);
2693 char *p; 2693 char *p;
2694 2694
2695 if (*buflen < dlen + 1) 2695 if (*buflen < dlen + 1)
2696 return -ENAMETOOLONG; 2696 return -ENAMETOOLONG;
2697 *buflen -= dlen + 1; 2697 *buflen -= dlen + 1;
2698 p = *buffer -= dlen + 1; 2698 p = *buffer -= dlen + 1;
2699 *p++ = '/'; 2699 *p++ = '/';
2700 while (dlen--) { 2700 while (dlen--) {
2701 char c = *dname++; 2701 char c = *dname++;
2702 if (!c) 2702 if (!c)
2703 break; 2703 break;
2704 *p++ = c; 2704 *p++ = c;
2705 } 2705 }
2706 return 0; 2706 return 0;
2707 } 2707 }
2708 2708
2709 /** 2709 /**
2710 * prepend_path - Prepend path string to a buffer 2710 * prepend_path - Prepend path string to a buffer
2711 * @path: the dentry/vfsmount to report 2711 * @path: the dentry/vfsmount to report
2712 * @root: root vfsmnt/dentry 2712 * @root: root vfsmnt/dentry
2713 * @buffer: pointer to the end of the buffer 2713 * @buffer: pointer to the end of the buffer
2714 * @buflen: pointer to buffer length 2714 * @buflen: pointer to buffer length
2715 * 2715 *
2716 * The function will first try to write out the pathname without taking any 2716 * The function will first try to write out the pathname without taking any
2717 * lock other than the RCU read lock to make sure that dentries won't go away. 2717 * lock other than the RCU read lock to make sure that dentries won't go away.
2718 * It only checks the sequence number of the global rename_lock as any change 2718 * It only checks the sequence number of the global rename_lock as any change
2719 * in the dentry's d_seq will be preceded by changes in the rename_lock 2719 * in the dentry's d_seq will be preceded by changes in the rename_lock
2720 * sequence number. If the sequence number had been changed, it will restart 2720 * sequence number. If the sequence number had been changed, it will restart
2721 * the whole pathname back-tracing sequence again by taking the rename_lock. 2721 * the whole pathname back-tracing sequence again by taking the rename_lock.
2722 * In this case, there is no need to take the RCU read lock as the recursive 2722 * In this case, there is no need to take the RCU read lock as the recursive
2723 * parent pointer references will keep the dentry chain alive as long as no 2723 * parent pointer references will keep the dentry chain alive as long as no
2724 * rename operation is performed. 2724 * rename operation is performed.
2725 */ 2725 */
2726 static int prepend_path(const struct path *path, 2726 static int prepend_path(const struct path *path,
2727 const struct path *root, 2727 const struct path *root,
2728 char **buffer, int *buflen) 2728 char **buffer, int *buflen)
2729 { 2729 {
2730 struct dentry *dentry = path->dentry; 2730 struct dentry *dentry = path->dentry;
2731 struct vfsmount *vfsmnt = path->mnt; 2731 struct vfsmount *vfsmnt = path->mnt;
2732 struct mount *mnt = real_mount(vfsmnt); 2732 struct mount *mnt = real_mount(vfsmnt);
2733 int error = 0; 2733 int error = 0;
2734 unsigned seq = 0; 2734 unsigned seq = 0;
2735 char *bptr; 2735 char *bptr;
2736 int blen; 2736 int blen;
2737 2737
2738 rcu_read_lock(); 2738 rcu_read_lock();
2739 restart: 2739 restart:
2740 bptr = *buffer; 2740 bptr = *buffer;
2741 blen = *buflen; 2741 blen = *buflen;
2742 read_seqbegin_or_lock(&rename_lock, &seq); 2742 read_seqbegin_or_lock(&rename_lock, &seq);
2743 while (dentry != root->dentry || vfsmnt != root->mnt) { 2743 while (dentry != root->dentry || vfsmnt != root->mnt) {
2744 struct dentry * parent; 2744 struct dentry * parent;
2745 2745
2746 if (dentry == vfsmnt->mnt_root || IS_ROOT(dentry)) { 2746 if (dentry == vfsmnt->mnt_root || IS_ROOT(dentry)) {
2747 /* Global root? */ 2747 /* Global root? */
2748 if (mnt_has_parent(mnt)) { 2748 if (mnt_has_parent(mnt)) {
2749 dentry = mnt->mnt_mountpoint; 2749 dentry = mnt->mnt_mountpoint;
2750 mnt = mnt->mnt_parent; 2750 mnt = mnt->mnt_parent;
2751 vfsmnt = &mnt->mnt; 2751 vfsmnt = &mnt->mnt;
2752 continue; 2752 continue;
2753 } 2753 }
2754 /* 2754 /*
2755 * Filesystems needing to implement special "root names" 2755 * Filesystems needing to implement special "root names"
2756 * should do so with ->d_dname() 2756 * should do so with ->d_dname()
2757 */ 2757 */
2758 if (IS_ROOT(dentry) && 2758 if (IS_ROOT(dentry) &&
2759 (dentry->d_name.len != 1 || 2759 (dentry->d_name.len != 1 ||
2760 dentry->d_name.name[0] != '/')) { 2760 dentry->d_name.name[0] != '/')) {
2761 WARN(1, "Root dentry has weird name <%.*s>\n", 2761 WARN(1, "Root dentry has weird name <%.*s>\n",
2762 (int) dentry->d_name.len, 2762 (int) dentry->d_name.len,
2763 dentry->d_name.name); 2763 dentry->d_name.name);
2764 } 2764 }
2765 if (!error) 2765 if (!error)
2766 error = is_mounted(vfsmnt) ? 1 : 2; 2766 error = is_mounted(vfsmnt) ? 1 : 2;
2767 break; 2767 break;
2768 } 2768 }
2769 parent = dentry->d_parent; 2769 parent = dentry->d_parent;
2770 prefetch(parent); 2770 prefetch(parent);
2771 error = prepend_name(&bptr, &blen, &dentry->d_name); 2771 error = prepend_name(&bptr, &blen, &dentry->d_name);
2772 if (error) 2772 if (error)
2773 break; 2773 break;
2774 2774
2775 dentry = parent; 2775 dentry = parent;
2776 } 2776 }
2777 if (!(seq & 1)) 2777 if (!(seq & 1))
2778 rcu_read_unlock(); 2778 rcu_read_unlock();
2779 if (need_seqretry(&rename_lock, seq)) { 2779 if (need_seqretry(&rename_lock, seq)) {
2780 seq = 1; 2780 seq = 1;
2781 goto restart; 2781 goto restart;
2782 } 2782 }
2783 done_seqretry(&rename_lock, seq); 2783 done_seqretry(&rename_lock, seq);
2784 2784
2785 if (error >= 0 && bptr == *buffer) { 2785 if (error >= 0 && bptr == *buffer) {
2786 if (--blen < 0) 2786 if (--blen < 0)
2787 error = -ENAMETOOLONG; 2787 error = -ENAMETOOLONG;
2788 else 2788 else
2789 *--bptr = '/'; 2789 *--bptr = '/';
2790 } 2790 }
2791 *buffer = bptr; 2791 *buffer = bptr;
2792 *buflen = blen; 2792 *buflen = blen;
2793 return error; 2793 return error;
2794 } 2794 }
2795 2795
2796 /** 2796 /**
2797 * __d_path - return the path of a dentry 2797 * __d_path - return the path of a dentry
2798 * @path: the dentry/vfsmount to report 2798 * @path: the dentry/vfsmount to report
2799 * @root: root vfsmnt/dentry 2799 * @root: root vfsmnt/dentry
2800 * @buf: buffer to return value in 2800 * @buf: buffer to return value in
2801 * @buflen: buffer length 2801 * @buflen: buffer length
2802 * 2802 *
2803 * Convert a dentry into an ASCII path name. 2803 * Convert a dentry into an ASCII path name.
2804 * 2804 *
2805 * Returns a pointer into the buffer or an error code if the 2805 * Returns a pointer into the buffer or an error code if the
2806 * path was too long. 2806 * path was too long.
2807 * 2807 *
2808 * "buflen" should be positive. 2808 * "buflen" should be positive.
2809 * 2809 *
2810 * If the path is not reachable from the supplied root, return %NULL. 2810 * If the path is not reachable from the supplied root, return %NULL.
2811 */ 2811 */
2812 char *__d_path(const struct path *path, 2812 char *__d_path(const struct path *path,
2813 const struct path *root, 2813 const struct path *root,
2814 char *buf, int buflen) 2814 char *buf, int buflen)
2815 { 2815 {
2816 char *res = buf + buflen; 2816 char *res = buf + buflen;
2817 int error; 2817 int error;
2818 2818
2819 prepend(&res, &buflen, "\0", 1); 2819 prepend(&res, &buflen, "\0", 1);
2820 br_read_lock(&vfsmount_lock); 2820 br_read_lock(&vfsmount_lock);
2821 error = prepend_path(path, root, &res, &buflen); 2821 error = prepend_path(path, root, &res, &buflen);
2822 br_read_unlock(&vfsmount_lock); 2822 br_read_unlock(&vfsmount_lock);
2823 2823
2824 if (error < 0) 2824 if (error < 0)
2825 return ERR_PTR(error); 2825 return ERR_PTR(error);
2826 if (error > 0) 2826 if (error > 0)
2827 return NULL; 2827 return NULL;
2828 return res; 2828 return res;
2829 } 2829 }
2830 2830
2831 char *d_absolute_path(const struct path *path, 2831 char *d_absolute_path(const struct path *path,
2832 char *buf, int buflen) 2832 char *buf, int buflen)
2833 { 2833 {
2834 struct path root = {}; 2834 struct path root = {};
2835 char *res = buf + buflen; 2835 char *res = buf + buflen;
2836 int error; 2836 int error;
2837 2837
2838 prepend(&res, &buflen, "\0", 1); 2838 prepend(&res, &buflen, "\0", 1);
2839 br_read_lock(&vfsmount_lock); 2839 br_read_lock(&vfsmount_lock);
2840 error = prepend_path(path, &root, &res, &buflen); 2840 error = prepend_path(path, &root, &res, &buflen);
2841 br_read_unlock(&vfsmount_lock); 2841 br_read_unlock(&vfsmount_lock);
2842 2842
2843 if (error > 1) 2843 if (error > 1)
2844 error = -EINVAL; 2844 error = -EINVAL;
2845 if (error < 0) 2845 if (error < 0)
2846 return ERR_PTR(error); 2846 return ERR_PTR(error);
2847 return res; 2847 return res;
2848 } 2848 }
2849 2849
2850 /* 2850 /*
2851 * same as __d_path but appends "(deleted)" for unlinked files. 2851 * same as __d_path but appends "(deleted)" for unlinked files.
2852 */ 2852 */
2853 static int path_with_deleted(const struct path *path, 2853 static int path_with_deleted(const struct path *path,
2854 const struct path *root, 2854 const struct path *root,
2855 char **buf, int *buflen) 2855 char **buf, int *buflen)
2856 { 2856 {
2857 prepend(buf, buflen, "\0", 1); 2857 prepend(buf, buflen, "\0", 1);
2858 if (d_unlinked(path->dentry)) { 2858 if (d_unlinked(path->dentry)) {
2859 int error = prepend(buf, buflen, " (deleted)", 10); 2859 int error = prepend(buf, buflen, " (deleted)", 10);
2860 if (error) 2860 if (error)
2861 return error; 2861 return error;
2862 } 2862 }
2863 2863
2864 return prepend_path(path, root, buf, buflen); 2864 return prepend_path(path, root, buf, buflen);
2865 } 2865 }
2866 2866
2867 static int prepend_unreachable(char **buffer, int *buflen) 2867 static int prepend_unreachable(char **buffer, int *buflen)
2868 { 2868 {
2869 return prepend(buffer, buflen, "(unreachable)", 13); 2869 return prepend(buffer, buflen, "(unreachable)", 13);
2870 } 2870 }
2871 2871
2872 /** 2872 /**
2873 * d_path - return the path of a dentry 2873 * d_path - return the path of a dentry
2874 * @path: path to report 2874 * @path: path to report
2875 * @buf: buffer to return value in 2875 * @buf: buffer to return value in
2876 * @buflen: buffer length 2876 * @buflen: buffer length
2877 * 2877 *
2878 * Convert a dentry into an ASCII path name. If the entry has been deleted 2878 * Convert a dentry into an ASCII path name. If the entry has been deleted
2879 * the string " (deleted)" is appended. Note that this is ambiguous. 2879 * the string " (deleted)" is appended. Note that this is ambiguous.
2880 * 2880 *
2881 * Returns a pointer into the buffer or an error code if the path was 2881 * Returns a pointer into the buffer or an error code if the path was
2882 * too long. Note: Callers should use the returned pointer, not the passed 2882 * too long. Note: Callers should use the returned pointer, not the passed
2883 * in buffer, to use the name! The implementation often starts at an offset 2883 * in buffer, to use the name! The implementation often starts at an offset
2884 * into the buffer, and may leave 0 bytes at the start. 2884 * into the buffer, and may leave 0 bytes at the start.
2885 * 2885 *
2886 * "buflen" should be positive. 2886 * "buflen" should be positive.
2887 */ 2887 */
2888 char *d_path(const struct path *path, char *buf, int buflen) 2888 char *d_path(const struct path *path, char *buf, int buflen)
2889 { 2889 {
2890 char *res = buf + buflen; 2890 char *res = buf + buflen;
2891 struct path root; 2891 struct path root;
2892 int error; 2892 int error;
2893 2893
2894 /* 2894 /*
2895 * We have various synthetic filesystems that never get mounted. On 2895 * We have various synthetic filesystems that never get mounted. On
2896 * these filesystems dentries are never used for lookup purposes, and 2896 * these filesystems dentries are never used for lookup purposes, and
2897 * thus don't need to be hashed. They also don't need a name until a 2897 * thus don't need to be hashed. They also don't need a name until a
2898 * user wants to identify the object in /proc/pid/fd/. The little hack 2898 * user wants to identify the object in /proc/pid/fd/. The little hack
2899 * below allows us to generate a name for these objects on demand: 2899 * below allows us to generate a name for these objects on demand:
2900 */ 2900 */
2901 if (path->dentry->d_op && path->dentry->d_op->d_dname) 2901 if (path->dentry->d_op && path->dentry->d_op->d_dname)
2902 return path->dentry->d_op->d_dname(path->dentry, buf, buflen); 2902 return path->dentry->d_op->d_dname(path->dentry, buf, buflen);
2903 2903
2904 get_fs_root(current->fs, &root); 2904 get_fs_root(current->fs, &root);
2905 br_read_lock(&vfsmount_lock); 2905 br_read_lock(&vfsmount_lock);
2906 error = path_with_deleted(path, &root, &res, &buflen); 2906 error = path_with_deleted(path, &root, &res, &buflen);
2907 br_read_unlock(&vfsmount_lock); 2907 br_read_unlock(&vfsmount_lock);
2908 if (error < 0) 2908 if (error < 0)
2909 res = ERR_PTR(error); 2909 res = ERR_PTR(error);
2910 path_put(&root); 2910 path_put(&root);
2911 return res; 2911 return res;
2912 } 2912 }
2913 EXPORT_SYMBOL(d_path); 2913 EXPORT_SYMBOL(d_path);
2914 2914
2915 /* 2915 /*
2916 * Helper function for dentry_operations.d_dname() members 2916 * Helper function for dentry_operations.d_dname() members
2917 */ 2917 */
2918 char *dynamic_dname(struct dentry *dentry, char *buffer, int buflen, 2918 char *dynamic_dname(struct dentry *dentry, char *buffer, int buflen,
2919 const char *fmt, ...) 2919 const char *fmt, ...)
2920 { 2920 {
2921 va_list args; 2921 va_list args;
2922 char temp[64]; 2922 char temp[64];
2923 int sz; 2923 int sz;
2924 2924
2925 va_start(args, fmt); 2925 va_start(args, fmt);
2926 sz = vsnprintf(temp, sizeof(temp), fmt, args) + 1; 2926 sz = vsnprintf(temp, sizeof(temp), fmt, args) + 1;
2927 va_end(args); 2927 va_end(args);
2928 2928
2929 if (sz > sizeof(temp) || sz > buflen) 2929 if (sz > sizeof(temp) || sz > buflen)
2930 return ERR_PTR(-ENAMETOOLONG); 2930 return ERR_PTR(-ENAMETOOLONG);
2931 2931
2932 buffer += buflen - sz; 2932 buffer += buflen - sz;
2933 return memcpy(buffer, temp, sz); 2933 return memcpy(buffer, temp, sz);
2934 } 2934 }
2935 2935
2936 char *simple_dname(struct dentry *dentry, char *buffer, int buflen) 2936 char *simple_dname(struct dentry *dentry, char *buffer, int buflen)
2937 { 2937 {
2938 char *end = buffer + buflen; 2938 char *end = buffer + buflen;
2939 /* these dentries are never renamed, so d_lock is not needed */ 2939 /* these dentries are never renamed, so d_lock is not needed */
2940 if (prepend(&end, &buflen, " (deleted)", 11) || 2940 if (prepend(&end, &buflen, " (deleted)", 11) ||
2941 prepend(&end, &buflen, dentry->d_name.name, dentry->d_name.len) || 2941 prepend(&end, &buflen, dentry->d_name.name, dentry->d_name.len) ||
2942 prepend(&end, &buflen, "/", 1)) 2942 prepend(&end, &buflen, "/", 1))
2943 end = ERR_PTR(-ENAMETOOLONG); 2943 end = ERR_PTR(-ENAMETOOLONG);
2944 return end; 2944 return end;
2945 } 2945 }
2946 2946
2947 /* 2947 /*
2948 * Write full pathname from the root of the filesystem into the buffer. 2948 * Write full pathname from the root of the filesystem into the buffer.
2949 */ 2949 */
2950 static char *__dentry_path(struct dentry *dentry, char *buf, int buflen) 2950 static char *__dentry_path(struct dentry *dentry, char *buf, int buflen)
2951 { 2951 {
2952 char *end, *retval; 2952 char *end, *retval;
2953 int len, seq = 0; 2953 int len, seq = 0;
2954 int error = 0; 2954 int error = 0;
2955 2955
2956 rcu_read_lock(); 2956 rcu_read_lock();
2957 restart: 2957 restart:
2958 end = buf + buflen; 2958 end = buf + buflen;
2959 len = buflen; 2959 len = buflen;
2960 prepend(&end, &len, "\0", 1); 2960 prepend(&end, &len, "\0", 1);
2961 if (buflen < 1) 2961 if (buflen < 1)
2962 goto Elong; 2962 goto Elong;
2963 /* Get '/' right */ 2963 /* Get '/' right */
2964 retval = end-1; 2964 retval = end-1;
2965 *retval = '/'; 2965 *retval = '/';
2966 read_seqbegin_or_lock(&rename_lock, &seq); 2966 read_seqbegin_or_lock(&rename_lock, &seq);
2967 while (!IS_ROOT(dentry)) { 2967 while (!IS_ROOT(dentry)) {
2968 struct dentry *parent = dentry->d_parent; 2968 struct dentry *parent = dentry->d_parent;
2969 int error; 2969 int error;
2970 2970
2971 prefetch(parent); 2971 prefetch(parent);
2972 error = prepend_name(&end, &len, &dentry->d_name); 2972 error = prepend_name(&end, &len, &dentry->d_name);
2973 if (error) 2973 if (error)
2974 break; 2974 break;
2975 2975
2976 retval = end; 2976 retval = end;
2977 dentry = parent; 2977 dentry = parent;
2978 } 2978 }
2979 if (!(seq & 1)) 2979 if (!(seq & 1))
2980 rcu_read_unlock(); 2980 rcu_read_unlock();
2981 if (need_seqretry(&rename_lock, seq)) { 2981 if (need_seqretry(&rename_lock, seq)) {
2982 seq = 1; 2982 seq = 1;
2983 goto restart; 2983 goto restart;
2984 } 2984 }
2985 done_seqretry(&rename_lock, seq); 2985 done_seqretry(&rename_lock, seq);
2986 if (error) 2986 if (error)
2987 goto Elong; 2987 goto Elong;
2988 return retval; 2988 return retval;
2989 Elong: 2989 Elong:
2990 return ERR_PTR(-ENAMETOOLONG); 2990 return ERR_PTR(-ENAMETOOLONG);
2991 } 2991 }
2992 2992
2993 char *dentry_path_raw(struct dentry *dentry, char *buf, int buflen) 2993 char *dentry_path_raw(struct dentry *dentry, char *buf, int buflen)
2994 { 2994 {
2995 return __dentry_path(dentry, buf, buflen); 2995 return __dentry_path(dentry, buf, buflen);
2996 } 2996 }
2997 EXPORT_SYMBOL(dentry_path_raw); 2997 EXPORT_SYMBOL(dentry_path_raw);
2998 2998
2999 char *dentry_path(struct dentry *dentry, char *buf, int buflen) 2999 char *dentry_path(struct dentry *dentry, char *buf, int buflen)
3000 { 3000 {
3001 char *p = NULL; 3001 char *p = NULL;
3002 char *retval; 3002 char *retval;
3003 3003
3004 if (d_unlinked(dentry)) { 3004 if (d_unlinked(dentry)) {
3005 p = buf + buflen; 3005 p = buf + buflen;
3006 if (prepend(&p, &buflen, "//deleted", 10) != 0) 3006 if (prepend(&p, &buflen, "//deleted", 10) != 0)
3007 goto Elong; 3007 goto Elong;
3008 buflen++; 3008 buflen++;
3009 } 3009 }
3010 retval = __dentry_path(dentry, buf, buflen); 3010 retval = __dentry_path(dentry, buf, buflen);
3011 if (!IS_ERR(retval) && p) 3011 if (!IS_ERR(retval) && p)
3012 *p = '/'; /* restore '/' overriden with '\0' */ 3012 *p = '/'; /* restore '/' overriden with '\0' */
3013 return retval; 3013 return retval;
3014 Elong: 3014 Elong:
3015 return ERR_PTR(-ENAMETOOLONG); 3015 return ERR_PTR(-ENAMETOOLONG);
3016 } 3016 }
3017 3017
3018 static inline void get_fs_root_and_pwd(struct fs_struct *fs, struct path *root,
3019 struct path *pwd)
3020 {
3021 spin_lock(&fs->lock);
3022 *root = fs->root;
3023 path_get(root);
3024 *pwd = fs->pwd;
3025 path_get(pwd);
3026 spin_unlock(&fs->lock);
3027 }
3028
3018 /* 3029 /*
3019 * NOTE! The user-level library version returns a 3030 * NOTE! The user-level library version returns a
3020 * character pointer. The kernel system call just 3031 * character pointer. The kernel system call just
3021 * returns the length of the buffer filled (which 3032 * returns the length of the buffer filled (which
3022 * includes the ending '\0' character), or a negative 3033 * includes the ending '\0' character), or a negative
3023 * error value. So libc would do something like 3034 * error value. So libc would do something like
3024 * 3035 *
3025 * char *getcwd(char * buf, size_t size) 3036 * char *getcwd(char * buf, size_t size)
3026 * { 3037 * {
3027 * int retval; 3038 * int retval;
3028 * 3039 *
3029 * retval = sys_getcwd(buf, size); 3040 * retval = sys_getcwd(buf, size);
3030 * if (retval >= 0) 3041 * if (retval >= 0)
3031 * return buf; 3042 * return buf;
3032 * errno = -retval; 3043 * errno = -retval;
3033 * return NULL; 3044 * return NULL;
3034 * } 3045 * }
3035 */ 3046 */
3036 SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size) 3047 SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size)
3037 { 3048 {
3038 int error; 3049 int error;
3039 struct path pwd, root; 3050 struct path pwd, root;
3040 char *page = (char *) __get_free_page(GFP_USER); 3051 char *page = (char *) __get_free_page(GFP_USER);
3041 3052
3042 if (!page) 3053 if (!page)
3043 return -ENOMEM; 3054 return -ENOMEM;
3044 3055
3045 get_fs_root_and_pwd(current->fs, &root, &pwd); 3056 get_fs_root_and_pwd(current->fs, &root, &pwd);
3046 3057
3047 error = -ENOENT; 3058 error = -ENOENT;
3048 br_read_lock(&vfsmount_lock); 3059 br_read_lock(&vfsmount_lock);
3049 if (!d_unlinked(pwd.dentry)) { 3060 if (!d_unlinked(pwd.dentry)) {
3050 unsigned long len; 3061 unsigned long len;
3051 char *cwd = page + PAGE_SIZE; 3062 char *cwd = page + PAGE_SIZE;
3052 int buflen = PAGE_SIZE; 3063 int buflen = PAGE_SIZE;
3053 3064
3054 prepend(&cwd, &buflen, "\0", 1); 3065 prepend(&cwd, &buflen, "\0", 1);
3055 error = prepend_path(&pwd, &root, &cwd, &buflen); 3066 error = prepend_path(&pwd, &root, &cwd, &buflen);
3056 br_read_unlock(&vfsmount_lock); 3067 br_read_unlock(&vfsmount_lock);
3057 3068
3058 if (error < 0) 3069 if (error < 0)
3059 goto out; 3070 goto out;
3060 3071
3061 /* Unreachable from current root */ 3072 /* Unreachable from current root */
3062 if (error > 0) { 3073 if (error > 0) {
3063 error = prepend_unreachable(&cwd, &buflen); 3074 error = prepend_unreachable(&cwd, &buflen);
3064 if (error) 3075 if (error)
3065 goto out; 3076 goto out;
3066 } 3077 }
3067 3078
3068 error = -ERANGE; 3079 error = -ERANGE;
3069 len = PAGE_SIZE + page - cwd; 3080 len = PAGE_SIZE + page - cwd;
3070 if (len <= size) { 3081 if (len <= size) {
3071 error = len; 3082 error = len;
3072 if (copy_to_user(buf, cwd, len)) 3083 if (copy_to_user(buf, cwd, len))
3073 error = -EFAULT; 3084 error = -EFAULT;
3074 } 3085 }
3075 } else { 3086 } else {
3076 br_read_unlock(&vfsmount_lock); 3087 br_read_unlock(&vfsmount_lock);
3077 } 3088 }
3078 3089
3079 out: 3090 out:
3080 path_put(&pwd); 3091 path_put(&pwd);
3081 path_put(&root); 3092 path_put(&root);
3082 free_page((unsigned long) page); 3093 free_page((unsigned long) page);
3083 return error; 3094 return error;
3084 } 3095 }
3085 3096
3086 /* 3097 /*
3087 * Test whether new_dentry is a subdirectory of old_dentry. 3098 * Test whether new_dentry is a subdirectory of old_dentry.
3088 * 3099 *
3089 * Trivially implemented using the dcache structure 3100 * Trivially implemented using the dcache structure
3090 */ 3101 */
3091 3102
3092 /** 3103 /**
3093 * is_subdir - is new dentry a subdirectory of old_dentry 3104 * is_subdir - is new dentry a subdirectory of old_dentry
3094 * @new_dentry: new dentry 3105 * @new_dentry: new dentry
3095 * @old_dentry: old dentry 3106 * @old_dentry: old dentry
3096 * 3107 *
3097 * Returns 1 if new_dentry is a subdirectory of the parent (at any depth). 3108 * Returns 1 if new_dentry is a subdirectory of the parent (at any depth).
3098 * Returns 0 otherwise. 3109 * Returns 0 otherwise.
3099 * Caller must ensure that "new_dentry" is pinned before calling is_subdir() 3110 * Caller must ensure that "new_dentry" is pinned before calling is_subdir()
3100 */ 3111 */
3101 3112
3102 int is_subdir(struct dentry *new_dentry, struct dentry *old_dentry) 3113 int is_subdir(struct dentry *new_dentry, struct dentry *old_dentry)
3103 { 3114 {
3104 int result; 3115 int result;
3105 unsigned seq; 3116 unsigned seq;
3106 3117
3107 if (new_dentry == old_dentry) 3118 if (new_dentry == old_dentry)
3108 return 1; 3119 return 1;
3109 3120
3110 do { 3121 do {
3111 /* for restarting inner loop in case of seq retry */ 3122 /* for restarting inner loop in case of seq retry */
3112 seq = read_seqbegin(&rename_lock); 3123 seq = read_seqbegin(&rename_lock);
3113 /* 3124 /*
3114 * Need rcu_readlock to protect against the d_parent trashing 3125 * Need rcu_readlock to protect against the d_parent trashing
3115 * due to d_move 3126 * due to d_move
3116 */ 3127 */
3117 rcu_read_lock(); 3128 rcu_read_lock();
3118 if (d_ancestor(old_dentry, new_dentry)) 3129 if (d_ancestor(old_dentry, new_dentry))
3119 result = 1; 3130 result = 1;
3120 else 3131 else
3121 result = 0; 3132 result = 0;
3122 rcu_read_unlock(); 3133 rcu_read_unlock();
3123 } while (read_seqretry(&rename_lock, seq)); 3134 } while (read_seqretry(&rename_lock, seq));
3124 3135
3125 return result; 3136 return result;
3126 } 3137 }
3127 3138
3128 static enum d_walk_ret d_genocide_kill(void *data, struct dentry *dentry) 3139 static enum d_walk_ret d_genocide_kill(void *data, struct dentry *dentry)
3129 { 3140 {
3130 struct dentry *root = data; 3141 struct dentry *root = data;
3131 if (dentry != root) { 3142 if (dentry != root) {
3132 if (d_unhashed(dentry) || !dentry->d_inode) 3143 if (d_unhashed(dentry) || !dentry->d_inode)
3133 return D_WALK_SKIP; 3144 return D_WALK_SKIP;
3134 3145
3135 if (!(dentry->d_flags & DCACHE_GENOCIDE)) { 3146 if (!(dentry->d_flags & DCACHE_GENOCIDE)) {
3136 dentry->d_flags |= DCACHE_GENOCIDE; 3147 dentry->d_flags |= DCACHE_GENOCIDE;
3137 dentry->d_lockref.count--; 3148 dentry->d_lockref.count--;
3138 } 3149 }
3139 } 3150 }
3140 return D_WALK_CONTINUE; 3151 return D_WALK_CONTINUE;
3141 } 3152 }
3142 3153
3143 void d_genocide(struct dentry *parent) 3154 void d_genocide(struct dentry *parent)
3144 { 3155 {
3145 d_walk(parent, parent, d_genocide_kill, NULL); 3156 d_walk(parent, parent, d_genocide_kill, NULL);
3146 } 3157 }
3147 3158
3148 void d_tmpfile(struct dentry *dentry, struct inode *inode) 3159 void d_tmpfile(struct dentry *dentry, struct inode *inode)
3149 { 3160 {
3150 inode_dec_link_count(inode); 3161 inode_dec_link_count(inode);
3151 BUG_ON(dentry->d_name.name != dentry->d_iname || 3162 BUG_ON(dentry->d_name.name != dentry->d_iname ||
3152 !hlist_unhashed(&dentry->d_alias) || 3163 !hlist_unhashed(&dentry->d_alias) ||
3153 !d_unlinked(dentry)); 3164 !d_unlinked(dentry));
3154 spin_lock(&dentry->d_parent->d_lock); 3165 spin_lock(&dentry->d_parent->d_lock);
3155 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); 3166 spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
3156 dentry->d_name.len = sprintf(dentry->d_iname, "#%llu", 3167 dentry->d_name.len = sprintf(dentry->d_iname, "#%llu",
3157 (unsigned long long)inode->i_ino); 3168 (unsigned long long)inode->i_ino);
3158 spin_unlock(&dentry->d_lock); 3169 spin_unlock(&dentry->d_lock);
3159 spin_unlock(&dentry->d_parent->d_lock); 3170 spin_unlock(&dentry->d_parent->d_lock);
3160 d_instantiate(dentry, inode); 3171 d_instantiate(dentry, inode);
3161 } 3172 }
3162 EXPORT_SYMBOL(d_tmpfile); 3173 EXPORT_SYMBOL(d_tmpfile);
3163 3174
3164 static __initdata unsigned long dhash_entries; 3175 static __initdata unsigned long dhash_entries;
3165 static int __init set_dhash_entries(char *str) 3176 static int __init set_dhash_entries(char *str)
3166 { 3177 {
3167 if (!str) 3178 if (!str)
3168 return 0; 3179 return 0;
3169 dhash_entries = simple_strtoul(str, &str, 0); 3180 dhash_entries = simple_strtoul(str, &str, 0);
3170 return 1; 3181 return 1;
3171 } 3182 }
3172 __setup("dhash_entries=", set_dhash_entries); 3183 __setup("dhash_entries=", set_dhash_entries);
3173 3184
3174 static void __init dcache_init_early(void) 3185 static void __init dcache_init_early(void)
3175 { 3186 {
3176 unsigned int loop; 3187 unsigned int loop;
3177 3188
3178 /* If hashes are distributed across NUMA nodes, defer 3189 /* If hashes are distributed across NUMA nodes, defer
3179 * hash allocation until vmalloc space is available. 3190 * hash allocation until vmalloc space is available.
3180 */ 3191 */
3181 if (hashdist) 3192 if (hashdist)
3182 return; 3193 return;
3183 3194
3184 dentry_hashtable = 3195 dentry_hashtable =
3185 alloc_large_system_hash("Dentry cache", 3196 alloc_large_system_hash("Dentry cache",
3186 sizeof(struct hlist_bl_head), 3197 sizeof(struct hlist_bl_head),
3187 dhash_entries, 3198 dhash_entries,
3188 13, 3199 13,
3189 HASH_EARLY, 3200 HASH_EARLY,
3190 &d_hash_shift, 3201 &d_hash_shift,
3191 &d_hash_mask, 3202 &d_hash_mask,
3192 0, 3203 0,
3193 0); 3204 0);
3194 3205
3195 for (loop = 0; loop < (1U << d_hash_shift); loop++) 3206 for (loop = 0; loop < (1U << d_hash_shift); loop++)
3196 INIT_HLIST_BL_HEAD(dentry_hashtable + loop); 3207 INIT_HLIST_BL_HEAD(dentry_hashtable + loop);
3197 } 3208 }
3198 3209
3199 static void __init dcache_init(void) 3210 static void __init dcache_init(void)
3200 { 3211 {
3201 unsigned int loop; 3212 unsigned int loop;
3202 3213
3203 /* 3214 /*
3204 * A constructor could be added for stable state like the lists, 3215 * A constructor could be added for stable state like the lists,
3205 * but it is probably not worth it because of the cache nature 3216 * but it is probably not worth it because of the cache nature
3206 * of the dcache. 3217 * of the dcache.
3207 */ 3218 */
3208 dentry_cache = KMEM_CACHE(dentry, 3219 dentry_cache = KMEM_CACHE(dentry,
3209 SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD); 3220 SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD);
3210 3221
3211 /* Hash may have been set up in dcache_init_early */ 3222 /* Hash may have been set up in dcache_init_early */
3212 if (!hashdist) 3223 if (!hashdist)
3213 return; 3224 return;
3214 3225
3215 dentry_hashtable = 3226 dentry_hashtable =
3216 alloc_large_system_hash("Dentry cache", 3227 alloc_large_system_hash("Dentry cache",
3217 sizeof(struct hlist_bl_head), 3228 sizeof(struct hlist_bl_head),
3218 dhash_entries, 3229 dhash_entries,
3219 13, 3230 13,
3220 0, 3231 0,
3221 &d_hash_shift, 3232 &d_hash_shift,
3222 &d_hash_mask, 3233 &d_hash_mask,
3223 0, 3234 0,
3224 0); 3235 0);
3225 3236
3226 for (loop = 0; loop < (1U << d_hash_shift); loop++) 3237 for (loop = 0; loop < (1U << d_hash_shift); loop++)
3227 INIT_HLIST_BL_HEAD(dentry_hashtable + loop); 3238 INIT_HLIST_BL_HEAD(dentry_hashtable + loop);
3228 } 3239 }
3229 3240
3230 /* SLAB cache for __getname() consumers */ 3241 /* SLAB cache for __getname() consumers */
3231 struct kmem_cache *names_cachep __read_mostly; 3242 struct kmem_cache *names_cachep __read_mostly;
3232 EXPORT_SYMBOL(names_cachep); 3243 EXPORT_SYMBOL(names_cachep);
3233 3244
3234 EXPORT_SYMBOL(d_genocide); 3245 EXPORT_SYMBOL(d_genocide);
3235 3246
3236 void __init vfs_caches_init_early(void) 3247 void __init vfs_caches_init_early(void)
3237 { 3248 {
3238 dcache_init_early(); 3249 dcache_init_early();
3239 inode_init_early(); 3250 inode_init_early();
3240 } 3251 }
3241 3252
3242 void __init vfs_caches_init(unsigned long mempages) 3253 void __init vfs_caches_init(unsigned long mempages)
3243 { 3254 {
3244 unsigned long reserve; 3255 unsigned long reserve;
3245 3256
3246 /* Base hash sizes on available memory, with a reserve equal to 3257 /* Base hash sizes on available memory, with a reserve equal to
3247 150% of current kernel size */ 3258 150% of current kernel size */
3248 3259
3249 reserve = min((mempages - nr_free_pages()) * 3/2, mempages - 1); 3260 reserve = min((mempages - nr_free_pages()) * 3/2, mempages - 1);
3250 mempages -= reserve; 3261 mempages -= reserve;
3251 3262
3252 names_cachep = kmem_cache_create("names_cache", PATH_MAX, 0, 3263 names_cachep = kmem_cache_create("names_cache", PATH_MAX, 0,
3253 SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); 3264 SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
3254 3265
3255 dcache_init(); 3266 dcache_init();
3256 inode_init(); 3267 inode_init();
3257 files_init(mempages); 3268 files_init(mempages);
3258 mnt_init(); 3269 mnt_init();
3259 bdev_cache_init(); 3270 bdev_cache_init();
3260 chrdev_init(); 3271 chrdev_init();
3261 } 3272 }
3262 3273
include/linux/fs_struct.h
Diff comments   View file @ 5762482
1 #ifndef _LINUX_FS_STRUCT_H 1 #ifndef _LINUX_FS_STRUCT_H
2 #define _LINUX_FS_STRUCT_H 2 #define _LINUX_FS_STRUCT_H
3 3
4 #include <linux/path.h> 4 #include <linux/path.h>
5 #include <linux/spinlock.h> 5 #include <linux/spinlock.h>
6 #include <linux/seqlock.h> 6 #include <linux/seqlock.h>
7 7
8 struct fs_struct { 8 struct fs_struct {
9 int users; 9 int users;
10 spinlock_t lock; 10 spinlock_t lock;
11 seqcount_t seq; 11 seqcount_t seq;
12 int umask; 12 int umask;
13 int in_exec; 13 int in_exec;
14 struct path root, pwd; 14 struct path root, pwd;
15 }; 15 };
16 16
17 extern struct kmem_cache *fs_cachep; 17 extern struct kmem_cache *fs_cachep;
18 18
19 extern void exit_fs(struct task_struct *); 19 extern void exit_fs(struct task_struct *);
20 extern void set_fs_root(struct fs_struct *, const struct path *); 20 extern void set_fs_root(struct fs_struct *, const struct path *);
21 extern void set_fs_pwd(struct fs_struct *, const struct path *); 21 extern void set_fs_pwd(struct fs_struct *, const struct path *);
22 extern struct fs_struct *copy_fs_struct(struct fs_struct *); 22 extern struct fs_struct *copy_fs_struct(struct fs_struct *);
23 extern void free_fs_struct(struct fs_struct *); 23 extern void free_fs_struct(struct fs_struct *);
24 extern int unshare_fs_struct(void); 24 extern int unshare_fs_struct(void);
25 25
26 static inline void get_fs_root(struct fs_struct *fs, struct path *root) 26 static inline void get_fs_root(struct fs_struct *fs, struct path *root)
27 { 27 {
28 spin_lock(&fs->lock); 28 spin_lock(&fs->lock);
29 *root = fs->root; 29 *root = fs->root;
30 path_get(root); 30 path_get(root);
31 spin_unlock(&fs->lock); 31 spin_unlock(&fs->lock);
32 } 32 }
33 33
34 static inline void get_fs_pwd(struct fs_struct *fs, struct path *pwd) 34 static inline void get_fs_pwd(struct fs_struct *fs, struct path *pwd)
35 { 35 {
36 spin_lock(&fs->lock); 36 spin_lock(&fs->lock);
37 *pwd = fs->pwd; 37 *pwd = fs->pwd;
38 path_get(pwd); 38 path_get(pwd);
39 spin_unlock(&fs->lock); 39 spin_unlock(&fs->lock);
40 } 40 }
41 41
42 static inline void get_fs_root_and_pwd(struct fs_struct *fs, struct path *root,
43 struct path *pwd)
44 {
45 spin_lock(&fs->lock);
46 *root = fs->root;
47 path_get(root);
48 *pwd = fs->pwd;
49 path_get(pwd);
50 spin_unlock(&fs->lock);
51 }
52
53 extern bool current_chrooted(void); 42 extern bool current_chrooted(void);
54 43
55 #endif /* _LINUX_FS_STRUCT_H */ 44 #endif /* _LINUX_FS_STRUCT_H */
56 45