Blame view
fs/dcache.c
89.4 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
/* * fs/dcache.c * * Complete reimplementation * (C) 1997 Thomas Schoebel-Theuer, * with heavy changes by Linus Torvalds */ /* * Notes on the allocation strategy: * * The dcache is a master of the icache - whenever a dcache entry * exists, the inode will always exist. "iput()" is done either when * the dcache entry is deleted or garbage collected. */ |
1da177e4c
|
16 17 18 19 |
#include <linux/syscalls.h> #include <linux/string.h> #include <linux/mm.h> #include <linux/fs.h> |
7a91bf7f5
|
20 |
#include <linux/fsnotify.h> |
1da177e4c
|
21 22 |
#include <linux/slab.h> #include <linux/init.h> |
1da177e4c
|
23 24 |
#include <linux/hash.h> #include <linux/cache.h> |
630d9c472
|
25 |
#include <linux/export.h> |
1da177e4c
|
26 27 28 29 30 31 32 |
#include <linux/mount.h> #include <linux/file.h> #include <asm/uaccess.h> #include <linux/security.h> #include <linux/seqlock.h> #include <linux/swap.h> #include <linux/bootmem.h> |
5ad4e53bd
|
33 |
#include <linux/fs_struct.h> |
613afbf83
|
34 |
#include <linux/hardirq.h> |
ceb5bdc2d
|
35 36 |
#include <linux/bit_spinlock.h> #include <linux/rculist_bl.h> |
268bb0ce3
|
37 |
#include <linux/prefetch.h> |
dd179946d
|
38 |
#include <linux/ratelimit.h> |
f60415675
|
39 |
#include <linux/list_lru.h> |
df4c0e36f
|
40 |
#include <linux/kasan.h> |
07f3f05c1
|
41 |
#include "internal.h" |
b2dba1af3
|
42 |
#include "mount.h" |
1da177e4c
|
43 |
|
789680d1e
|
44 45 |
/* * Usage: |
873feea09
|
46 |
* dcache->d_inode->i_lock protects: |
946e51f2b
|
47 |
* - i_dentry, d_u.d_alias, d_inode of aliases |
ceb5bdc2d
|
48 49 50 51 |
* dcache_hash_bucket lock protects: * - the dcache hash table * s_anon bl list spinlock protects: * - the s_anon list (see __d_drop) |
19156840e
|
52 |
* dentry->d_sb->s_dentry_lru_lock protects: |
230445078
|
53 54 55 56 57 |
* - the dcache lru lists and counters * d_lock protects: * - d_flags * - d_name * - d_lru |
b7ab39f63
|
58 |
* - d_count |
da5029563
|
59 |
* - d_unhashed() |
2fd6b7f50
|
60 61 |
* - d_parent and d_subdirs * - childrens' d_child and d_parent |
946e51f2b
|
62 |
* - d_u.d_alias, d_inode |
789680d1e
|
63 64 |
* * Ordering: |
873feea09
|
65 |
* dentry->d_inode->i_lock |
b5c84bf6f
|
66 |
* dentry->d_lock |
19156840e
|
67 |
* dentry->d_sb->s_dentry_lru_lock |
ceb5bdc2d
|
68 69 |
* dcache_hash_bucket lock * s_anon lock |
789680d1e
|
70 |
* |
da5029563
|
71 72 73 74 75 76 77 |
* If there is an ancestor relationship: * dentry->d_parent->...->d_parent->d_lock * ... * dentry->d_parent->d_lock * dentry->d_lock * * If no ancestor relationship: |
789680d1e
|
78 79 80 81 |
* if (dentry1 < dentry2) * dentry1->d_lock * dentry2->d_lock */ |
fa3536cc1
|
82 |
int sysctl_vfs_cache_pressure __read_mostly = 100; |
1da177e4c
|
83 |
EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure); |
74c3cbe33
|
84 |
__cacheline_aligned_in_smp DEFINE_SEQLOCK(rename_lock); |
1da177e4c
|
85 |
|
949854d02
|
86 |
EXPORT_SYMBOL(rename_lock); |
1da177e4c
|
87 |
|
e18b890bb
|
88 |
static struct kmem_cache *dentry_cache __read_mostly; |
1da177e4c
|
89 |
|
1da177e4c
|
90 91 92 93 94 95 96 97 |
/* * This is the single most critical data structure when it comes * to the dcache: the hashtable for lookups. Somebody should try * to make this good - I've just made it work. * * This hash-function tries to avoid losing too many bits of hash * information, yet avoid using a prime hash-size or similar. */ |
1da177e4c
|
98 |
|
fa3536cc1
|
99 100 |
static unsigned int d_hash_mask __read_mostly; static unsigned int d_hash_shift __read_mostly; |
ceb5bdc2d
|
101 |
|
b07ad9967
|
102 |
static struct hlist_bl_head *dentry_hashtable __read_mostly; |
ceb5bdc2d
|
103 |
|
8966be903
|
104 |
static inline struct hlist_bl_head *d_hash(const struct dentry *parent, |
6d7d1a0dc
|
105 |
unsigned int hash) |
ceb5bdc2d
|
106 |
{ |
6d7d1a0dc
|
107 |
hash += (unsigned long) parent / L1_CACHE_BYTES; |
99d263d4c
|
108 |
return dentry_hashtable + hash_32(hash, d_hash_shift); |
ceb5bdc2d
|
109 |
} |
1da177e4c
|
110 111 112 113 |
/* Statistics gathering. */ struct dentry_stat_t dentry_stat = { .age_limit = 45, }; |
3942c07cc
|
114 |
static DEFINE_PER_CPU(long, nr_dentry); |
62d36c770
|
115 |
static DEFINE_PER_CPU(long, nr_dentry_unused); |
312d3ca85
|
116 117 |
#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS) |
62d36c770
|
118 119 120 121 122 123 124 125 126 127 128 129 130 |
/* * Here we resort to our own counters instead of using generic per-cpu counters * for consistency with what the vfs inode code does. We are expected to harvest * better code and performance by having our own specialized counters. * * Please note that the loop is done over all possible CPUs, not over all online * CPUs. The reason for this is that we don't want to play games with CPUs going * on and off. If one of them goes off, we will just keep their counters. * * glommer: See cffbc8a for details, and if you ever intend to change this, * please update all vfs counters to match. */ |
3942c07cc
|
131 |
static long get_nr_dentry(void) |
3e880fb5e
|
132 133 |
{ int i; |
3942c07cc
|
134 |
long sum = 0; |
3e880fb5e
|
135 136 137 138 |
for_each_possible_cpu(i) sum += per_cpu(nr_dentry, i); return sum < 0 ? 0 : sum; } |
62d36c770
|
139 140 141 142 143 144 145 146 |
static long get_nr_dentry_unused(void) { int i; long sum = 0; for_each_possible_cpu(i) sum += per_cpu(nr_dentry_unused, i); return sum < 0 ? 0 : sum; } |
1f7e0616c
|
147 |
int proc_nr_dentry(struct ctl_table *table, int write, void __user *buffer, |
312d3ca85
|
148 149 |
size_t *lenp, loff_t *ppos) { |
3e880fb5e
|
150 |
dentry_stat.nr_dentry = get_nr_dentry(); |
62d36c770
|
151 |
dentry_stat.nr_unused = get_nr_dentry_unused(); |
3942c07cc
|
152 |
return proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
312d3ca85
|
153 154 |
} #endif |
5483f18e9
|
155 156 157 158 |
/* * Compare 2 name strings, return 0 if they match, otherwise non-zero. * The strings are both count bytes long, and count is non-zero. */ |
e419b4cc5
|
159 160 161 162 163 164 165 166 167 168 169 170 |
#ifdef CONFIG_DCACHE_WORD_ACCESS #include <asm/word-at-a-time.h> /* * NOTE! 'cs' and 'scount' come from a dentry, so it has a * aligned allocation for this particular component. We don't * strictly need the load_unaligned_zeropad() safety, but it * doesn't hurt either. * * In contrast, 'ct' and 'tcount' can be from a pathname, and do * need the careful unaligned handling. */ |
94753db5e
|
171 |
static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount) |
5483f18e9
|
172 |
{ |
bfcfaa77b
|
173 |
unsigned long a,b,mask; |
bfcfaa77b
|
174 175 |
for (;;) { |
12f8ad4b0
|
176 |
a = *(unsigned long *)cs; |
e419b4cc5
|
177 |
b = load_unaligned_zeropad(ct); |
bfcfaa77b
|
178 179 180 181 182 183 184 185 186 187 |
if (tcount < sizeof(unsigned long)) break; if (unlikely(a != b)) return 1; cs += sizeof(unsigned long); ct += sizeof(unsigned long); tcount -= sizeof(unsigned long); if (!tcount) return 0; } |
a5c21dcef
|
188 |
mask = bytemask_from_count(tcount); |
bfcfaa77b
|
189 |
return unlikely(!!((a ^ b) & mask)); |
e419b4cc5
|
190 |
} |
bfcfaa77b
|
191 |
#else |
e419b4cc5
|
192 |
|
94753db5e
|
193 |
static inline int dentry_string_cmp(const unsigned char *cs, const unsigned char *ct, unsigned tcount) |
e419b4cc5
|
194 |
{ |
5483f18e9
|
195 196 197 198 199 200 201 202 203 |
do { if (*cs != *ct) return 1; cs++; ct++; tcount--; } while (tcount); return 0; } |
e419b4cc5
|
204 |
#endif |
94753db5e
|
205 206 |
static inline int dentry_cmp(const struct dentry *dentry, const unsigned char *ct, unsigned tcount) { |
6326c71fd
|
207 |
const unsigned char *cs; |
94753db5e
|
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 |
/* * Be careful about RCU walk racing with rename: * use ACCESS_ONCE to fetch the name pointer. * * NOTE! Even if a rename will mean that the length * was not loaded atomically, we don't care. The * RCU walk will check the sequence count eventually, * and catch it. And we won't overrun the buffer, * because we're reading the name pointer atomically, * and a dentry name is guaranteed to be properly * terminated with a NUL byte. * * End result: even if 'len' is wrong, we'll exit * early because the data cannot match (there can * be no NUL in the ct/tcount data) */ |
6326c71fd
|
224 225 226 |
cs = ACCESS_ONCE(dentry->d_name.name); smp_read_barrier_depends(); return dentry_string_cmp(cs, ct, tcount); |
94753db5e
|
227 |
} |
8d85b4845
|
228 229 230 231 232 233 234 235 236 237 238 239 |
struct external_name { union { atomic_t count; struct rcu_head head; } u; unsigned char name[]; }; static inline struct external_name *external_name(struct dentry *dentry) { return container_of(dentry->d_name.name, struct external_name, name[0]); } |
9c82ab9c9
|
240 |
static void __d_free(struct rcu_head *head) |
1da177e4c
|
241 |
{ |
9c82ab9c9
|
242 |
struct dentry *dentry = container_of(head, struct dentry, d_u.d_rcu); |
8d85b4845
|
243 244 245 246 247 248 |
kmem_cache_free(dentry_cache, dentry); } static void __d_free_external(struct rcu_head *head) { struct dentry *dentry = container_of(head, struct dentry, d_u.d_rcu); |
8d85b4845
|
249 |
kfree(external_name(dentry)); |
1da177e4c
|
250 251 |
kmem_cache_free(dentry_cache, dentry); } |
810bb1726
|
252 253 254 255 |
static inline int dname_external(const struct dentry *dentry) { return dentry->d_name.name != dentry->d_iname; } |
4bf46a272
|
256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 |
/* * Make sure other CPUs see the inode attached before the type is set. */ static inline void __d_set_inode_and_type(struct dentry *dentry, struct inode *inode, unsigned type_flags) { unsigned flags; dentry->d_inode = inode; smp_wmb(); flags = READ_ONCE(dentry->d_flags); flags &= ~(DCACHE_ENTRY_TYPE | DCACHE_FALLTHRU); flags |= type_flags; WRITE_ONCE(dentry->d_flags, flags); } /* * Ideally, we want to make sure that other CPUs see the flags cleared before * the inode is detached, but this is really a violation of RCU principles * since the ordering suggests we should always set inode before flags. * * We should instead replace or discard the entire dentry - but that sucks * performancewise on mass deletion/rename. */ static inline void __d_clear_type_and_inode(struct dentry *dentry) { unsigned flags = READ_ONCE(dentry->d_flags); flags &= ~(DCACHE_ENTRY_TYPE | DCACHE_FALLTHRU); WRITE_ONCE(dentry->d_flags, flags); smp_wmb(); dentry->d_inode = NULL; } |
b4f0354e9
|
290 291 |
static void dentry_free(struct dentry *dentry) { |
946e51f2b
|
292 |
WARN_ON(!hlist_unhashed(&dentry->d_u.d_alias)); |
8d85b4845
|
293 294 295 296 297 298 299 |
if (unlikely(dname_external(dentry))) { struct external_name *p = external_name(dentry); if (likely(atomic_dec_and_test(&p->u.count))) { call_rcu(&dentry->d_u.d_rcu, __d_free_external); return; } } |
b4f0354e9
|
300 301 302 303 304 305 |
/* if dentry was never visible to RCU, immediate free is OK */ if (!(dentry->d_flags & DCACHE_RCUACCESS)) __d_free(&dentry->d_u.d_rcu); else call_rcu(&dentry->d_u.d_rcu, __d_free); } |
31e6b01f4
|
306 |
/** |
a7c6f571f
|
307 |
* dentry_rcuwalk_invalidate - invalidate in-progress rcu-walk lookups |
ff5fdb614
|
308 |
* @dentry: the target dentry |
31e6b01f4
|
309 310 311 312 |
* After this call, in-progress rcu-walk path lookup will fail. This * should be called after unhashing, and after changing d_inode (if * the dentry has not already been unhashed). */ |
a7c6f571f
|
313 |
static inline void dentry_rcuwalk_invalidate(struct dentry *dentry) |
31e6b01f4
|
314 |
{ |
a7c6f571f
|
315 316 317 |
lockdep_assert_held(&dentry->d_lock); /* Go through am invalidation barrier */ write_seqcount_invalidate(&dentry->d_seq); |
31e6b01f4
|
318 |
} |
1da177e4c
|
319 320 |
/* * Release the dentry's inode, using the filesystem |
31e6b01f4
|
321 322 |
* d_iput() operation if defined. Dentry has no refcount * and is unhashed. |
1da177e4c
|
323 |
*/ |
858119e15
|
324 |
static void dentry_iput(struct dentry * dentry) |
31f3e0b3a
|
325 |
__releases(dentry->d_lock) |
873feea09
|
326 |
__releases(dentry->d_inode->i_lock) |
1da177e4c
|
327 328 329 |
{ struct inode *inode = dentry->d_inode; if (inode) { |
4bf46a272
|
330 |
__d_clear_type_and_inode(dentry); |
946e51f2b
|
331 |
hlist_del_init(&dentry->d_u.d_alias); |
1da177e4c
|
332 |
spin_unlock(&dentry->d_lock); |
873feea09
|
333 |
spin_unlock(&inode->i_lock); |
f805fbdaa
|
334 335 |
if (!inode->i_nlink) fsnotify_inoderemove(inode); |
1da177e4c
|
336 337 338 339 340 341 |
if (dentry->d_op && dentry->d_op->d_iput) dentry->d_op->d_iput(dentry, inode); else iput(inode); } else { spin_unlock(&dentry->d_lock); |
1da177e4c
|
342 343 |
} } |
da3bbdd46
|
344 |
/* |
31e6b01f4
|
345 346 347 348 349 |
* Release the dentry's inode, using the filesystem * d_iput() operation if defined. dentry remains in-use. */ static void dentry_unlink_inode(struct dentry * dentry) __releases(dentry->d_lock) |
873feea09
|
350 |
__releases(dentry->d_inode->i_lock) |
31e6b01f4
|
351 352 |
{ struct inode *inode = dentry->d_inode; |
4bf46a272
|
353 |
__d_clear_type_and_inode(dentry); |
946e51f2b
|
354 |
hlist_del_init(&dentry->d_u.d_alias); |
a7c6f571f
|
355 |
dentry_rcuwalk_invalidate(dentry); |
31e6b01f4
|
356 |
spin_unlock(&dentry->d_lock); |
873feea09
|
357 |
spin_unlock(&inode->i_lock); |
31e6b01f4
|
358 359 360 361 362 363 364 365 366 |
if (!inode->i_nlink) fsnotify_inoderemove(inode); if (dentry->d_op && dentry->d_op->d_iput) dentry->d_op->d_iput(dentry, inode); else iput(inode); } /* |
89dc77bcd
|
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 |
* The DCACHE_LRU_LIST bit is set whenever the 'd_lru' entry * is in use - which includes both the "real" per-superblock * LRU list _and_ the DCACHE_SHRINK_LIST use. * * The DCACHE_SHRINK_LIST bit is set whenever the dentry is * on the shrink list (ie not on the superblock LRU list). * * The per-cpu "nr_dentry_unused" counters are updated with * the DCACHE_LRU_LIST bit. * * These helper functions make sure we always follow the * rules. d_lock must be held by the caller. */ #define D_FLAG_VERIFY(dentry,x) WARN_ON_ONCE(((dentry)->d_flags & (DCACHE_LRU_LIST | DCACHE_SHRINK_LIST)) != (x)) static void d_lru_add(struct dentry *dentry) { D_FLAG_VERIFY(dentry, 0); dentry->d_flags |= DCACHE_LRU_LIST; this_cpu_inc(nr_dentry_unused); WARN_ON_ONCE(!list_lru_add(&dentry->d_sb->s_dentry_lru, &dentry->d_lru)); } static void d_lru_del(struct dentry *dentry) { D_FLAG_VERIFY(dentry, DCACHE_LRU_LIST); dentry->d_flags &= ~DCACHE_LRU_LIST; this_cpu_dec(nr_dentry_unused); WARN_ON_ONCE(!list_lru_del(&dentry->d_sb->s_dentry_lru, &dentry->d_lru)); } static void d_shrink_del(struct dentry *dentry) { D_FLAG_VERIFY(dentry, DCACHE_SHRINK_LIST | DCACHE_LRU_LIST); list_del_init(&dentry->d_lru); dentry->d_flags &= ~(DCACHE_SHRINK_LIST | DCACHE_LRU_LIST); this_cpu_dec(nr_dentry_unused); } static void d_shrink_add(struct dentry *dentry, struct list_head *list) { D_FLAG_VERIFY(dentry, 0); list_add(&dentry->d_lru, list); dentry->d_flags |= DCACHE_SHRINK_LIST | DCACHE_LRU_LIST; this_cpu_inc(nr_dentry_unused); } /* * These can only be called under the global LRU lock, ie during the * callback for freeing the LRU list. "isolate" removes it from the * LRU lists entirely, while shrink_move moves it to the indicated * private list. */ |
3f97b1632
|
419 |
static void d_lru_isolate(struct list_lru_one *lru, struct dentry *dentry) |
89dc77bcd
|
420 421 422 423 |
{ D_FLAG_VERIFY(dentry, DCACHE_LRU_LIST); dentry->d_flags &= ~DCACHE_LRU_LIST; this_cpu_dec(nr_dentry_unused); |
3f97b1632
|
424 |
list_lru_isolate(lru, &dentry->d_lru); |
89dc77bcd
|
425 |
} |
3f97b1632
|
426 427 |
static void d_lru_shrink_move(struct list_lru_one *lru, struct dentry *dentry, struct list_head *list) |
89dc77bcd
|
428 429 430 |
{ D_FLAG_VERIFY(dentry, DCACHE_LRU_LIST); dentry->d_flags |= DCACHE_SHRINK_LIST; |
3f97b1632
|
431 |
list_lru_isolate_move(lru, &dentry->d_lru, list); |
89dc77bcd
|
432 433 434 |
} /* |
f60415675
|
435 |
* dentry_lru_(add|del)_list) must be called with d_lock held. |
da3bbdd46
|
436 437 438 |
*/ static void dentry_lru_add(struct dentry *dentry) { |
89dc77bcd
|
439 440 |
if (unlikely(!(dentry->d_flags & DCACHE_LRU_LIST))) d_lru_add(dentry); |
da3bbdd46
|
441 |
} |
d52b90864
|
442 |
/** |
789680d1e
|
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 |
* d_drop - drop a dentry * @dentry: dentry to drop * * d_drop() unhashes the entry from the parent dentry hashes, so that it won't * be found through a VFS lookup any more. Note that this is different from * deleting the dentry - d_delete will try to mark the dentry negative if * possible, giving a successful _negative_ lookup, while d_drop will * just make the cache lookup fail. * * d_drop() is used mainly for stuff that wants to invalidate a dentry for some * reason (NFS timeouts or autofs deletes). * * __d_drop requires dentry->d_lock. */ void __d_drop(struct dentry *dentry) { |
dea3667bc
|
459 |
if (!d_unhashed(dentry)) { |
b61625d24
|
460 |
struct hlist_bl_head *b; |
7632e465f
|
461 462 463 464 465 466 |
/* * Hashed dentries are normally on the dentry hashtable, * with the exception of those newly allocated by * d_obtain_alias, which are always IS_ROOT: */ if (unlikely(IS_ROOT(dentry))) |
b61625d24
|
467 468 469 470 471 472 473 474 |
b = &dentry->d_sb->s_anon; else b = d_hash(dentry->d_parent, dentry->d_name.hash); hlist_bl_lock(b); __hlist_bl_del(&dentry->d_hash); dentry->d_hash.pprev = NULL; hlist_bl_unlock(b); |
a7c6f571f
|
475 |
dentry_rcuwalk_invalidate(dentry); |
789680d1e
|
476 477 478 479 480 481 |
} } EXPORT_SYMBOL(__d_drop); void d_drop(struct dentry *dentry) { |
789680d1e
|
482 483 484 |
spin_lock(&dentry->d_lock); __d_drop(dentry); spin_unlock(&dentry->d_lock); |
789680d1e
|
485 486 |
} EXPORT_SYMBOL(d_drop); |
e55fd0115
|
487 |
static void __dentry_kill(struct dentry *dentry) |
77812a1ef
|
488 |
{ |
41edf278f
|
489 490 |
struct dentry *parent = NULL; bool can_free = true; |
41edf278f
|
491 |
if (!IS_ROOT(dentry)) |
77812a1ef
|
492 |
parent = dentry->d_parent; |
31e6b01f4
|
493 |
|
0d98439ea
|
494 495 496 497 |
/* * The dentry is now unrecoverably dead to the world. */ lockref_mark_dead(&dentry->d_lockref); |
f0023bc61
|
498 |
/* |
f0023bc61
|
499 500 501 |
* inform the fs via d_prune that this dentry is about to be * unhashed and destroyed. */ |
292662014
|
502 |
if (dentry->d_flags & DCACHE_OP_PRUNE) |
61572bb1f
|
503 |
dentry->d_op->d_prune(dentry); |
01b603519
|
504 505 506 |
if (dentry->d_flags & DCACHE_LRU_LIST) { if (!(dentry->d_flags & DCACHE_SHRINK_LIST)) d_lru_del(dentry); |
01b603519
|
507 |
} |
77812a1ef
|
508 509 |
/* if it was on the hash then remove it */ __d_drop(dentry); |
ca5358ef7
|
510 |
__list_del_entry(&dentry->d_child); |
03b3b889e
|
511 512 513 514 515 516 517 518 519 520 521 522 |
/* * Inform d_walk() that we are no longer attached to the * dentry tree */ dentry->d_flags |= DCACHE_DENTRY_KILLED; if (parent) spin_unlock(&parent->d_lock); dentry_iput(dentry); /* * dentry_iput drops the locks, at which point nobody (except * transient RCU lookups) can reach this dentry. */ |
360f54796
|
523 |
BUG_ON(dentry->d_lockref.count > 0); |
03b3b889e
|
524 525 526 |
this_cpu_dec(nr_dentry); if (dentry->d_op && dentry->d_op->d_release) dentry->d_op->d_release(dentry); |
41edf278f
|
527 528 529 530 531 532 |
spin_lock(&dentry->d_lock); if (dentry->d_flags & DCACHE_SHRINK_LIST) { dentry->d_flags |= DCACHE_MAY_FREE; can_free = false; } spin_unlock(&dentry->d_lock); |
41edf278f
|
533 534 |
if (likely(can_free)) dentry_free(dentry); |
e55fd0115
|
535 536 537 538 539 540 541 542 |
} /* * Finish off a dentry we've decided to kill. * dentry->d_lock must be held, returns with it unlocked. * If ref is non-zero, then decrement the refcount too. * Returns dentry requiring refcount drop, or NULL if we're done. */ |
8cbf74da4
|
543 |
static struct dentry *dentry_kill(struct dentry *dentry) |
e55fd0115
|
544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 |
__releases(dentry->d_lock) { struct inode *inode = dentry->d_inode; struct dentry *parent = NULL; if (inode && unlikely(!spin_trylock(&inode->i_lock))) goto failed; if (!IS_ROOT(dentry)) { parent = dentry->d_parent; if (unlikely(!spin_trylock(&parent->d_lock))) { if (inode) spin_unlock(&inode->i_lock); goto failed; } } __dentry_kill(dentry); |
03b3b889e
|
562 |
return parent; |
e55fd0115
|
563 564 |
failed: |
8cbf74da4
|
565 566 |
spin_unlock(&dentry->d_lock); cpu_relax(); |
e55fd0115
|
567 |
return dentry; /* try again with same dentry */ |
77812a1ef
|
568 |
} |
046b961b4
|
569 570 571 572 573 |
static inline struct dentry *lock_parent(struct dentry *dentry) { struct dentry *parent = dentry->d_parent; if (IS_ROOT(dentry)) return NULL; |
360f54796
|
574 |
if (unlikely(dentry->d_lockref.count < 0)) |
c2338f2dc
|
575 |
return NULL; |
046b961b4
|
576 577 |
if (likely(spin_trylock(&parent->d_lock))) return parent; |
046b961b4
|
578 |
rcu_read_lock(); |
c2338f2dc
|
579 |
spin_unlock(&dentry->d_lock); |
046b961b4
|
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 |
again: parent = ACCESS_ONCE(dentry->d_parent); spin_lock(&parent->d_lock); /* * We can't blindly lock dentry until we are sure * that we won't violate the locking order. * Any changes of dentry->d_parent must have * been done with parent->d_lock held, so * spin_lock() above is enough of a barrier * for checking if it's still our child. */ if (unlikely(parent != dentry->d_parent)) { spin_unlock(&parent->d_lock); goto again; } rcu_read_unlock(); if (parent != dentry) |
9f12600fe
|
597 |
spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); |
046b961b4
|
598 599 600 601 |
else parent = NULL; return parent; } |
360f54796
|
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 |
/* * Try to do a lockless dput(), and return whether that was successful. * * If unsuccessful, we return false, having already taken the dentry lock. * * The caller needs to hold the RCU read lock, so that the dentry is * guaranteed to stay around even if the refcount goes down to zero! */ static inline bool fast_dput(struct dentry *dentry) { int ret; unsigned int d_flags; /* * If we have a d_op->d_delete() operation, we sould not |
75a6f82a0
|
617 |
* let the dentry count go to zero, so use "put_or_lock". |
360f54796
|
618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 |
*/ if (unlikely(dentry->d_flags & DCACHE_OP_DELETE)) return lockref_put_or_lock(&dentry->d_lockref); /* * .. otherwise, we can try to just decrement the * lockref optimistically. */ ret = lockref_put_return(&dentry->d_lockref); /* * If the lockref_put_return() failed due to the lock being held * by somebody else, the fast path has failed. We will need to * get the lock, and then check the count again. */ if (unlikely(ret < 0)) { spin_lock(&dentry->d_lock); if (dentry->d_lockref.count > 1) { dentry->d_lockref.count--; spin_unlock(&dentry->d_lock); return 1; } return 0; } /* * If we weren't the last ref, we're done. */ if (ret) return 1; /* * Careful, careful. The reference count went down * to zero, but we don't hold the dentry lock, so * somebody else could get it again, and do another * dput(), and we need to not race with that. * * However, there is a very special and common case * where we don't care, because there is nothing to * do: the dentry is still hashed, it does not have * a 'delete' op, and it's referenced and already on * the LRU list. * * NOTE! Since we aren't locked, these values are * not "stable". However, it is sufficient that at * some point after we dropped the reference the * dentry was hashed and the flags had the proper * value. Other dentry users may have re-gotten * a reference to the dentry and change that, but * our work is done - we can leave the dentry * around with a zero refcount. */ smp_rmb(); d_flags = ACCESS_ONCE(dentry->d_flags); |
75a6f82a0
|
672 |
d_flags &= DCACHE_REFERENCED | DCACHE_LRU_LIST | DCACHE_DISCONNECTED; |
360f54796
|
673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 |
/* Nothing to do? Dropping the reference was all we needed? */ if (d_flags == (DCACHE_REFERENCED | DCACHE_LRU_LIST) && !d_unhashed(dentry)) return 1; /* * Not the fast normal case? Get the lock. We've already decremented * the refcount, but we'll need to re-check the situation after * getting the lock. */ spin_lock(&dentry->d_lock); /* * Did somebody else grab a reference to it in the meantime, and * we're no longer the last user after all? Alternatively, somebody * else could have killed it and marked it dead. Either way, we * don't need to do anything else. */ if (dentry->d_lockref.count) { spin_unlock(&dentry->d_lock); return 1; } /* * Re-get the reference we optimistically dropped. We hold the * lock, and we just tested that it was zero, so we can just * set it to 1. */ dentry->d_lockref.count = 1; return 0; } |
1da177e4c
|
704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 |
/* * This is dput * * This is complicated by the fact that we do not want to put * dentries that are no longer on any hash chain on the unused * list: we'd much rather just get rid of them immediately. * * However, that implies that we have to traverse the dentry * tree upwards to the parents which might _also_ now be * scheduled for deletion (it may have been only waiting for * its last child to go away). * * This tail recursion is done by hand as we don't want to depend * on the compiler to always get this right (gcc generally doesn't). * Real recursion would eat up our stack space. */ /* * dput - release a dentry * @dentry: dentry to release * * Release a dentry. This will drop the usage count and if appropriate * call the dentry unlink method as well as removing it from the queues and * releasing its resources. If the parent dentries were scheduled for release * they too may now get deleted. |
1da177e4c
|
729 |
*/ |
1da177e4c
|
730 731 |
void dput(struct dentry *dentry) { |
8aab6a273
|
732 |
if (unlikely(!dentry)) |
1da177e4c
|
733 734 735 |
return; repeat: |
360f54796
|
736 737 738 |
rcu_read_lock(); if (likely(fast_dput(dentry))) { rcu_read_unlock(); |
1da177e4c
|
739 |
return; |
360f54796
|
740 741 742 743 |
} /* Slow case: now with the dentry lock held */ rcu_read_unlock(); |
1da177e4c
|
744 |
|
8aab6a273
|
745 746 747 |
/* Unreachable? Get rid of it */ if (unlikely(d_unhashed(dentry))) goto kill_it; |
75a6f82a0
|
748 749 |
if (unlikely(dentry->d_flags & DCACHE_DISCONNECTED)) goto kill_it; |
8aab6a273
|
750 |
if (unlikely(dentry->d_flags & DCACHE_OP_DELETE)) { |
1da177e4c
|
751 |
if (dentry->d_op->d_delete(dentry)) |
61f3dee4a
|
752 |
goto kill_it; |
1da177e4c
|
753 |
} |
265ac9023
|
754 |
|
358eec182
|
755 756 |
if (!(dentry->d_flags & DCACHE_REFERENCED)) dentry->d_flags |= DCACHE_REFERENCED; |
a4633357a
|
757 |
dentry_lru_add(dentry); |
265ac9023
|
758 |
|
98474236f
|
759 |
dentry->d_lockref.count--; |
61f3dee4a
|
760 |
spin_unlock(&dentry->d_lock); |
1da177e4c
|
761 |
return; |
d52b90864
|
762 |
kill_it: |
8cbf74da4
|
763 |
dentry = dentry_kill(dentry); |
d52b90864
|
764 765 |
if (dentry) goto repeat; |
1da177e4c
|
766 |
} |
ec4f86059
|
767 |
EXPORT_SYMBOL(dput); |
1da177e4c
|
768 |
|
1da177e4c
|
769 |
|
b5c84bf6f
|
770 |
/* This must be called with d_lock held */ |
dc0474be3
|
771 |
static inline void __dget_dlock(struct dentry *dentry) |
230445078
|
772 |
{ |
98474236f
|
773 |
dentry->d_lockref.count++; |
230445078
|
774 |
} |
dc0474be3
|
775 |
static inline void __dget(struct dentry *dentry) |
1da177e4c
|
776 |
{ |
98474236f
|
777 |
lockref_get(&dentry->d_lockref); |
1da177e4c
|
778 |
} |
b7ab39f63
|
779 780 |
struct dentry *dget_parent(struct dentry *dentry) { |
df3d0bbcd
|
781 |
int gotref; |
b7ab39f63
|
782 |
struct dentry *ret; |
df3d0bbcd
|
783 784 785 786 787 788 789 790 791 792 793 794 795 |
/* * Do optimistic parent lookup without any * locking. */ rcu_read_lock(); ret = ACCESS_ONCE(dentry->d_parent); gotref = lockref_get_not_zero(&ret->d_lockref); rcu_read_unlock(); if (likely(gotref)) { if (likely(ret == ACCESS_ONCE(dentry->d_parent))) return ret; dput(ret); } |
b7ab39f63
|
796 |
repeat: |
a734eb458
|
797 798 799 800 801 |
/* * Don't need rcu_dereference because we re-check it was correct under * the lock. */ rcu_read_lock(); |
b7ab39f63
|
802 |
ret = dentry->d_parent; |
a734eb458
|
803 804 805 806 |
spin_lock(&ret->d_lock); if (unlikely(ret != dentry->d_parent)) { spin_unlock(&ret->d_lock); rcu_read_unlock(); |
b7ab39f63
|
807 808 |
goto repeat; } |
a734eb458
|
809 |
rcu_read_unlock(); |
98474236f
|
810 811 |
BUG_ON(!ret->d_lockref.count); ret->d_lockref.count++; |
b7ab39f63
|
812 |
spin_unlock(&ret->d_lock); |
b7ab39f63
|
813 814 815 |
return ret; } EXPORT_SYMBOL(dget_parent); |
1da177e4c
|
816 817 818 |
/** * d_find_alias - grab a hashed alias of inode * @inode: inode in question |
1da177e4c
|
819 820 821 822 823 |
* * If inode has a hashed alias, or is a directory and has any alias, * acquire the reference to alias and return it. Otherwise return NULL. * Notice that if inode is a directory there can be only one alias and * it can be unhashed only if it has no children, or if it is the root |
3ccb354d6
|
824 825 |
* of a filesystem, or if the directory was renamed and d_revalidate * was the first vfs operation to notice. |
1da177e4c
|
826 |
* |
21c0d8fdd
|
827 |
* If the inode has an IS_ROOT, DCACHE_DISCONNECTED alias, then prefer |
52ed46f0f
|
828 |
* any other hashed alias over that one. |
1da177e4c
|
829 |
*/ |
52ed46f0f
|
830 |
static struct dentry *__d_find_alias(struct inode *inode) |
1da177e4c
|
831 |
{ |
da5029563
|
832 |
struct dentry *alias, *discon_alias; |
1da177e4c
|
833 |
|
da5029563
|
834 835 |
again: discon_alias = NULL; |
946e51f2b
|
836 |
hlist_for_each_entry(alias, &inode->i_dentry, d_u.d_alias) { |
da5029563
|
837 |
spin_lock(&alias->d_lock); |
1da177e4c
|
838 |
if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) { |
21c0d8fdd
|
839 |
if (IS_ROOT(alias) && |
da5029563
|
840 |
(alias->d_flags & DCACHE_DISCONNECTED)) { |
1da177e4c
|
841 |
discon_alias = alias; |
52ed46f0f
|
842 |
} else { |
dc0474be3
|
843 |
__dget_dlock(alias); |
da5029563
|
844 845 846 847 848 849 850 851 852 853 |
spin_unlock(&alias->d_lock); return alias; } } spin_unlock(&alias->d_lock); } if (discon_alias) { alias = discon_alias; spin_lock(&alias->d_lock); if (S_ISDIR(inode->i_mode) || !d_unhashed(alias)) { |
8d80d7dab
|
854 855 856 |
__dget_dlock(alias); spin_unlock(&alias->d_lock); return alias; |
1da177e4c
|
857 |
} |
da5029563
|
858 859 |
spin_unlock(&alias->d_lock); goto again; |
1da177e4c
|
860 |
} |
da5029563
|
861 |
return NULL; |
1da177e4c
|
862 |
} |
da5029563
|
863 |
struct dentry *d_find_alias(struct inode *inode) |
1da177e4c
|
864 |
{ |
214fda1f6
|
865 |
struct dentry *de = NULL; |
b3d9b7a3c
|
866 |
if (!hlist_empty(&inode->i_dentry)) { |
873feea09
|
867 |
spin_lock(&inode->i_lock); |
52ed46f0f
|
868 |
de = __d_find_alias(inode); |
873feea09
|
869 |
spin_unlock(&inode->i_lock); |
214fda1f6
|
870 |
} |
1da177e4c
|
871 872 |
return de; } |
ec4f86059
|
873 |
EXPORT_SYMBOL(d_find_alias); |
1da177e4c
|
874 875 876 877 878 879 880 |
/* * Try to kill dentries associated with this inode. * WARNING: you must own a reference to inode. */ void d_prune_aliases(struct inode *inode) { |
0cdca3f98
|
881 |
struct dentry *dentry; |
1da177e4c
|
882 |
restart: |
873feea09
|
883 |
spin_lock(&inode->i_lock); |
946e51f2b
|
884 |
hlist_for_each_entry(dentry, &inode->i_dentry, d_u.d_alias) { |
1da177e4c
|
885 |
spin_lock(&dentry->d_lock); |
98474236f
|
886 |
if (!dentry->d_lockref.count) { |
29355c390
|
887 888 889 |
struct dentry *parent = lock_parent(dentry); if (likely(!dentry->d_lockref.count)) { __dentry_kill(dentry); |
4a7795d35
|
890 |
dput(parent); |
29355c390
|
891 892 893 894 |
goto restart; } if (parent) spin_unlock(&parent->d_lock); |
1da177e4c
|
895 896 897 |
} spin_unlock(&dentry->d_lock); } |
873feea09
|
898 |
spin_unlock(&inode->i_lock); |
1da177e4c
|
899 |
} |
ec4f86059
|
900 |
EXPORT_SYMBOL(d_prune_aliases); |
1da177e4c
|
901 |
|
3049cfe24
|
902 |
static void shrink_dentry_list(struct list_head *list) |
1da177e4c
|
903 |
{ |
5c47e6d0a
|
904 |
struct dentry *dentry, *parent; |
da3bbdd46
|
905 |
|
60942f2f2
|
906 |
while (!list_empty(list)) { |
ff2fde992
|
907 |
struct inode *inode; |
60942f2f2
|
908 |
dentry = list_entry(list->prev, struct dentry, d_lru); |
ec33679d7
|
909 |
spin_lock(&dentry->d_lock); |
046b961b4
|
910 |
parent = lock_parent(dentry); |
1da177e4c
|
911 |
/* |
dd1f6b2e4
|
912 913 914 915 |
* The dispose list is isolated and dentries are not accounted * to the LRU here, so we can simply remove it from the list * here regardless of whether it is referenced or not. */ |
89dc77bcd
|
916 |
d_shrink_del(dentry); |
dd1f6b2e4
|
917 918 |
/* |
1da177e4c
|
919 |
* We found an inuse dentry which was not removed from |
dd1f6b2e4
|
920 |
* the LRU because of laziness during lookup. Do not free it. |
1da177e4c
|
921 |
*/ |
360f54796
|
922 |
if (dentry->d_lockref.count > 0) { |
da3bbdd46
|
923 |
spin_unlock(&dentry->d_lock); |
046b961b4
|
924 925 |
if (parent) spin_unlock(&parent->d_lock); |
1da177e4c
|
926 927 |
continue; } |
77812a1ef
|
928 |
|
64fd72e0a
|
929 930 931 932 |
if (unlikely(dentry->d_flags & DCACHE_DENTRY_KILLED)) { bool can_free = dentry->d_flags & DCACHE_MAY_FREE; spin_unlock(&dentry->d_lock); |
046b961b4
|
933 934 |
if (parent) spin_unlock(&parent->d_lock); |
64fd72e0a
|
935 936 937 938 |
if (can_free) dentry_free(dentry); continue; } |
ff2fde992
|
939 940 |
inode = dentry->d_inode; if (inode && unlikely(!spin_trylock(&inode->i_lock))) { |
89dc77bcd
|
941 |
d_shrink_add(dentry, list); |
dd1f6b2e4
|
942 |
spin_unlock(&dentry->d_lock); |
046b961b4
|
943 944 |
if (parent) spin_unlock(&parent->d_lock); |
5c47e6d0a
|
945 |
continue; |
dd1f6b2e4
|
946 |
} |
ff2fde992
|
947 |
|
ff2fde992
|
948 |
__dentry_kill(dentry); |
046b961b4
|
949 |
|
5c47e6d0a
|
950 951 952 953 954 955 956 |
/* * We need to prune ancestors too. This is necessary to prevent * quadratic behavior of shrink_dcache_parent(), but is also * expected to be beneficial in reducing dentry cache * fragmentation. */ dentry = parent; |
b2b80195d
|
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 |
while (dentry && !lockref_put_or_lock(&dentry->d_lockref)) { parent = lock_parent(dentry); if (dentry->d_lockref.count != 1) { dentry->d_lockref.count--; spin_unlock(&dentry->d_lock); if (parent) spin_unlock(&parent->d_lock); break; } inode = dentry->d_inode; /* can't be NULL */ if (unlikely(!spin_trylock(&inode->i_lock))) { spin_unlock(&dentry->d_lock); if (parent) spin_unlock(&parent->d_lock); cpu_relax(); continue; } __dentry_kill(dentry); dentry = parent; } |
da3bbdd46
|
977 |
} |
3049cfe24
|
978 |
} |
3f97b1632
|
979 980 |
static enum lru_status dentry_lru_isolate(struct list_head *item, struct list_lru_one *lru, spinlock_t *lru_lock, void *arg) |
f60415675
|
981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 |
{ struct list_head *freeable = arg; struct dentry *dentry = container_of(item, struct dentry, d_lru); /* * we are inverting the lru lock/dentry->d_lock here, * so use a trylock. If we fail to get the lock, just skip * it */ if (!spin_trylock(&dentry->d_lock)) return LRU_SKIP; /* * Referenced dentries are still in use. If they have active * counts, just remove them from the LRU. Otherwise give them * another pass through the LRU. */ if (dentry->d_lockref.count) { |
3f97b1632
|
1000 |
d_lru_isolate(lru, dentry); |
f60415675
|
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 |
spin_unlock(&dentry->d_lock); return LRU_REMOVED; } if (dentry->d_flags & DCACHE_REFERENCED) { dentry->d_flags &= ~DCACHE_REFERENCED; spin_unlock(&dentry->d_lock); /* * The list move itself will be made by the common LRU code. At * this point, we've dropped the dentry->d_lock but keep the * lru lock. This is safe to do, since every list movement is * protected by the lru lock even if both locks are held. * * This is guaranteed by the fact that all LRU management * functions are intermediated by the LRU API calls like * list_lru_add and list_lru_del. List movement in this file * only ever occur through this functions or through callbacks * like this one, that are called from the LRU API. * * The only exceptions to this are functions like * shrink_dentry_list, and code that first checks for the * DCACHE_SHRINK_LIST flag. Those are guaranteed to be * operating only with stack provided lists after they are * properly isolated from the main list. It is thus, always a * local access. */ return LRU_ROTATE; } |
3f97b1632
|
1030 |
d_lru_shrink_move(lru, dentry, freeable); |
f60415675
|
1031 1032 1033 1034 |
spin_unlock(&dentry->d_lock); return LRU_REMOVED; } |
3049cfe24
|
1035 |
/** |
b48f03b31
|
1036 1037 |
* prune_dcache_sb - shrink the dcache * @sb: superblock |
503c358cf
|
1038 |
* @sc: shrink control, passed to list_lru_shrink_walk() |
b48f03b31
|
1039 |
* |
503c358cf
|
1040 1041 |
* Attempt to shrink the superblock dcache LRU by @sc->nr_to_scan entries. This * is done when we need more memory and called from the superblock shrinker |
b48f03b31
|
1042 |
* function. |
3049cfe24
|
1043 |
* |
b48f03b31
|
1044 1045 |
* This function may fail to free any resources if all the dentries are in * use. |
3049cfe24
|
1046 |
*/ |
503c358cf
|
1047 |
long prune_dcache_sb(struct super_block *sb, struct shrink_control *sc) |
3049cfe24
|
1048 |
{ |
f60415675
|
1049 1050 |
LIST_HEAD(dispose); long freed; |
3049cfe24
|
1051 |
|
503c358cf
|
1052 1053 |
freed = list_lru_shrink_walk(&sb->s_dentry_lru, sc, dentry_lru_isolate, &dispose); |
f60415675
|
1054 |
shrink_dentry_list(&dispose); |
0a234c6dc
|
1055 |
return freed; |
da3bbdd46
|
1056 |
} |
230445078
|
1057 |
|
4e717f5c1
|
1058 |
static enum lru_status dentry_lru_isolate_shrink(struct list_head *item, |
3f97b1632
|
1059 |
struct list_lru_one *lru, spinlock_t *lru_lock, void *arg) |
dd1f6b2e4
|
1060 |
{ |
4e717f5c1
|
1061 1062 |
struct list_head *freeable = arg; struct dentry *dentry = container_of(item, struct dentry, d_lru); |
dd1f6b2e4
|
1063 |
|
4e717f5c1
|
1064 1065 1066 1067 1068 1069 1070 |
/* * we are inverting the lru lock/dentry->d_lock here, * so use a trylock. If we fail to get the lock, just skip * it */ if (!spin_trylock(&dentry->d_lock)) return LRU_SKIP; |
3f97b1632
|
1071 |
d_lru_shrink_move(lru, dentry, freeable); |
4e717f5c1
|
1072 |
spin_unlock(&dentry->d_lock); |
ec33679d7
|
1073 |
|
4e717f5c1
|
1074 |
return LRU_REMOVED; |
da3bbdd46
|
1075 |
} |
4e717f5c1
|
1076 |
|
da3bbdd46
|
1077 |
/** |
1da177e4c
|
1078 1079 1080 |
* shrink_dcache_sb - shrink dcache for a superblock * @sb: superblock * |
3049cfe24
|
1081 1082 |
* Shrink the dcache for the specified super block. This is used to free * the dcache before unmounting a file system. |
1da177e4c
|
1083 |
*/ |
3049cfe24
|
1084 |
void shrink_dcache_sb(struct super_block *sb) |
1da177e4c
|
1085 |
{ |
4e717f5c1
|
1086 1087 1088 1089 1090 1091 1092 |
long freed; do { LIST_HEAD(dispose); freed = list_lru_walk(&sb->s_dentry_lru, dentry_lru_isolate_shrink, &dispose, UINT_MAX); |
3049cfe24
|
1093 |
|
4e717f5c1
|
1094 1095 1096 |
this_cpu_sub(nr_dentry_unused, freed); shrink_dentry_list(&dispose); } while (freed > 0); |
1da177e4c
|
1097 |
} |
ec4f86059
|
1098 |
EXPORT_SYMBOL(shrink_dcache_sb); |
1da177e4c
|
1099 |
|
db14fc3ab
|
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 |
/** * enum d_walk_ret - action to talke during tree walk * @D_WALK_CONTINUE: contrinue walk * @D_WALK_QUIT: quit walk * @D_WALK_NORETRY: quit when retry is needed * @D_WALK_SKIP: skip this dentry and its children */ enum d_walk_ret { D_WALK_CONTINUE, D_WALK_QUIT, D_WALK_NORETRY, D_WALK_SKIP, }; |
c826cb7df
|
1113 |
|
1da177e4c
|
1114 |
/** |
db14fc3ab
|
1115 1116 1117 1118 1119 |
* d_walk - walk the dentry tree * @parent: start of walk * @data: data passed to @enter() and @finish() * @enter: callback when first entering the dentry * @finish: callback when successfully finished the walk |
1da177e4c
|
1120 |
* |
db14fc3ab
|
1121 |
* The @enter() and @finish() callbacks are called with d_lock held. |
1da177e4c
|
1122 |
*/ |
db14fc3ab
|
1123 1124 1125 |
static void d_walk(struct dentry *parent, void *data, enum d_walk_ret (*enter)(void *, struct dentry *), void (*finish)(void *)) |
1da177e4c
|
1126 |
{ |
949854d02
|
1127 |
struct dentry *this_parent; |
1da177e4c
|
1128 |
struct list_head *next; |
48f5ec21d
|
1129 |
unsigned seq = 0; |
db14fc3ab
|
1130 1131 |
enum d_walk_ret ret; bool retry = true; |
949854d02
|
1132 |
|
58db63d08
|
1133 |
again: |
48f5ec21d
|
1134 |
read_seqbegin_or_lock(&rename_lock, &seq); |
58db63d08
|
1135 |
this_parent = parent; |
2fd6b7f50
|
1136 |
spin_lock(&this_parent->d_lock); |
db14fc3ab
|
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 |
ret = enter(data, this_parent); switch (ret) { case D_WALK_CONTINUE: break; case D_WALK_QUIT: case D_WALK_SKIP: goto out_unlock; case D_WALK_NORETRY: retry = false; break; } |
1da177e4c
|
1149 1150 1151 1152 1153 |
repeat: next = this_parent->d_subdirs.next; resume: while (next != &this_parent->d_subdirs) { struct list_head *tmp = next; |
946e51f2b
|
1154 |
struct dentry *dentry = list_entry(tmp, struct dentry, d_child); |
1da177e4c
|
1155 |
next = tmp->next; |
2fd6b7f50
|
1156 1157 |
spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); |
db14fc3ab
|
1158 1159 1160 1161 1162 1163 |
ret = enter(data, dentry); switch (ret) { case D_WALK_CONTINUE: break; case D_WALK_QUIT: |
2fd6b7f50
|
1164 |
spin_unlock(&dentry->d_lock); |
db14fc3ab
|
1165 1166 1167 1168 1169 1170 1171 |
goto out_unlock; case D_WALK_NORETRY: retry = false; break; case D_WALK_SKIP: spin_unlock(&dentry->d_lock); continue; |
2fd6b7f50
|
1172 |
} |
db14fc3ab
|
1173 |
|
1da177e4c
|
1174 |
if (!list_empty(&dentry->d_subdirs)) { |
2fd6b7f50
|
1175 1176 |
spin_unlock(&this_parent->d_lock); spin_release(&dentry->d_lock.dep_map, 1, _RET_IP_); |
1da177e4c
|
1177 |
this_parent = dentry; |
2fd6b7f50
|
1178 |
spin_acquire(&this_parent->d_lock.dep_map, 0, 1, _RET_IP_); |
1da177e4c
|
1179 1180 |
goto repeat; } |
2fd6b7f50
|
1181 |
spin_unlock(&dentry->d_lock); |
1da177e4c
|
1182 1183 1184 1185 |
} /* * All done at this level ... ascend and resume the search. */ |
ca5358ef7
|
1186 1187 |
rcu_read_lock(); ascend: |
1da177e4c
|
1188 |
if (this_parent != parent) { |
c826cb7df
|
1189 |
struct dentry *child = this_parent; |
31dec1327
|
1190 |
this_parent = child->d_parent; |
31dec1327
|
1191 1192 |
spin_unlock(&child->d_lock); spin_lock(&this_parent->d_lock); |
ca5358ef7
|
1193 1194 |
/* might go back up the wrong parent if we have had a rename. */ if (need_seqretry(&rename_lock, seq)) |
949854d02
|
1195 |
goto rename_retry; |
2159184ea
|
1196 1197 1198 |
/* go into the first sibling still alive */ do { next = child->d_child.next; |
ca5358ef7
|
1199 1200 1201 |
if (next == &this_parent->d_subdirs) goto ascend; child = list_entry(next, struct dentry, d_child); |
2159184ea
|
1202 |
} while (unlikely(child->d_flags & DCACHE_DENTRY_KILLED)); |
31dec1327
|
1203 |
rcu_read_unlock(); |
1da177e4c
|
1204 1205 |
goto resume; } |
ca5358ef7
|
1206 |
if (need_seqretry(&rename_lock, seq)) |
949854d02
|
1207 |
goto rename_retry; |
ca5358ef7
|
1208 |
rcu_read_unlock(); |
db14fc3ab
|
1209 1210 1211 1212 1213 |
if (finish) finish(data); out_unlock: spin_unlock(&this_parent->d_lock); |
48f5ec21d
|
1214 |
done_seqretry(&rename_lock, seq); |
db14fc3ab
|
1215 |
return; |
58db63d08
|
1216 1217 |
rename_retry: |
ca5358ef7
|
1218 1219 1220 |
spin_unlock(&this_parent->d_lock); rcu_read_unlock(); BUG_ON(seq & 1); |
db14fc3ab
|
1221 1222 |
if (!retry) return; |
48f5ec21d
|
1223 |
seq = 1; |
58db63d08
|
1224 |
goto again; |
1da177e4c
|
1225 |
} |
db14fc3ab
|
1226 1227 1228 1229 1230 1231 |
/* * Search for at least 1 mount point in the dentry's subdirs. * We descend to the next level whenever the d_subdirs * list is non-empty and continue searching. */ |
db14fc3ab
|
1232 1233 1234 1235 1236 1237 1238 1239 1240 |
static enum d_walk_ret check_mount(void *data, struct dentry *dentry) { int *ret = data; if (d_mountpoint(dentry)) { *ret = 1; return D_WALK_QUIT; } return D_WALK_CONTINUE; } |
69c88dc7d
|
1241 1242 1243 1244 1245 1246 1247 |
/** * have_submounts - check for mounts over a dentry * @parent: dentry to check. * * Return true if the parent or its subdirectories contain * a mount point */ |
db14fc3ab
|
1248 1249 1250 1251 1252 1253 1254 1255 |
int have_submounts(struct dentry *parent) { int ret = 0; d_walk(parent, &ret, check_mount, NULL); return ret; } |
ec4f86059
|
1256 |
EXPORT_SYMBOL(have_submounts); |
1da177e4c
|
1257 1258 |
/* |
eed810076
|
1259 1260 1261 1262 |
* Called by mount code to set a mountpoint and check if the mountpoint is * reachable (e.g. NFS can unhash a directory dentry and then the complete * subtree can become unreachable). * |
1ffe46d11
|
1263 |
* Only one of d_invalidate() and d_set_mounted() must succeed. For |
eed810076
|
1264 1265 1266 1267 1268 1269 1270 1271 |
* this reason take rename_lock and d_lock on dentry and ancestors. */ int d_set_mounted(struct dentry *dentry) { struct dentry *p; int ret = -ENOENT; write_seqlock(&rename_lock); for (p = dentry->d_parent; !IS_ROOT(p); p = p->d_parent) { |
1ffe46d11
|
1272 |
/* Need exclusion wrt. d_invalidate() */ |
eed810076
|
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 |
spin_lock(&p->d_lock); if (unlikely(d_unhashed(p))) { spin_unlock(&p->d_lock); goto out; } spin_unlock(&p->d_lock); } spin_lock(&dentry->d_lock); if (!d_unlinked(dentry)) { dentry->d_flags |= DCACHE_MOUNTED; ret = 0; } spin_unlock(&dentry->d_lock); out: write_sequnlock(&rename_lock); return ret; } /* |
fd5179094
|
1292 |
* Search the dentry child list of the specified parent, |
1da177e4c
|
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 |
* and move any unused dentries to the end of the unused * list for prune_dcache(). We descend to the next level * whenever the d_subdirs list is non-empty and continue * searching. * * It returns zero iff there are no unused children, * otherwise it returns the number of children moved to * the end of the unused list. This may not be the total * number of unused children, because select_parent can * drop the lock and return early due to latency * constraints. */ |
1da177e4c
|
1305 |
|
db14fc3ab
|
1306 1307 1308 1309 1310 |
struct select_data { struct dentry *start; struct list_head dispose; int found; }; |
230445078
|
1311 |
|
db14fc3ab
|
1312 1313 1314 1315 |
static enum d_walk_ret select_collect(void *_data, struct dentry *dentry) { struct select_data *data = _data; enum d_walk_ret ret = D_WALK_CONTINUE; |
1da177e4c
|
1316 |
|
db14fc3ab
|
1317 1318 |
if (data->start == dentry) goto out; |
2fd6b7f50
|
1319 |
|
fe91522a7
|
1320 |
if (dentry->d_flags & DCACHE_SHRINK_LIST) { |
db14fc3ab
|
1321 |
data->found++; |
fe91522a7
|
1322 1323 1324 1325 1326 1327 1328 |
} else { if (dentry->d_flags & DCACHE_LRU_LIST) d_lru_del(dentry); if (!dentry->d_lockref.count) { d_shrink_add(dentry, &data->dispose); data->found++; } |
1da177e4c
|
1329 |
} |
db14fc3ab
|
1330 1331 1332 1333 1334 |
/* * We can return to the caller if we have found some (this * ensures forward progress). We'll be coming back to find * the rest. */ |
fe91522a7
|
1335 1336 |
if (!list_empty(&data->dispose)) ret = need_resched() ? D_WALK_QUIT : D_WALK_NORETRY; |
1da177e4c
|
1337 |
out: |
db14fc3ab
|
1338 |
return ret; |
1da177e4c
|
1339 1340 1341 1342 1343 1344 1345 1346 |
} /** * shrink_dcache_parent - prune dcache * @parent: parent of entries to prune * * Prune the dcache to remove unused children of the parent dentry. */ |
db14fc3ab
|
1347 |
void shrink_dcache_parent(struct dentry *parent) |
1da177e4c
|
1348 |
{ |
db14fc3ab
|
1349 1350 |
for (;;) { struct select_data data; |
1da177e4c
|
1351 |
|
db14fc3ab
|
1352 1353 1354 1355 1356 1357 1358 1359 1360 |
INIT_LIST_HEAD(&data.dispose); data.start = parent; data.found = 0; d_walk(parent, &data, select_collect, NULL); if (!data.found) break; shrink_dentry_list(&data.dispose); |
421348f1c
|
1361 1362 |
cond_resched(); } |
1da177e4c
|
1363 |
} |
ec4f86059
|
1364 |
EXPORT_SYMBOL(shrink_dcache_parent); |
1da177e4c
|
1365 |
|
9c8c10e26
|
1366 |
static enum d_walk_ret umount_check(void *_data, struct dentry *dentry) |
42c326082
|
1367 |
{ |
9c8c10e26
|
1368 1369 1370 |
/* it has busy descendents; complain about those instead */ if (!list_empty(&dentry->d_subdirs)) return D_WALK_CONTINUE; |
42c326082
|
1371 |
|
9c8c10e26
|
1372 1373 1374 1375 1376 1377 1378 |
/* root with refcount 1 is fine */ if (dentry == _data && dentry->d_lockref.count == 1) return D_WALK_CONTINUE; printk(KERN_ERR "BUG: Dentry %p{i=%lx,n=%pd} " " still in use (%d) [unmount of %s %s] ", |
42c326082
|
1379 1380 1381 |
dentry, dentry->d_inode ? dentry->d_inode->i_ino : 0UL, |
9c8c10e26
|
1382 |
dentry, |
42c326082
|
1383 1384 1385 |
dentry->d_lockref.count, dentry->d_sb->s_type->name, dentry->d_sb->s_id); |
9c8c10e26
|
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 |
WARN_ON(1); return D_WALK_CONTINUE; } static void do_one_tree(struct dentry *dentry) { shrink_dcache_parent(dentry); d_walk(dentry, dentry, umount_check, NULL); d_drop(dentry); dput(dentry); |
42c326082
|
1396 1397 1398 1399 1400 1401 1402 1403 |
} /* * destroy the dentries attached to a superblock on unmounting */ void shrink_dcache_for_umount(struct super_block *sb) { struct dentry *dentry; |
9c8c10e26
|
1404 |
WARN(down_read_trylock(&sb->s_umount), "s_umount should've been locked"); |
42c326082
|
1405 1406 1407 |
dentry = sb->s_root; sb->s_root = NULL; |
9c8c10e26
|
1408 |
do_one_tree(dentry); |
42c326082
|
1409 1410 |
while (!hlist_bl_empty(&sb->s_anon)) { |
9c8c10e26
|
1411 1412 |
dentry = dget(hlist_bl_entry(hlist_bl_first(&sb->s_anon), struct dentry, d_hash)); do_one_tree(dentry); |
42c326082
|
1413 1414 |
} } |
8ed936b56
|
1415 1416 1417 1418 1419 |
struct detach_data { struct select_data select; struct dentry *mountpoint; }; static enum d_walk_ret detach_and_collect(void *_data, struct dentry *dentry) |
848ac114e
|
1420 |
{ |
8ed936b56
|
1421 |
struct detach_data *data = _data; |
848ac114e
|
1422 1423 |
if (d_mountpoint(dentry)) { |
8ed936b56
|
1424 1425 |
__dget_dlock(dentry); data->mountpoint = dentry; |
848ac114e
|
1426 1427 |
return D_WALK_QUIT; } |
8ed936b56
|
1428 |
return select_collect(&data->select, dentry); |
848ac114e
|
1429 1430 1431 1432 |
} static void check_and_drop(void *_data) { |
8ed936b56
|
1433 |
struct detach_data *data = _data; |
848ac114e
|
1434 |
|
8ed936b56
|
1435 1436 |
if (!data->mountpoint && !data->select.found) __d_drop(data->select.start); |
848ac114e
|
1437 1438 1439 |
} /** |
1ffe46d11
|
1440 1441 1442 |
* d_invalidate - detach submounts, prune dcache, and drop * @dentry: dentry to invalidate (aka detach, prune and drop) * |
1ffe46d11
|
1443 |
* no dcache lock. |
848ac114e
|
1444 |
* |
8ed936b56
|
1445 1446 1447 |
* The final d_drop is done as an atomic operation relative to * rename_lock ensuring there are no races with d_set_mounted. This * ensures there are no unhashed dentries on the path to a mountpoint. |
848ac114e
|
1448 |
*/ |
5542aa2fa
|
1449 |
void d_invalidate(struct dentry *dentry) |
848ac114e
|
1450 |
{ |
1ffe46d11
|
1451 1452 1453 1454 1455 1456 |
/* * If it's already been dropped, return OK. */ spin_lock(&dentry->d_lock); if (d_unhashed(dentry)) { spin_unlock(&dentry->d_lock); |
5542aa2fa
|
1457 |
return; |
1ffe46d11
|
1458 1459 |
} spin_unlock(&dentry->d_lock); |
848ac114e
|
1460 1461 1462 |
/* Negative dentries can be dropped without further checks */ if (!dentry->d_inode) { d_drop(dentry); |
5542aa2fa
|
1463 |
return; |
848ac114e
|
1464 1465 1466 |
} for (;;) { |
8ed936b56
|
1467 |
struct detach_data data; |
848ac114e
|
1468 |
|
8ed936b56
|
1469 1470 1471 1472 1473 1474 |
data.mountpoint = NULL; INIT_LIST_HEAD(&data.select.dispose); data.select.start = dentry; data.select.found = 0; d_walk(dentry, &data, detach_and_collect, check_and_drop); |
848ac114e
|
1475 |
|
8ed936b56
|
1476 1477 |
if (data.select.found) shrink_dentry_list(&data.select.dispose); |
848ac114e
|
1478 |
|
8ed936b56
|
1479 1480 1481 1482 |
if (data.mountpoint) { detach_mounts(data.mountpoint); dput(data.mountpoint); } |
848ac114e
|
1483 |
|
8ed936b56
|
1484 |
if (!data.mountpoint && !data.select.found) |
848ac114e
|
1485 1486 1487 1488 |
break; cond_resched(); } |
848ac114e
|
1489 |
} |
1ffe46d11
|
1490 |
EXPORT_SYMBOL(d_invalidate); |
848ac114e
|
1491 |
|
1da177e4c
|
1492 |
/** |
a4464dbc0
|
1493 1494 |
* __d_alloc - allocate a dcache entry * @sb: filesystem it will belong to |
1da177e4c
|
1495 1496 1497 1498 1499 1500 1501 |
* @name: qstr of the name * * Allocates a dentry. It returns %NULL if there is insufficient memory * available. On a success the dentry is returned. The name passed in is * copied and the copy passed in may be reused after this call. */ |
a4464dbc0
|
1502 |
struct dentry *__d_alloc(struct super_block *sb, const struct qstr *name) |
1da177e4c
|
1503 1504 1505 |
{ struct dentry *dentry; char *dname; |
e12ba74d8
|
1506 |
dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL); |
1da177e4c
|
1507 1508 |
if (!dentry) return NULL; |
6326c71fd
|
1509 1510 1511 1512 1513 1514 1515 |
/* * We guarantee that the inline name is always NUL-terminated. * This way the memcpy() done by the name switching in rename * will still always have a NUL at the end, even if we might * be overwriting an internal NUL character */ dentry->d_iname[DNAME_INLINE_LEN-1] = 0; |
1da177e4c
|
1516 |
if (name->len > DNAME_INLINE_LEN-1) { |
8d85b4845
|
1517 1518 1519 |
size_t size = offsetof(struct external_name, name[1]); struct external_name *p = kmalloc(size + name->len, GFP_KERNEL); if (!p) { |
1da177e4c
|
1520 1521 1522 |
kmem_cache_free(dentry_cache, dentry); return NULL; } |
8d85b4845
|
1523 1524 |
atomic_set(&p->u.count, 1); dname = p->name; |
df4c0e36f
|
1525 1526 1527 |
if (IS_ENABLED(CONFIG_DCACHE_WORD_ACCESS)) kasan_unpoison_shadow(dname, round_up(name->len + 1, sizeof(unsigned long))); |
1da177e4c
|
1528 1529 1530 |
} else { dname = dentry->d_iname; } |
1da177e4c
|
1531 1532 1533 1534 1535 |
dentry->d_name.len = name->len; dentry->d_name.hash = name->hash; memcpy(dname, name->name, name->len); dname[name->len] = 0; |
6326c71fd
|
1536 1537 1538 |
/* Make sure we always see the terminating NUL character */ smp_wmb(); dentry->d_name.name = dname; |
98474236f
|
1539 |
dentry->d_lockref.count = 1; |
dea3667bc
|
1540 |
dentry->d_flags = 0; |
1da177e4c
|
1541 |
spin_lock_init(&dentry->d_lock); |
31e6b01f4
|
1542 |
seqcount_init(&dentry->d_seq); |
1da177e4c
|
1543 |
dentry->d_inode = NULL; |
a4464dbc0
|
1544 1545 |
dentry->d_parent = dentry; dentry->d_sb = sb; |
1da177e4c
|
1546 1547 |
dentry->d_op = NULL; dentry->d_fsdata = NULL; |
ceb5bdc2d
|
1548 |
INIT_HLIST_BL_NODE(&dentry->d_hash); |
1da177e4c
|
1549 1550 |
INIT_LIST_HEAD(&dentry->d_lru); INIT_LIST_HEAD(&dentry->d_subdirs); |
946e51f2b
|
1551 1552 |
INIT_HLIST_NODE(&dentry->d_u.d_alias); INIT_LIST_HEAD(&dentry->d_child); |
a4464dbc0
|
1553 |
d_set_d_op(dentry, dentry->d_sb->s_d_op); |
1da177e4c
|
1554 |
|
3e880fb5e
|
1555 |
this_cpu_inc(nr_dentry); |
312d3ca85
|
1556 |
|
1da177e4c
|
1557 1558 |
return dentry; } |
a4464dbc0
|
1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 |
/** * d_alloc - allocate a dcache entry * @parent: parent of entry to allocate * @name: qstr of the name * * Allocates a dentry. It returns %NULL if there is insufficient memory * available. On a success the dentry is returned. The name passed in is * copied and the copy passed in may be reused after this call. */ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name) { struct dentry *dentry = __d_alloc(parent->d_sb, name); if (!dentry) return NULL; spin_lock(&parent->d_lock); /* * don't need child lock because it is not subject * to concurrency here */ __dget_dlock(parent); dentry->d_parent = parent; |
946e51f2b
|
1582 |
list_add(&dentry->d_child, &parent->d_subdirs); |
a4464dbc0
|
1583 1584 1585 1586 |
spin_unlock(&parent->d_lock); return dentry; } |
ec4f86059
|
1587 |
EXPORT_SYMBOL(d_alloc); |
1da177e4c
|
1588 |
|
e1a24bb0a
|
1589 1590 1591 1592 1593 1594 1595 1596 |
/** * d_alloc_pseudo - allocate a dentry (for lookup-less filesystems) * @sb: the superblock * @name: qstr of the name * * For a filesystem that just pins its dentries in memory and never * performs lookups at all, return an unhashed IS_ROOT dentry. */ |
4b936885a
|
1597 1598 |
struct dentry *d_alloc_pseudo(struct super_block *sb, const struct qstr *name) { |
e1a24bb0a
|
1599 |
return __d_alloc(sb, name); |
4b936885a
|
1600 1601 |
} EXPORT_SYMBOL(d_alloc_pseudo); |
1da177e4c
|
1602 1603 1604 1605 1606 1607 1608 1609 1610 |
struct dentry *d_alloc_name(struct dentry *parent, const char *name) { struct qstr q; q.name = name; q.len = strlen(name); q.hash = full_name_hash(q.name, q.len); return d_alloc(parent, &q); } |
ef26ca97e
|
1611 |
EXPORT_SYMBOL(d_alloc_name); |
1da177e4c
|
1612 |
|
fb045adb9
|
1613 1614 |
void d_set_d_op(struct dentry *dentry, const struct dentry_operations *op) { |
6f7f7caab
|
1615 1616 |
WARN_ON_ONCE(dentry->d_op); WARN_ON_ONCE(dentry->d_flags & (DCACHE_OP_HASH | |
fb045adb9
|
1617 1618 |
DCACHE_OP_COMPARE | DCACHE_OP_REVALIDATE | |
ecf3d1f1a
|
1619 |
DCACHE_OP_WEAK_REVALIDATE | |
4bacc9c92
|
1620 1621 |
DCACHE_OP_DELETE | DCACHE_OP_SELECT_INODE)); |
fb045adb9
|
1622 1623 1624 1625 1626 1627 1628 1629 1630 |
dentry->d_op = op; if (!op) return; if (op->d_hash) dentry->d_flags |= DCACHE_OP_HASH; if (op->d_compare) dentry->d_flags |= DCACHE_OP_COMPARE; if (op->d_revalidate) dentry->d_flags |= DCACHE_OP_REVALIDATE; |
ecf3d1f1a
|
1631 1632 |
if (op->d_weak_revalidate) dentry->d_flags |= DCACHE_OP_WEAK_REVALIDATE; |
fb045adb9
|
1633 1634 |
if (op->d_delete) dentry->d_flags |= DCACHE_OP_DELETE; |
f0023bc61
|
1635 1636 |
if (op->d_prune) dentry->d_flags |= DCACHE_OP_PRUNE; |
4bacc9c92
|
1637 1638 |
if (op->d_select_inode) dentry->d_flags |= DCACHE_OP_SELECT_INODE; |
fb045adb9
|
1639 1640 1641 |
} EXPORT_SYMBOL(d_set_d_op); |
df1a085af
|
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 |
/* * d_set_fallthru - Mark a dentry as falling through to a lower layer * @dentry - The dentry to mark * * Mark a dentry as falling through to the lower layer (as set with * d_pin_lower()). This flag may be recorded on the medium. */ void d_set_fallthru(struct dentry *dentry) { spin_lock(&dentry->d_lock); dentry->d_flags |= DCACHE_FALLTHRU; spin_unlock(&dentry->d_lock); } EXPORT_SYMBOL(d_set_fallthru); |
b18825a7c
|
1657 1658 |
static unsigned d_flags_for_inode(struct inode *inode) { |
44bdb5e5f
|
1659 |
unsigned add_flags = DCACHE_REGULAR_TYPE; |
b18825a7c
|
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 |
if (!inode) return DCACHE_MISS_TYPE; if (S_ISDIR(inode->i_mode)) { add_flags = DCACHE_DIRECTORY_TYPE; if (unlikely(!(inode->i_opflags & IOP_LOOKUP))) { if (unlikely(!inode->i_op->lookup)) add_flags = DCACHE_AUTODIR_TYPE; else inode->i_opflags |= IOP_LOOKUP; } |
44bdb5e5f
|
1672 1673 1674 1675 1676 |
goto type_determined; } if (unlikely(!(inode->i_opflags & IOP_NOFOLLOW))) { if (unlikely(inode->i_op->follow_link)) { |
b18825a7c
|
1677 |
add_flags = DCACHE_SYMLINK_TYPE; |
44bdb5e5f
|
1678 1679 1680 |
goto type_determined; } inode->i_opflags |= IOP_NOFOLLOW; |
b18825a7c
|
1681 |
} |
44bdb5e5f
|
1682 1683 1684 1685 |
if (unlikely(!S_ISREG(inode->i_mode))) add_flags = DCACHE_SPECIAL_TYPE; type_determined: |
b18825a7c
|
1686 1687 1688 1689 |
if (unlikely(IS_AUTOMOUNT(inode))) add_flags |= DCACHE_NEED_AUTOMOUNT; return add_flags; } |
360da9002
|
1690 1691 |
static void __d_instantiate(struct dentry *dentry, struct inode *inode) { |
b18825a7c
|
1692 |
unsigned add_flags = d_flags_for_inode(inode); |
b23fb0a60
|
1693 |
spin_lock(&dentry->d_lock); |
b18825a7c
|
1694 |
if (inode) |
946e51f2b
|
1695 |
hlist_add_head(&dentry->d_u.d_alias, &inode->i_dentry); |
4bf46a272
|
1696 |
__d_set_inode_and_type(dentry, inode, add_flags); |
a7c6f571f
|
1697 |
dentry_rcuwalk_invalidate(dentry); |
b23fb0a60
|
1698 |
spin_unlock(&dentry->d_lock); |
360da9002
|
1699 1700 |
fsnotify_d_instantiate(dentry, inode); } |
1da177e4c
|
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 |
/** * d_instantiate - fill in inode information for a dentry * @entry: dentry to complete * @inode: inode to attach to this dentry * * Fill in inode information in the entry. * * This turns negative dentries into productive full members * of society. * * NOTE! This assumes that the inode count has been incremented * (or otherwise set) by the caller to indicate that it is now * in use by the dcache. */ void d_instantiate(struct dentry *entry, struct inode * inode) { |
946e51f2b
|
1718 |
BUG_ON(!hlist_unhashed(&entry->d_u.d_alias)); |
873feea09
|
1719 1720 |
if (inode) spin_lock(&inode->i_lock); |
360da9002
|
1721 |
__d_instantiate(entry, inode); |
873feea09
|
1722 1723 |
if (inode) spin_unlock(&inode->i_lock); |
1da177e4c
|
1724 1725 |
security_d_instantiate(entry, inode); } |
ec4f86059
|
1726 |
EXPORT_SYMBOL(d_instantiate); |
1da177e4c
|
1727 1728 1729 1730 1731 1732 1733 1734 |
/** * d_instantiate_unique - instantiate a non-aliased dentry * @entry: dentry to instantiate * @inode: inode to attach to this dentry * * Fill in inode information in the entry. On success, it returns NULL. * If an unhashed alias of "entry" already exists, then we return the |
e866cfa93
|
1735 |
* aliased dentry instead and drop one reference to inode. |
1da177e4c
|
1736 1737 1738 |
* * Note that in order to avoid conflicts with rename() etc, the caller * had better be holding the parent directory semaphore. |
e866cfa93
|
1739 1740 1741 1742 |
* * This also assumes that the inode count has been incremented * (or otherwise set) by the caller to indicate that it is now * in use by the dcache. |
1da177e4c
|
1743 |
*/ |
770bfad84
|
1744 1745 |
static struct dentry *__d_instantiate_unique(struct dentry *entry, struct inode *inode) |
1da177e4c
|
1746 1747 1748 1749 1750 |
{ struct dentry *alias; int len = entry->d_name.len; const char *name = entry->d_name.name; unsigned int hash = entry->d_name.hash; |
770bfad84
|
1751 |
if (!inode) { |
360da9002
|
1752 |
__d_instantiate(entry, NULL); |
770bfad84
|
1753 1754 |
return NULL; } |
946e51f2b
|
1755 |
hlist_for_each_entry(alias, &inode->i_dentry, d_u.d_alias) { |
9abca3608
|
1756 1757 1758 1759 1760 |
/* * Don't need alias->d_lock here, because aliases with * d_parent == entry->d_parent are not subject to name or * parent changes, because the parent inode i_mutex is held. */ |
12f8ad4b0
|
1761 |
if (alias->d_name.hash != hash) |
1da177e4c
|
1762 1763 1764 |
continue; if (alias->d_parent != entry->d_parent) continue; |
ee983e896
|
1765 1766 |
if (alias->d_name.len != len) continue; |
12f8ad4b0
|
1767 |
if (dentry_cmp(alias, name, len)) |
1da177e4c
|
1768 |
continue; |
dc0474be3
|
1769 |
__dget(alias); |
1da177e4c
|
1770 1771 |
return alias; } |
770bfad84
|
1772 |
|
360da9002
|
1773 |
__d_instantiate(entry, inode); |
1da177e4c
|
1774 1775 |
return NULL; } |
770bfad84
|
1776 1777 1778 1779 |
struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode) { struct dentry *result; |
946e51f2b
|
1780 |
BUG_ON(!hlist_unhashed(&entry->d_u.d_alias)); |
770bfad84
|
1781 |
|
873feea09
|
1782 1783 |
if (inode) spin_lock(&inode->i_lock); |
770bfad84
|
1784 |
result = __d_instantiate_unique(entry, inode); |
873feea09
|
1785 1786 |
if (inode) spin_unlock(&inode->i_lock); |
770bfad84
|
1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 |
if (!result) { security_d_instantiate(entry, inode); return NULL; } BUG_ON(!d_unhashed(result)); iput(inode); return result; } |
1da177e4c
|
1797 |
EXPORT_SYMBOL(d_instantiate_unique); |
b70a80e7a
|
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 |
/** * d_instantiate_no_diralias - instantiate a non-aliased dentry * @entry: dentry to complete * @inode: inode to attach to this dentry * * Fill in inode information in the entry. If a directory alias is found, then * return an error (and drop inode). Together with d_materialise_unique() this * guarantees that a directory inode may never have more than one alias. */ int d_instantiate_no_diralias(struct dentry *entry, struct inode *inode) { |
946e51f2b
|
1809 |
BUG_ON(!hlist_unhashed(&entry->d_u.d_alias)); |
b70a80e7a
|
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 |
spin_lock(&inode->i_lock); if (S_ISDIR(inode->i_mode) && !hlist_empty(&inode->i_dentry)) { spin_unlock(&inode->i_lock); iput(inode); return -EBUSY; } __d_instantiate(entry, inode); spin_unlock(&inode->i_lock); security_d_instantiate(entry, inode); return 0; } EXPORT_SYMBOL(d_instantiate_no_diralias); |
adc0e91ab
|
1824 1825 1826 1827 1828 |
struct dentry *d_make_root(struct inode *root_inode) { struct dentry *res = NULL; if (root_inode) { |
26fe57502
|
1829 |
static const struct qstr name = QSTR_INIT("/", 1); |
adc0e91ab
|
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 |
res = __d_alloc(root_inode->i_sb, &name); if (res) d_instantiate(res, root_inode); else iput(root_inode); } return res; } EXPORT_SYMBOL(d_make_root); |
d891eedbc
|
1840 1841 1842 |
static struct dentry * __d_find_any_alias(struct inode *inode) { struct dentry *alias; |
b3d9b7a3c
|
1843 |
if (hlist_empty(&inode->i_dentry)) |
d891eedbc
|
1844 |
return NULL; |
946e51f2b
|
1845 |
alias = hlist_entry(inode->i_dentry.first, struct dentry, d_u.d_alias); |
d891eedbc
|
1846 1847 1848 |
__dget(alias); return alias; } |
46f72b349
|
1849 1850 1851 1852 1853 1854 1855 1856 |
/** * d_find_any_alias - find any alias for a given inode * @inode: inode to find an alias for * * If any aliases exist for the given inode, take and return a * reference for one of them. If no aliases exist, return %NULL. */ struct dentry *d_find_any_alias(struct inode *inode) |
d891eedbc
|
1857 1858 1859 1860 1861 1862 1863 1864 |
{ struct dentry *de; spin_lock(&inode->i_lock); de = __d_find_any_alias(inode); spin_unlock(&inode->i_lock); return de; } |
46f72b349
|
1865 |
EXPORT_SYMBOL(d_find_any_alias); |
d891eedbc
|
1866 |
|
49c7dd287
|
1867 |
static struct dentry *__d_obtain_alias(struct inode *inode, int disconnected) |
4ea3ada29
|
1868 |
{ |
b911a6bde
|
1869 |
static const struct qstr anonstring = QSTR_INIT("/", 1); |
9308a6128
|
1870 1871 |
struct dentry *tmp; struct dentry *res; |
b18825a7c
|
1872 |
unsigned add_flags; |
4ea3ada29
|
1873 1874 |
if (!inode) |
440037287
|
1875 |
return ERR_PTR(-ESTALE); |
4ea3ada29
|
1876 1877 |
if (IS_ERR(inode)) return ERR_CAST(inode); |
d891eedbc
|
1878 |
res = d_find_any_alias(inode); |
9308a6128
|
1879 1880 |
if (res) goto out_iput; |
a4464dbc0
|
1881 |
tmp = __d_alloc(inode->i_sb, &anonstring); |
9308a6128
|
1882 1883 1884 |
if (!tmp) { res = ERR_PTR(-ENOMEM); goto out_iput; |
4ea3ada29
|
1885 |
} |
b5c84bf6f
|
1886 |
|
873feea09
|
1887 |
spin_lock(&inode->i_lock); |
d891eedbc
|
1888 |
res = __d_find_any_alias(inode); |
9308a6128
|
1889 |
if (res) { |
873feea09
|
1890 |
spin_unlock(&inode->i_lock); |
9308a6128
|
1891 1892 1893 1894 1895 |
dput(tmp); goto out_iput; } /* attach a disconnected dentry */ |
1a0a397e4
|
1896 1897 1898 1899 |
add_flags = d_flags_for_inode(inode); if (disconnected) add_flags |= DCACHE_DISCONNECTED; |
b18825a7c
|
1900 |
|
9308a6128
|
1901 |
spin_lock(&tmp->d_lock); |
4bf46a272
|
1902 |
__d_set_inode_and_type(tmp, inode, add_flags); |
946e51f2b
|
1903 |
hlist_add_head(&tmp->d_u.d_alias, &inode->i_dentry); |
1879fd6a2
|
1904 |
hlist_bl_lock(&tmp->d_sb->s_anon); |
ceb5bdc2d
|
1905 |
hlist_bl_add_head(&tmp->d_hash, &tmp->d_sb->s_anon); |
1879fd6a2
|
1906 |
hlist_bl_unlock(&tmp->d_sb->s_anon); |
9308a6128
|
1907 |
spin_unlock(&tmp->d_lock); |
873feea09
|
1908 |
spin_unlock(&inode->i_lock); |
24ff6663c
|
1909 |
security_d_instantiate(tmp, inode); |
9308a6128
|
1910 |
|
9308a6128
|
1911 1912 1913 |
return tmp; out_iput: |
24ff6663c
|
1914 1915 |
if (res && !IS_ERR(res)) security_d_instantiate(res, inode); |
9308a6128
|
1916 1917 |
iput(inode); return res; |
4ea3ada29
|
1918 |
} |
1a0a397e4
|
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 |
/** * d_obtain_alias - find or allocate a DISCONNECTED dentry for a given inode * @inode: inode to allocate the dentry for * * Obtain a dentry for an inode resulting from NFS filehandle conversion or * similar open by handle operations. The returned dentry may be anonymous, * or may have a full name (if the inode was already in the cache). * * When called on a directory inode, we must ensure that the inode only ever * has one dentry. If a dentry is found, that is returned instead of * allocating a new one. * * On successful return, the reference to the inode has been transferred * to the dentry. In case of an error the reference on the inode is released. * To make it easier to use in export operations a %NULL or IS_ERR inode may * be passed in and the error will be propagated to the return value, * with a %NULL @inode replaced by ERR_PTR(-ESTALE). */ struct dentry *d_obtain_alias(struct inode *inode) { return __d_obtain_alias(inode, 1); } |
adc487204
|
1942 |
EXPORT_SYMBOL(d_obtain_alias); |
1da177e4c
|
1943 1944 |
/** |
1a0a397e4
|
1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 |
* d_obtain_root - find or allocate a dentry for a given inode * @inode: inode to allocate the dentry for * * Obtain an IS_ROOT dentry for the root of a filesystem. * * We must ensure that directory inodes only ever have one dentry. If a * dentry is found, that is returned instead of allocating a new one. * * On successful return, the reference to the inode has been transferred * to the dentry. In case of an error the reference on the inode is * released. A %NULL or IS_ERR inode may be passed in and will be the * error will be propagate to the return value, with a %NULL @inode * replaced by ERR_PTR(-ESTALE). */ struct dentry *d_obtain_root(struct inode *inode) { return __d_obtain_alias(inode, 0); } EXPORT_SYMBOL(d_obtain_root); /** |
9403540c0
|
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 |
* d_add_ci - lookup or allocate new dentry with case-exact name * @inode: the inode case-insensitive lookup has found * @dentry: the negative dentry that was passed to the parent's lookup func * @name: the case-exact name to be associated with the returned dentry * * This is to avoid filling the dcache with case-insensitive names to the * same inode, only the actual correct case is stored in the dcache for * case-insensitive filesystems. * * For a case-insensitive lookup match and if the the case-exact dentry * already exists in in the dcache, use it and return it. * * If no entry exists with the exact case name, allocate new dentry with * the exact case, and return the spliced entry. */ |
e45b590b9
|
1981 |
struct dentry *d_add_ci(struct dentry *dentry, struct inode *inode, |
9403540c0
|
1982 1983 |
struct qstr *name) { |
9403540c0
|
1984 1985 |
struct dentry *found; struct dentry *new; |
b6520c819
|
1986 1987 1988 1989 |
/* * First check if a dentry matching the name already exists, * if not go ahead and create it now. */ |
9403540c0
|
1990 |
found = d_hash_and_lookup(dentry->d_parent, name); |
9403540c0
|
1991 1992 1993 |
if (!found) { new = d_alloc(dentry->d_parent, name); if (!new) { |
4f522a247
|
1994 |
found = ERR_PTR(-ENOMEM); |
427c77d46
|
1995 1996 1997 1998 1999 2000 2001 |
} else { found = d_splice_alias(inode, new); if (found) { dput(new); return found; } return new; |
9403540c0
|
2002 |
} |
9403540c0
|
2003 |
} |
9403540c0
|
2004 |
iput(inode); |
4f522a247
|
2005 |
return found; |
9403540c0
|
2006 |
} |
ec4f86059
|
2007 |
EXPORT_SYMBOL(d_add_ci); |
1da177e4c
|
2008 |
|
12f8ad4b0
|
2009 2010 2011 2012 |
/* * Do the slow-case of the dentry name compare. * * Unlike the dentry_cmp() function, we need to atomically |
da53be12b
|
2013 |
* load the name and length information, so that the |
12f8ad4b0
|
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 |
* filesystem can rely on them, and can use the 'name' and * 'len' information without worrying about walking off the * end of memory etc. * * Thus the read_seqcount_retry() and the "duplicate" info * in arguments (the low-level filesystem should not look * at the dentry inode or name contents directly, since * rename can change them while we're in RCU mode). */ enum slow_d_compare { D_COMP_OK, D_COMP_NOMATCH, D_COMP_SEQRETRY, }; static noinline enum slow_d_compare slow_dentry_cmp( const struct dentry *parent, |
12f8ad4b0
|
2031 2032 2033 2034 2035 2036 |
struct dentry *dentry, unsigned int seq, const struct qstr *name) { int tlen = dentry->d_name.len; const char *tname = dentry->d_name.name; |
12f8ad4b0
|
2037 2038 2039 2040 2041 |
if (read_seqcount_retry(&dentry->d_seq, seq)) { cpu_relax(); return D_COMP_SEQRETRY; } |
da53be12b
|
2042 |
if (parent->d_op->d_compare(parent, dentry, tlen, tname, name)) |
12f8ad4b0
|
2043 2044 2045 |
return D_COMP_NOMATCH; return D_COMP_OK; } |
1da177e4c
|
2046 |
/** |
31e6b01f4
|
2047 2048 2049 |
* __d_lookup_rcu - search for a dentry (racy, store-free) * @parent: parent dentry * @name: qstr of name we wish to find |
1f1e6e523
|
2050 |
* @seqp: returns d_seq value at the point where the dentry was found |
31e6b01f4
|
2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 |
* Returns: dentry, or NULL * * __d_lookup_rcu is the dcache lookup function for rcu-walk name * resolution (store-free path walking) design described in * Documentation/filesystems/path-lookup.txt. * * This is not to be used outside core vfs. * * __d_lookup_rcu must only be used in rcu-walk mode, ie. with vfsmount lock * held, and rcu_read_lock held. The returned dentry must not be stored into * without taking d_lock and checking d_seq sequence count against @seq * returned here. * |
15570086b
|
2064 |
* A refcount may be taken on the found dentry with the d_rcu_to_refcount |
31e6b01f4
|
2065 2066 2067 2068 2069 2070 |
* function. * * Alternatively, __d_lookup_rcu may be called again to look up the child of * the returned dentry, so long as its parent's seqlock is checked after the * child is looked up. Thus, an interlocking stepping of sequence lock checks * is formed, giving integrity down the path walk. |
12f8ad4b0
|
2071 2072 2073 |
* * NOTE! The caller *has* to check the resulting dentry against the sequence * number we've returned before using any of the resulting dentry state! |
31e6b01f4
|
2074 |
*/ |
8966be903
|
2075 2076 |
struct dentry *__d_lookup_rcu(const struct dentry *parent, const struct qstr *name, |
da53be12b
|
2077 |
unsigned *seqp) |
31e6b01f4
|
2078 |
{ |
26fe57502
|
2079 |
u64 hashlen = name->hash_len; |
31e6b01f4
|
2080 |
const unsigned char *str = name->name; |
26fe57502
|
2081 |
struct hlist_bl_head *b = d_hash(parent, hashlen_hash(hashlen)); |
ceb5bdc2d
|
2082 |
struct hlist_bl_node *node; |
31e6b01f4
|
2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 |
struct dentry *dentry; /* * Note: There is significant duplication with __d_lookup_rcu which is * required to prevent single threaded performance regressions * especially on architectures where smp_rmb (in seqcounts) are costly. * Keep the two functions in sync. */ /* * The hash list is protected using RCU. * * Carefully use d_seq when comparing a candidate dentry, to avoid * races with d_move(). * * It is possible that concurrent renames can mess up our list * walk here and result in missing our dentry, resulting in the * false-negative result. d_lookup() protects against concurrent * renames using rename_lock seqlock. * |
b0a4bb830
|
2103 |
* See Documentation/filesystems/path-lookup.txt for more details. |
31e6b01f4
|
2104 |
*/ |
b07ad9967
|
2105 |
hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) { |
8966be903
|
2106 |
unsigned seq; |
31e6b01f4
|
2107 |
|
31e6b01f4
|
2108 |
seqretry: |
12f8ad4b0
|
2109 2110 |
/* * The dentry sequence count protects us from concurrent |
da53be12b
|
2111 |
* renames, and thus protects parent and name fields. |
12f8ad4b0
|
2112 2113 |
* * The caller must perform a seqcount check in order |
da53be12b
|
2114 |
* to do anything useful with the returned dentry. |
12f8ad4b0
|
2115 2116 2117 2118 2119 2120 2121 2122 2123 |
* * NOTE! We do a "raw" seqcount_begin here. That means that * we don't wait for the sequence count to stabilize if it * is in the middle of a sequence change. If we do the slow * dentry compare, we will do seqretries until it is stable, * and if we end up with a successful lookup, we actually * want to exit RCU lookup anyway. */ seq = raw_seqcount_begin(&dentry->d_seq); |
31e6b01f4
|
2124 2125 |
if (dentry->d_parent != parent) continue; |
2e321806b
|
2126 2127 |
if (d_unhashed(dentry)) continue; |
12f8ad4b0
|
2128 |
|
830c0f0ed
|
2129 |
if (unlikely(parent->d_flags & DCACHE_OP_COMPARE)) { |
26fe57502
|
2130 2131 |
if (dentry->d_name.hash != hashlen_hash(hashlen)) continue; |
da53be12b
|
2132 2133 |
*seqp = seq; switch (slow_dentry_cmp(parent, dentry, seq, name)) { |
12f8ad4b0
|
2134 2135 2136 |
case D_COMP_OK: return dentry; case D_COMP_NOMATCH: |
31e6b01f4
|
2137 |
continue; |
12f8ad4b0
|
2138 2139 2140 |
default: goto seqretry; } |
31e6b01f4
|
2141 |
} |
12f8ad4b0
|
2142 |
|
26fe57502
|
2143 |
if (dentry->d_name.hash_len != hashlen) |
ee983e896
|
2144 |
continue; |
da53be12b
|
2145 |
*seqp = seq; |
26fe57502
|
2146 |
if (!dentry_cmp(dentry, str, hashlen_len(hashlen))) |
12f8ad4b0
|
2147 |
return dentry; |
31e6b01f4
|
2148 2149 2150 2151 2152 |
} return NULL; } /** |
1da177e4c
|
2153 2154 2155 |
* d_lookup - search for a dentry * @parent: parent dentry * @name: qstr of name we wish to find |
b04f784e5
|
2156 |
* Returns: dentry, or NULL |
1da177e4c
|
2157 |
* |
b04f784e5
|
2158 2159 2160 2161 |
* d_lookup searches the children of the parent dentry for the name in * question. If the dentry is found its reference count is incremented and the * dentry is returned. The caller must use dput to free the entry when it has * finished using it. %NULL is returned if the dentry does not exist. |
1da177e4c
|
2162 |
*/ |
da2d8455e
|
2163 |
struct dentry *d_lookup(const struct dentry *parent, const struct qstr *name) |
1da177e4c
|
2164 |
{ |
31e6b01f4
|
2165 |
struct dentry *dentry; |
949854d02
|
2166 |
unsigned seq; |
1da177e4c
|
2167 |
|
b8314f930
|
2168 2169 2170 2171 |
do { seq = read_seqbegin(&rename_lock); dentry = __d_lookup(parent, name); if (dentry) |
1da177e4c
|
2172 2173 2174 2175 |
break; } while (read_seqretry(&rename_lock, seq)); return dentry; } |
ec4f86059
|
2176 |
EXPORT_SYMBOL(d_lookup); |
1da177e4c
|
2177 |
|
31e6b01f4
|
2178 |
/** |
b04f784e5
|
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 |
* __d_lookup - search for a dentry (racy) * @parent: parent dentry * @name: qstr of name we wish to find * Returns: dentry, or NULL * * __d_lookup is like d_lookup, however it may (rarely) return a * false-negative result due to unrelated rename activity. * * __d_lookup is slightly faster by avoiding rename_lock read seqlock, * however it must be used carefully, eg. with a following d_lookup in * the case of failure. * * __d_lookup callers must be commented. */ |
a713ca2ab
|
2193 |
struct dentry *__d_lookup(const struct dentry *parent, const struct qstr *name) |
1da177e4c
|
2194 2195 2196 2197 |
{ unsigned int len = name->len; unsigned int hash = name->hash; const unsigned char *str = name->name; |
b07ad9967
|
2198 |
struct hlist_bl_head *b = d_hash(parent, hash); |
ceb5bdc2d
|
2199 |
struct hlist_bl_node *node; |
31e6b01f4
|
2200 |
struct dentry *found = NULL; |
665a7583f
|
2201 |
struct dentry *dentry; |
1da177e4c
|
2202 |
|
b04f784e5
|
2203 |
/* |
31e6b01f4
|
2204 2205 2206 2207 2208 2209 2210 |
* Note: There is significant duplication with __d_lookup_rcu which is * required to prevent single threaded performance regressions * especially on architectures where smp_rmb (in seqcounts) are costly. * Keep the two functions in sync. */ /* |
b04f784e5
|
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 |
* The hash list is protected using RCU. * * Take d_lock when comparing a candidate dentry, to avoid races * with d_move(). * * It is possible that concurrent renames can mess up our list * walk here and result in missing our dentry, resulting in the * false-negative result. d_lookup() protects against concurrent * renames using rename_lock seqlock. * |
b0a4bb830
|
2221 |
* See Documentation/filesystems/path-lookup.txt for more details. |
b04f784e5
|
2222 |
*/ |
1da177e4c
|
2223 2224 |
rcu_read_lock(); |
b07ad9967
|
2225 |
hlist_bl_for_each_entry_rcu(dentry, node, b, d_hash) { |
1da177e4c
|
2226 |
|
1da177e4c
|
2227 2228 |
if (dentry->d_name.hash != hash) continue; |
1da177e4c
|
2229 2230 |
spin_lock(&dentry->d_lock); |
1da177e4c
|
2231 2232 |
if (dentry->d_parent != parent) goto next; |
d0185c088
|
2233 2234 |
if (d_unhashed(dentry)) goto next; |
1da177e4c
|
2235 2236 2237 2238 |
/* * It is safe to compare names since d_move() cannot * change the qstr (protected by d_lock). */ |
fb045adb9
|
2239 |
if (parent->d_flags & DCACHE_OP_COMPARE) { |
12f8ad4b0
|
2240 2241 |
int tlen = dentry->d_name.len; const char *tname = dentry->d_name.name; |
da53be12b
|
2242 |
if (parent->d_op->d_compare(parent, dentry, tlen, tname, name)) |
1da177e4c
|
2243 2244 |
goto next; } else { |
ee983e896
|
2245 2246 |
if (dentry->d_name.len != len) goto next; |
12f8ad4b0
|
2247 |
if (dentry_cmp(dentry, str, len)) |
1da177e4c
|
2248 2249 |
goto next; } |
98474236f
|
2250 |
dentry->d_lockref.count++; |
d0185c088
|
2251 |
found = dentry; |
1da177e4c
|
2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 |
spin_unlock(&dentry->d_lock); break; next: spin_unlock(&dentry->d_lock); } rcu_read_unlock(); return found; } /** |
3e7e241f8
|
2263 2264 2265 2266 |
* d_hash_and_lookup - hash the qstr then search for a dentry * @dir: Directory to search in * @name: qstr of name we wish to find * |
4f522a247
|
2267 |
* On lookup failure NULL is returned; on bad name - ERR_PTR(-error) |
3e7e241f8
|
2268 2269 2270 |
*/ struct dentry *d_hash_and_lookup(struct dentry *dir, struct qstr *name) { |
3e7e241f8
|
2271 2272 2273 2274 2275 2276 |
/* * Check for a fs-specific hash function. Note that we must * calculate the standard hash first, as the d_op->d_hash() * routine may choose to leave the hash value unchanged. */ name->hash = full_name_hash(name->name, name->len); |
fb045adb9
|
2277 |
if (dir->d_flags & DCACHE_OP_HASH) { |
da53be12b
|
2278 |
int err = dir->d_op->d_hash(dir, name); |
4f522a247
|
2279 2280 |
if (unlikely(err < 0)) return ERR_PTR(err); |
3e7e241f8
|
2281 |
} |
4f522a247
|
2282 |
return d_lookup(dir, name); |
3e7e241f8
|
2283 |
} |
4f522a247
|
2284 |
EXPORT_SYMBOL(d_hash_and_lookup); |
3e7e241f8
|
2285 |
|
1da177e4c
|
2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 |
/* * When a file is deleted, we have two options: * - turn this dentry into a negative dentry * - unhash this dentry and free it. * * Usually, we want to just turn this into * a negative dentry, but if anybody else is * currently using the dentry or the inode * we can't do that and we fall back on removing * it from the hash queues and waiting for * it to be deleted later when it has no users */ /** * d_delete - delete a dentry * @dentry: The dentry to delete * * Turn the dentry into a negative dentry if possible, otherwise * remove it from the hash queues so it can be deleted later */ void d_delete(struct dentry * dentry) { |
873feea09
|
2309 |
struct inode *inode; |
7a91bf7f5
|
2310 |
int isdir = 0; |
1da177e4c
|
2311 2312 2313 |
/* * Are we the only user? */ |
357f8e658
|
2314 |
again: |
1da177e4c
|
2315 |
spin_lock(&dentry->d_lock); |
873feea09
|
2316 2317 |
inode = dentry->d_inode; isdir = S_ISDIR(inode->i_mode); |
98474236f
|
2318 |
if (dentry->d_lockref.count == 1) { |
1fe0c0230
|
2319 |
if (!spin_trylock(&inode->i_lock)) { |
357f8e658
|
2320 2321 2322 2323 |
spin_unlock(&dentry->d_lock); cpu_relax(); goto again; } |
13e3c5e5b
|
2324 |
dentry->d_flags &= ~DCACHE_CANT_MOUNT; |
31e6b01f4
|
2325 |
dentry_unlink_inode(dentry); |
7a91bf7f5
|
2326 |
fsnotify_nameremove(dentry, isdir); |
1da177e4c
|
2327 2328 2329 2330 2331 2332 2333 |
return; } if (!d_unhashed(dentry)) __d_drop(dentry); spin_unlock(&dentry->d_lock); |
7a91bf7f5
|
2334 2335 |
fsnotify_nameremove(dentry, isdir); |
1da177e4c
|
2336 |
} |
ec4f86059
|
2337 |
EXPORT_SYMBOL(d_delete); |
1da177e4c
|
2338 |
|
b07ad9967
|
2339 |
static void __d_rehash(struct dentry * entry, struct hlist_bl_head *b) |
1da177e4c
|
2340 |
{ |
ceb5bdc2d
|
2341 |
BUG_ON(!d_unhashed(entry)); |
1879fd6a2
|
2342 |
hlist_bl_lock(b); |
dea3667bc
|
2343 |
entry->d_flags |= DCACHE_RCUACCESS; |
b07ad9967
|
2344 |
hlist_bl_add_head_rcu(&entry->d_hash, b); |
1879fd6a2
|
2345 |
hlist_bl_unlock(b); |
1da177e4c
|
2346 |
} |
770bfad84
|
2347 2348 2349 2350 |
static void _d_rehash(struct dentry * entry) { __d_rehash(entry, d_hash(entry->d_parent, entry->d_name.hash)); } |
1da177e4c
|
2351 2352 2353 2354 2355 2356 2357 2358 2359 |
/** * d_rehash - add an entry back to the hash * @entry: dentry to add to the hash * * Adds a dentry to the hash according to its name. */ void d_rehash(struct dentry * entry) { |
1da177e4c
|
2360 |
spin_lock(&entry->d_lock); |
770bfad84
|
2361 |
_d_rehash(entry); |
1da177e4c
|
2362 |
spin_unlock(&entry->d_lock); |
1da177e4c
|
2363 |
} |
ec4f86059
|
2364 |
EXPORT_SYMBOL(d_rehash); |
1da177e4c
|
2365 |
|
fb2d5b86a
|
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 |
/** * dentry_update_name_case - update case insensitive dentry with a new name * @dentry: dentry to be updated * @name: new name * * Update a case insensitive dentry with new case of name. * * dentry must have been returned by d_lookup with name @name. Old and new * name lengths must match (ie. no d_compare which allows mismatched name * lengths). * * Parent inode i_mutex must be held over d_lookup and into this call (to * keep renames and concurrent inserts, and readdir(2) away). */ void dentry_update_name_case(struct dentry *dentry, struct qstr *name) { |
7ebfa57f6
|
2382 |
BUG_ON(!mutex_is_locked(&dentry->d_parent->d_inode->i_mutex)); |
fb2d5b86a
|
2383 |
BUG_ON(dentry->d_name.len != name->len); /* d_lookup gives this */ |
fb2d5b86a
|
2384 |
spin_lock(&dentry->d_lock); |
31e6b01f4
|
2385 |
write_seqcount_begin(&dentry->d_seq); |
fb2d5b86a
|
2386 |
memcpy((unsigned char *)dentry->d_name.name, name->name, name->len); |
31e6b01f4
|
2387 |
write_seqcount_end(&dentry->d_seq); |
fb2d5b86a
|
2388 |
spin_unlock(&dentry->d_lock); |
fb2d5b86a
|
2389 2390 |
} EXPORT_SYMBOL(dentry_update_name_case); |
8d85b4845
|
2391 |
static void swap_names(struct dentry *dentry, struct dentry *target) |
1da177e4c
|
2392 |
{ |
8d85b4845
|
2393 2394 |
if (unlikely(dname_external(target))) { if (unlikely(dname_external(dentry))) { |
1da177e4c
|
2395 2396 2397 |
/* * Both external: swap the pointers */ |
9a8d5bb4a
|
2398 |
swap(target->d_name.name, dentry->d_name.name); |
1da177e4c
|
2399 2400 2401 2402 2403 |
} else { /* * dentry:internal, target:external. Steal target's * storage and make target internal. */ |
321bcf921
|
2404 2405 |
memcpy(target->d_iname, dentry->d_name.name, dentry->d_name.len + 1); |
1da177e4c
|
2406 2407 2408 2409 |
dentry->d_name.name = target->d_name.name; target->d_name.name = target->d_iname; } } else { |
8d85b4845
|
2410 |
if (unlikely(dname_external(dentry))) { |
1da177e4c
|
2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 |
/* * dentry:external, target:internal. Give dentry's * storage to target and make dentry internal */ memcpy(dentry->d_iname, target->d_name.name, target->d_name.len + 1); target->d_name.name = dentry->d_name.name; dentry->d_name.name = dentry->d_iname; } else { /* |
da1ce0670
|
2421 |
* Both are internal. |
1da177e4c
|
2422 |
*/ |
da1ce0670
|
2423 2424 |
unsigned int i; BUILD_BUG_ON(!IS_ALIGNED(DNAME_INLINE_LEN, sizeof(long))); |
08d4f7722
|
2425 2426 |
kmemcheck_mark_initialized(dentry->d_iname, DNAME_INLINE_LEN); kmemcheck_mark_initialized(target->d_iname, DNAME_INLINE_LEN); |
da1ce0670
|
2427 2428 2429 2430 |
for (i = 0; i < DNAME_INLINE_LEN / sizeof(long); i++) { swap(((long *) &dentry->d_iname)[i], ((long *) &target->d_iname)[i]); } |
1da177e4c
|
2431 2432 |
} } |
a28ddb87c
|
2433 |
swap(dentry->d_name.hash_len, target->d_name.hash_len); |
1da177e4c
|
2434 |
} |
8d85b4845
|
2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 |
static void copy_name(struct dentry *dentry, struct dentry *target) { struct external_name *old_name = NULL; if (unlikely(dname_external(dentry))) old_name = external_name(dentry); if (unlikely(dname_external(target))) { atomic_inc(&external_name(target)->u.count); dentry->d_name = target->d_name; } else { memcpy(dentry->d_iname, target->d_name.name, target->d_name.len + 1); dentry->d_name.name = dentry->d_iname; dentry->d_name.hash_len = target->d_name.hash_len; } if (old_name && likely(atomic_dec_and_test(&old_name->u.count))) kfree_rcu(old_name, u.head); } |
2fd6b7f50
|
2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 |
static void dentry_lock_for_move(struct dentry *dentry, struct dentry *target) { /* * XXXX: do we really need to take target->d_lock? */ if (IS_ROOT(dentry) || dentry->d_parent == target->d_parent) spin_lock(&target->d_parent->d_lock); else { if (d_ancestor(dentry->d_parent, target->d_parent)) { spin_lock(&dentry->d_parent->d_lock); spin_lock_nested(&target->d_parent->d_lock, DENTRY_D_LOCK_NESTED); } else { spin_lock(&target->d_parent->d_lock); spin_lock_nested(&dentry->d_parent->d_lock, DENTRY_D_LOCK_NESTED); } } if (target < dentry) { spin_lock_nested(&target->d_lock, 2); spin_lock_nested(&dentry->d_lock, 3); } else { spin_lock_nested(&dentry->d_lock, 2); spin_lock_nested(&target->d_lock, 3); } } |
986c01942
|
2478 |
static void dentry_unlock_for_move(struct dentry *dentry, struct dentry *target) |
2fd6b7f50
|
2479 2480 2481 2482 2483 |
{ if (target->d_parent != dentry->d_parent) spin_unlock(&dentry->d_parent->d_lock); if (target->d_parent != target) spin_unlock(&target->d_parent->d_lock); |
986c01942
|
2484 2485 |
spin_unlock(&target->d_lock); spin_unlock(&dentry->d_lock); |
2fd6b7f50
|
2486 |
} |
1da177e4c
|
2487 |
/* |
2fd6b7f50
|
2488 2489 2490 |
* When switching names, the actual string doesn't strictly have to * be preserved in the target - because we're dropping the target * anyway. As such, we can just do a simple memcpy() to copy over |
d2fa4a847
|
2491 2492 2493 2494 2495 2496 2497 2498 2499 |
* the new name before we switch, unless we are going to rehash * it. Note that if we *do* unhash the target, we are not allowed * to rehash it without giving it a new name/hash key - whether * we swap or overwrite the names here, resulting name won't match * the reality in filesystem; it's only there for d_path() purposes. * Note that all of this is happening under rename_lock, so the * any hash lookup seeing it in the middle of manipulations will * be discarded anyway. So we do not care what happens to the hash * key in that case. |
1da177e4c
|
2500 |
*/ |
9eaef27b3
|
2501 |
/* |
183675011
|
2502 |
* __d_move - move a dentry |
1da177e4c
|
2503 2504 |
* @dentry: entry to move * @target: new dentry |
da1ce0670
|
2505 |
* @exchange: exchange the two dentries |
1da177e4c
|
2506 2507 |
* * Update the dcache to reflect the move of a file name. Negative |
c46c88774
|
2508 2509 2510 |
* dcache entries should not be moved in this way. Caller must hold * rename_lock, the i_mutex of the source and target directories, * and the sb->s_vfs_rename_mutex if they differ. See lock_rename(). |
1da177e4c
|
2511 |
*/ |
da1ce0670
|
2512 2513 |
static void __d_move(struct dentry *dentry, struct dentry *target, bool exchange) |
1da177e4c
|
2514 |
{ |
1da177e4c
|
2515 2516 2517 |
if (!dentry->d_inode) printk(KERN_WARNING "VFS: moving negative dcache entry "); |
2fd6b7f50
|
2518 2519 |
BUG_ON(d_ancestor(dentry, target)); BUG_ON(d_ancestor(target, dentry)); |
2fd6b7f50
|
2520 |
dentry_lock_for_move(dentry, target); |
1da177e4c
|
2521 |
|
31e6b01f4
|
2522 |
write_seqcount_begin(&dentry->d_seq); |
1ca7d67cf
|
2523 |
write_seqcount_begin_nested(&target->d_seq, DENTRY_D_LOCK_NESTED); |
31e6b01f4
|
2524 |
|
ceb5bdc2d
|
2525 2526 2527 2528 2529 2530 2531 |
/* __d_drop does write_seqcount_barrier, but they're OK to nest. */ /* * Move the dentry to the target hash queue. Don't bother checking * for the same hash queue because of how unlikely it is. */ __d_drop(dentry); |
789680d1e
|
2532 |
__d_rehash(dentry, d_hash(target->d_parent, target->d_name.hash)); |
1da177e4c
|
2533 |
|
da1ce0670
|
2534 2535 2536 2537 |
/* * Unhash the target (d_delete() is not usable here). If exchanging * the two dentries, then rehash onto the other's hash queue. */ |
1da177e4c
|
2538 |
__d_drop(target); |
da1ce0670
|
2539 2540 2541 2542 |
if (exchange) { __d_rehash(target, d_hash(dentry->d_parent, dentry->d_name.hash)); } |
1da177e4c
|
2543 |
|
1da177e4c
|
2544 |
/* Switch the names.. */ |
8d85b4845
|
2545 2546 2547 2548 |
if (exchange) swap_names(dentry, target); else copy_name(dentry, target); |
1da177e4c
|
2549 |
|
63cf427a5
|
2550 |
/* ... and switch them in the tree */ |
1da177e4c
|
2551 |
if (IS_ROOT(dentry)) { |
63cf427a5
|
2552 |
/* splicing a tree */ |
1da177e4c
|
2553 2554 |
dentry->d_parent = target->d_parent; target->d_parent = target; |
946e51f2b
|
2555 2556 |
list_del_init(&target->d_child); list_move(&dentry->d_child, &dentry->d_parent->d_subdirs); |
1da177e4c
|
2557 |
} else { |
63cf427a5
|
2558 |
/* swapping two dentries */ |
9a8d5bb4a
|
2559 |
swap(dentry->d_parent, target->d_parent); |
946e51f2b
|
2560 2561 |
list_move(&target->d_child, &target->d_parent->d_subdirs); list_move(&dentry->d_child, &dentry->d_parent->d_subdirs); |
63cf427a5
|
2562 2563 2564 |
if (exchange) fsnotify_d_move(target); fsnotify_d_move(dentry); |
1da177e4c
|
2565 |
} |
31e6b01f4
|
2566 2567 |
write_seqcount_end(&target->d_seq); write_seqcount_end(&dentry->d_seq); |
986c01942
|
2568 |
dentry_unlock_for_move(dentry, target); |
183675011
|
2569 2570 2571 2572 2573 2574 2575 2576 |
} /* * d_move - move a dentry * @dentry: entry to move * @target: new dentry * * Update the dcache to reflect the move of a file name. Negative |
c46c88774
|
2577 2578 |
* dcache entries should not be moved in this way. See the locking * requirements for __d_move. |
183675011
|
2579 2580 2581 2582 |
*/ void d_move(struct dentry *dentry, struct dentry *target) { write_seqlock(&rename_lock); |
da1ce0670
|
2583 |
__d_move(dentry, target, false); |
1da177e4c
|
2584 |
write_sequnlock(&rename_lock); |
9eaef27b3
|
2585 |
} |
ec4f86059
|
2586 |
EXPORT_SYMBOL(d_move); |
1da177e4c
|
2587 |
|
da1ce0670
|
2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 |
/* * d_exchange - exchange two dentries * @dentry1: first dentry * @dentry2: second dentry */ void d_exchange(struct dentry *dentry1, struct dentry *dentry2) { write_seqlock(&rename_lock); WARN_ON(!dentry1->d_inode); WARN_ON(!dentry2->d_inode); WARN_ON(IS_ROOT(dentry1)); WARN_ON(IS_ROOT(dentry2)); __d_move(dentry1, dentry2, true); write_sequnlock(&rename_lock); } |
e2761a116
|
2606 2607 2608 2609 2610 2611 2612 |
/** * d_ancestor - search for an ancestor * @p1: ancestor dentry * @p2: child dentry * * Returns the ancestor dentry of p2 which is a child of p1, if p1 is * an ancestor of p2, else NULL. |
9eaef27b3
|
2613 |
*/ |
e2761a116
|
2614 |
struct dentry *d_ancestor(struct dentry *p1, struct dentry *p2) |
9eaef27b3
|
2615 2616 |
{ struct dentry *p; |
871c0067d
|
2617 |
for (p = p2; !IS_ROOT(p); p = p->d_parent) { |
9eaef27b3
|
2618 |
if (p->d_parent == p1) |
e2761a116
|
2619 |
return p; |
9eaef27b3
|
2620 |
} |
e2761a116
|
2621 |
return NULL; |
9eaef27b3
|
2622 2623 2624 2625 2626 2627 |
} /* * This helper attempts to cope with remotely renamed directories * * It assumes that the caller is already holding |
a03e283bf
|
2628 |
* dentry->d_parent->d_inode->i_mutex, and rename_lock |
9eaef27b3
|
2629 2630 2631 |
* * Note: If ever the locking in lock_rename() changes, then please * remember to update this too... |
9eaef27b3
|
2632 |
*/ |
b5ae6b15b
|
2633 |
static int __d_unalias(struct inode *inode, |
873feea09
|
2634 |
struct dentry *dentry, struct dentry *alias) |
9eaef27b3
|
2635 2636 |
{ struct mutex *m1 = NULL, *m2 = NULL; |
3d330dc17
|
2637 |
int ret = -ESTALE; |
9eaef27b3
|
2638 2639 2640 2641 |
/* If alias and dentry share a parent, then no extra locks required */ if (alias->d_parent == dentry->d_parent) goto out_unalias; |
9eaef27b3
|
2642 |
/* See lock_rename() */ |
9eaef27b3
|
2643 2644 2645 2646 2647 2648 2649 |
if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex)) goto out_err; m1 = &dentry->d_sb->s_vfs_rename_mutex; if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex)) goto out_err; m2 = &alias->d_parent->d_inode->i_mutex; out_unalias: |
8ed936b56
|
2650 |
__d_move(alias, dentry, false); |
b5ae6b15b
|
2651 |
ret = 0; |
9eaef27b3
|
2652 |
out_err: |
9eaef27b3
|
2653 2654 2655 2656 2657 2658 |
if (m2) mutex_unlock(m2); if (m1) mutex_unlock(m1); return ret; } |
770bfad84
|
2659 |
/** |
3f70bd51c
|
2660 2661 2662 2663 |
* d_splice_alias - splice a disconnected dentry into the tree if one exists * @inode: the inode which may have a disconnected dentry * @dentry: a negative dentry which we want to point to the inode. * |
da093a9b7
|
2664 2665 2666 |
* If inode is a directory and has an IS_ROOT alias, then d_move that in * place of the given dentry and return it, else simply d_add the inode * to the dentry and return NULL. |
3f70bd51c
|
2667 |
* |
908790fa3
|
2668 2669 2670 |
* If a non-IS_ROOT directory is found, the filesystem is corrupt, and * we should error out: directories can't have multiple aliases. * |
3f70bd51c
|
2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 |
* This is needed in the lookup routine of any filesystem that is exportable * (via knfsd) so that we can build dcache paths to directories effectively. * * If a dentry was found and moved, then it is returned. Otherwise NULL * is returned. This matches the expected return value of ->lookup. * * Cluster filesystems may call this function with a negative, hashed dentry. * In that case, we know that the inode will be a regular file, and also this * will only occur during atomic_open. So we need to check for the dentry * being already hashed only in the final case. */ struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry) { |
3f70bd51c
|
2684 2685 |
if (IS_ERR(inode)) return ERR_CAST(inode); |
770bfad84
|
2686 |
BUG_ON(!d_unhashed(dentry)); |
770bfad84
|
2687 |
if (!inode) { |
360da9002
|
2688 |
__d_instantiate(dentry, NULL); |
b5ae6b15b
|
2689 |
goto out; |
770bfad84
|
2690 |
} |
873feea09
|
2691 |
spin_lock(&inode->i_lock); |
9eaef27b3
|
2692 |
if (S_ISDIR(inode->i_mode)) { |
b5ae6b15b
|
2693 2694 |
struct dentry *new = __d_find_any_alias(inode); if (unlikely(new)) { |
a03e283bf
|
2695 2696 |
/* The reference to new ensures it remains an alias */ spin_unlock(&inode->i_lock); |
183675011
|
2697 |
write_seqlock(&rename_lock); |
b5ae6b15b
|
2698 2699 |
if (unlikely(d_ancestor(new, dentry))) { write_sequnlock(&rename_lock); |
b5ae6b15b
|
2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 |
dput(new); new = ERR_PTR(-ELOOP); pr_warn_ratelimited( "VFS: Lookup of '%s' in %s %s" " would have caused loop ", dentry->d_name.name, inode->i_sb->s_type->name, inode->i_sb->s_id); } else if (!IS_ROOT(new)) { int err = __d_unalias(inode, dentry, new); |
183675011
|
2711 |
write_sequnlock(&rename_lock); |
b5ae6b15b
|
2712 2713 2714 2715 |
if (err) { dput(new); new = ERR_PTR(err); } |
183675011
|
2716 |
} else { |
b5ae6b15b
|
2717 2718 |
__d_move(new, dentry, false); write_sequnlock(&rename_lock); |
b5ae6b15b
|
2719 |
security_d_instantiate(new, inode); |
dd179946d
|
2720 |
} |
b5ae6b15b
|
2721 2722 |
iput(inode); return new; |
9eaef27b3
|
2723 |
} |
770bfad84
|
2724 |
} |
b5ae6b15b
|
2725 2726 |
/* already taking inode->i_lock, so d_add() by hand */ __d_instantiate(dentry, inode); |
873feea09
|
2727 |
spin_unlock(&inode->i_lock); |
b5ae6b15b
|
2728 2729 2730 2731 |
out: security_d_instantiate(dentry, inode); d_rehash(dentry); return NULL; |
770bfad84
|
2732 |
} |
b5ae6b15b
|
2733 |
EXPORT_SYMBOL(d_splice_alias); |
770bfad84
|
2734 |
|
cdd16d026
|
2735 |
static int prepend(char **buffer, int *buflen, const char *str, int namelen) |
6092d0481
|
2736 2737 2738 2739 2740 2741 2742 2743 |
{ *buflen -= namelen; if (*buflen < 0) return -ENAMETOOLONG; *buffer -= namelen; memcpy(*buffer, str, namelen); return 0; } |
232d2d60a
|
2744 2745 |
/** * prepend_name - prepend a pathname in front of current buffer pointer |
181299772
|
2746 2747 2748 |
* @buffer: buffer pointer * @buflen: allocated length of the buffer * @name: name string and length qstr structure |
232d2d60a
|
2749 2750 2751 2752 2753 2754 2755 2756 2757 |
* * With RCU path tracing, it may race with d_move(). Use ACCESS_ONCE() to * make sure that either the old or the new name pointer and length are * fetched. However, there may be mismatch between length and pointer. * The length cannot be trusted, we need to copy it byte-by-byte until * the length is reached or a null byte is found. It also prepends "/" at * the beginning of the name. The sequence number check at the caller will * retry it again when a d_move() does happen. So any garbage in the buffer * due to mismatched pointer and length will be discarded. |
6d13f6944
|
2758 2759 2760 |
* * Data dependency barrier is needed to make sure that we see that terminating * NUL. Alpha strikes again, film at 11... |
232d2d60a
|
2761 |
*/ |
cdd16d026
|
2762 2763 |
static int prepend_name(char **buffer, int *buflen, struct qstr *name) { |
232d2d60a
|
2764 2765 2766 |
const char *dname = ACCESS_ONCE(name->name); u32 dlen = ACCESS_ONCE(name->len); char *p; |
6d13f6944
|
2767 |
smp_read_barrier_depends(); |
232d2d60a
|
2768 |
*buflen -= dlen + 1; |
e825196d4
|
2769 2770 |
if (*buflen < 0) return -ENAMETOOLONG; |
232d2d60a
|
2771 2772 2773 2774 2775 2776 2777 2778 2779 |
p = *buffer -= dlen + 1; *p++ = '/'; while (dlen--) { char c = *dname++; if (!c) break; *p++ = c; } return 0; |
cdd16d026
|
2780 |
} |
1da177e4c
|
2781 |
/** |
208898c17
|
2782 |
* prepend_path - Prepend path string to a buffer |
9d1bc6013
|
2783 |
* @path: the dentry/vfsmount to report |
02125a826
|
2784 |
* @root: root vfsmnt/dentry |
f2eb6575d
|
2785 2786 |
* @buffer: pointer to the end of the buffer * @buflen: pointer to buffer length |
552ce544e
|
2787 |
* |
181299772
|
2788 2789 2790 2791 2792 2793 2794 2795 2796 |
* The function will first try to write out the pathname without taking any * lock other than the RCU read lock to make sure that dentries won't go away. * It only checks the sequence number of the global rename_lock as any change * in the dentry's d_seq will be preceded by changes in the rename_lock * sequence number. If the sequence number had been changed, it will restart * the whole pathname back-tracing sequence again by taking the rename_lock. * In this case, there is no need to take the RCU read lock as the recursive * parent pointer references will keep the dentry chain alive as long as no * rename operation is performed. |
1da177e4c
|
2797 |
*/ |
02125a826
|
2798 2799 |
static int prepend_path(const struct path *path, const struct path *root, |
f2eb6575d
|
2800 |
char **buffer, int *buflen) |
1da177e4c
|
2801 |
{ |
ede4cebce
|
2802 2803 2804 |
struct dentry *dentry; struct vfsmount *vfsmnt; struct mount *mnt; |
f2eb6575d
|
2805 |
int error = 0; |
48a066e72
|
2806 |
unsigned seq, m_seq = 0; |
232d2d60a
|
2807 2808 |
char *bptr; int blen; |
6092d0481
|
2809 |
|
48f5ec21d
|
2810 |
rcu_read_lock(); |
48a066e72
|
2811 2812 2813 |
restart_mnt: read_seqbegin_or_lock(&mount_lock, &m_seq); seq = 0; |
4ec6c2aea
|
2814 |
rcu_read_lock(); |
232d2d60a
|
2815 2816 2817 |
restart: bptr = *buffer; blen = *buflen; |
48a066e72
|
2818 |
error = 0; |
ede4cebce
|
2819 2820 2821 |
dentry = path->dentry; vfsmnt = path->mnt; mnt = real_mount(vfsmnt); |
232d2d60a
|
2822 |
read_seqbegin_or_lock(&rename_lock, &seq); |
f2eb6575d
|
2823 |
while (dentry != root->dentry || vfsmnt != root->mnt) { |
1da177e4c
|
2824 |
struct dentry * parent; |
1da177e4c
|
2825 |
if (dentry == vfsmnt->mnt_root || IS_ROOT(dentry)) { |
48a066e72
|
2826 |
struct mount *parent = ACCESS_ONCE(mnt->mnt_parent); |
cde93be45
|
2827 2828 2829 2830 2831 2832 2833 |
/* Escaped? */ if (dentry != vfsmnt->mnt_root) { bptr = *buffer; blen = *buflen; error = 3; break; } |
552ce544e
|
2834 |
/* Global root? */ |
48a066e72
|
2835 2836 2837 |
if (mnt != parent) { dentry = ACCESS_ONCE(mnt->mnt_mountpoint); mnt = parent; |
232d2d60a
|
2838 2839 2840 |
vfsmnt = &mnt->mnt; continue; } |
232d2d60a
|
2841 2842 2843 |
if (!error) error = is_mounted(vfsmnt) ? 1 : 2; break; |
1da177e4c
|
2844 2845 2846 |
} parent = dentry->d_parent; prefetch(parent); |
232d2d60a
|
2847 |
error = prepend_name(&bptr, &blen, &dentry->d_name); |
f2eb6575d
|
2848 2849 |
if (error) break; |
1da177e4c
|
2850 2851 |
dentry = parent; } |
48f5ec21d
|
2852 2853 2854 2855 |
if (!(seq & 1)) rcu_read_unlock(); if (need_seqretry(&rename_lock, seq)) { seq = 1; |
232d2d60a
|
2856 |
goto restart; |
48f5ec21d
|
2857 2858 |
} done_seqretry(&rename_lock, seq); |
4ec6c2aea
|
2859 2860 2861 |
if (!(m_seq & 1)) rcu_read_unlock(); |
48a066e72
|
2862 2863 2864 2865 2866 |
if (need_seqretry(&mount_lock, m_seq)) { m_seq = 1; goto restart_mnt; } done_seqretry(&mount_lock, m_seq); |
1da177e4c
|
2867 |
|
232d2d60a
|
2868 2869 2870 2871 2872 2873 2874 2875 |
if (error >= 0 && bptr == *buffer) { if (--blen < 0) error = -ENAMETOOLONG; else *--bptr = '/'; } *buffer = bptr; *buflen = blen; |
7ea600b53
|
2876 |
return error; |
f2eb6575d
|
2877 |
} |
be285c712
|
2878 |
|
f2eb6575d
|
2879 2880 2881 |
/** * __d_path - return the path of a dentry * @path: the dentry/vfsmount to report |
02125a826
|
2882 |
* @root: root vfsmnt/dentry |
cd956a1c0
|
2883 |
* @buf: buffer to return value in |
f2eb6575d
|
2884 2885 |
* @buflen: buffer length * |
ffd1f4ed5
|
2886 |
* Convert a dentry into an ASCII path name. |
f2eb6575d
|
2887 2888 2889 2890 |
* * Returns a pointer into the buffer or an error code if the * path was too long. * |
be148247c
|
2891 |
* "buflen" should be positive. |
f2eb6575d
|
2892 |
* |
02125a826
|
2893 |
* If the path is not reachable from the supplied root, return %NULL. |
f2eb6575d
|
2894 |
*/ |
02125a826
|
2895 2896 |
char *__d_path(const struct path *path, const struct path *root, |
f2eb6575d
|
2897 2898 2899 2900 2901 2902 |
char *buf, int buflen) { char *res = buf + buflen; int error; prepend(&res, &buflen, "\0", 1); |
f2eb6575d
|
2903 |
error = prepend_path(path, root, &res, &buflen); |
be148247c
|
2904 |
|
02125a826
|
2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 |
if (error < 0) return ERR_PTR(error); if (error > 0) return NULL; return res; } char *d_absolute_path(const struct path *path, char *buf, int buflen) { struct path root = {}; char *res = buf + buflen; int error; prepend(&res, &buflen, "\0", 1); |
02125a826
|
2920 |
error = prepend_path(path, &root, &res, &buflen); |
02125a826
|
2921 2922 2923 2924 |
if (error > 1) error = -EINVAL; if (error < 0) |
f2eb6575d
|
2925 |
return ERR_PTR(error); |
f2eb6575d
|
2926 |
return res; |
1da177e4c
|
2927 |
} |
ffd1f4ed5
|
2928 2929 2930 |
/* * same as __d_path but appends "(deleted)" for unlinked files. */ |
02125a826
|
2931 2932 2933 |
static int path_with_deleted(const struct path *path, const struct path *root, char **buf, int *buflen) |
ffd1f4ed5
|
2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 |
{ prepend(buf, buflen, "\0", 1); if (d_unlinked(path->dentry)) { int error = prepend(buf, buflen, " (deleted)", 10); if (error) return error; } return prepend_path(path, root, buf, buflen); } |
8df9d1a41
|
2944 2945 2946 2947 |
static int prepend_unreachable(char **buffer, int *buflen) { return prepend(buffer, buflen, "(unreachable)", 13); } |
68f0d9d92
|
2948 2949 2950 2951 2952 2953 2954 2955 2956 |
static void get_fs_root_rcu(struct fs_struct *fs, struct path *root) { unsigned seq; do { seq = read_seqcount_begin(&fs->seq); *root = fs->root; } while (read_seqcount_retry(&fs->seq, seq)); } |
a03a8a709
|
2957 2958 |
/** * d_path - return the path of a dentry |
cf28b4863
|
2959 |
* @path: path to report |
a03a8a709
|
2960 2961 2962 2963 2964 2965 |
* @buf: buffer to return value in * @buflen: buffer length * * Convert a dentry into an ASCII path name. If the entry has been deleted * the string " (deleted)" is appended. Note that this is ambiguous. * |
52afeefb9
|
2966 2967 2968 2969 |
* Returns a pointer into the buffer or an error code if the path was * too long. Note: Callers should use the returned pointer, not the passed * in buffer, to use the name! The implementation often starts at an offset * into the buffer, and may leave 0 bytes at the start. |
a03a8a709
|
2970 |
* |
31f3e0b3a
|
2971 |
* "buflen" should be positive. |
a03a8a709
|
2972 |
*/ |
20d4fdc1a
|
2973 |
char *d_path(const struct path *path, char *buf, int buflen) |
1da177e4c
|
2974 |
{ |
ffd1f4ed5
|
2975 |
char *res = buf + buflen; |
6ac08c39a
|
2976 |
struct path root; |
ffd1f4ed5
|
2977 |
int error; |
1da177e4c
|
2978 |
|
c23fbb6bc
|
2979 2980 2981 2982 2983 2984 |
/* * We have various synthetic filesystems that never get mounted. On * these filesystems dentries are never used for lookup purposes, and * thus don't need to be hashed. They also don't need a name until a * user wants to identify the object in /proc/pid/fd/. The little hack * below allows us to generate a name for these objects on demand: |
f48cfddc6
|
2985 2986 2987 2988 |
* * Some pseudo inodes are mountable. When they are mounted * path->dentry == path->mnt->mnt_root. In that case don't call d_dname * and instead have d_path return the mounted path. |
c23fbb6bc
|
2989 |
*/ |
f48cfddc6
|
2990 2991 |
if (path->dentry->d_op && path->dentry->d_op->d_dname && (!IS_ROOT(path->dentry) || path->dentry != path->mnt->mnt_root)) |
cf28b4863
|
2992 |
return path->dentry->d_op->d_dname(path->dentry, buf, buflen); |
c23fbb6bc
|
2993 |
|
68f0d9d92
|
2994 2995 |
rcu_read_lock(); get_fs_root_rcu(current->fs, &root); |
02125a826
|
2996 |
error = path_with_deleted(path, &root, &res, &buflen); |
68f0d9d92
|
2997 |
rcu_read_unlock(); |
02125a826
|
2998 |
if (error < 0) |
ffd1f4ed5
|
2999 |
res = ERR_PTR(error); |
1da177e4c
|
3000 3001 |
return res; } |
ec4f86059
|
3002 |
EXPORT_SYMBOL(d_path); |
1da177e4c
|
3003 3004 |
/* |
c23fbb6bc
|
3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 |
* Helper function for dentry_operations.d_dname() members */ char *dynamic_dname(struct dentry *dentry, char *buffer, int buflen, const char *fmt, ...) { va_list args; char temp[64]; int sz; va_start(args, fmt); sz = vsnprintf(temp, sizeof(temp), fmt, args) + 1; va_end(args); if (sz > sizeof(temp) || sz > buflen) return ERR_PTR(-ENAMETOOLONG); buffer += buflen - sz; return memcpy(buffer, temp, sz); } |
118b23022
|
3024 3025 3026 3027 3028 |
char *simple_dname(struct dentry *dentry, char *buffer, int buflen) { char *end = buffer + buflen; /* these dentries are never renamed, so d_lock is not needed */ if (prepend(&end, &buflen, " (deleted)", 11) || |
232d2d60a
|
3029 |
prepend(&end, &buflen, dentry->d_name.name, dentry->d_name.len) || |
118b23022
|
3030 3031 |
prepend(&end, &buflen, "/", 1)) end = ERR_PTR(-ENAMETOOLONG); |
232d2d60a
|
3032 |
return end; |
118b23022
|
3033 |
} |
31bbe16f6
|
3034 |
EXPORT_SYMBOL(simple_dname); |
118b23022
|
3035 |
|
c23fbb6bc
|
3036 |
/* |
6092d0481
|
3037 3038 |
* Write full pathname from the root of the filesystem into the buffer. */ |
f65008015
|
3039 |
static char *__dentry_path(struct dentry *d, char *buf, int buflen) |
6092d0481
|
3040 |
{ |
f65008015
|
3041 |
struct dentry *dentry; |
232d2d60a
|
3042 3043 3044 |
char *end, *retval; int len, seq = 0; int error = 0; |
6092d0481
|
3045 |
|
f65008015
|
3046 3047 |
if (buflen < 2) goto Elong; |
48f5ec21d
|
3048 |
rcu_read_lock(); |
232d2d60a
|
3049 |
restart: |
f65008015
|
3050 |
dentry = d; |
232d2d60a
|
3051 3052 3053 |
end = buf + buflen; len = buflen; prepend(&end, &len, "\0", 1); |
6092d0481
|
3054 3055 3056 |
/* Get '/' right */ retval = end-1; *retval = '/'; |
232d2d60a
|
3057 |
read_seqbegin_or_lock(&rename_lock, &seq); |
cdd16d026
|
3058 3059 |
while (!IS_ROOT(dentry)) { struct dentry *parent = dentry->d_parent; |
6092d0481
|
3060 |
|
6092d0481
|
3061 |
prefetch(parent); |
232d2d60a
|
3062 3063 3064 |
error = prepend_name(&end, &len, &dentry->d_name); if (error) break; |
6092d0481
|
3065 3066 3067 3068 |
retval = end; dentry = parent; } |
48f5ec21d
|
3069 3070 3071 3072 |
if (!(seq & 1)) rcu_read_unlock(); if (need_seqretry(&rename_lock, seq)) { seq = 1; |
232d2d60a
|
3073 |
goto restart; |
48f5ec21d
|
3074 3075 |
} done_seqretry(&rename_lock, seq); |
232d2d60a
|
3076 3077 |
if (error) goto Elong; |
c103135c1
|
3078 3079 3080 3081 |
return retval; Elong: return ERR_PTR(-ENAMETOOLONG); } |
ec2447c27
|
3082 3083 3084 |
char *dentry_path_raw(struct dentry *dentry, char *buf, int buflen) { |
232d2d60a
|
3085 |
return __dentry_path(dentry, buf, buflen); |
ec2447c27
|
3086 3087 |
} EXPORT_SYMBOL(dentry_path_raw); |
c103135c1
|
3088 3089 3090 3091 3092 |
char *dentry_path(struct dentry *dentry, char *buf, int buflen) { char *p = NULL; char *retval; |
c103135c1
|
3093 3094 3095 3096 3097 3098 3099 |
if (d_unlinked(dentry)) { p = buf + buflen; if (prepend(&p, &buflen, "//deleted", 10) != 0) goto Elong; buflen++; } retval = __dentry_path(dentry, buf, buflen); |
c103135c1
|
3100 3101 |
if (!IS_ERR(retval) && p) *p = '/'; /* restore '/' overriden with '\0' */ |
6092d0481
|
3102 3103 |
return retval; Elong: |
6092d0481
|
3104 3105 |
return ERR_PTR(-ENAMETOOLONG); } |
8b19e3418
|
3106 3107 |
static void get_fs_root_and_pwd_rcu(struct fs_struct *fs, struct path *root, struct path *pwd) |
5762482f5
|
3108 |
{ |
8b19e3418
|
3109 3110 3111 3112 3113 3114 3115 |
unsigned seq; do { seq = read_seqcount_begin(&fs->seq); *root = fs->root; *pwd = fs->pwd; } while (read_seqcount_retry(&fs->seq, seq)); |
5762482f5
|
3116 |
} |
6092d0481
|
3117 |
/* |
1da177e4c
|
3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 |
* NOTE! The user-level library version returns a * character pointer. The kernel system call just * returns the length of the buffer filled (which * includes the ending '\0' character), or a negative * error value. So libc would do something like * * char *getcwd(char * buf, size_t size) * { * int retval; * * retval = sys_getcwd(buf, size); * if (retval >= 0) * return buf; * errno = -retval; * return NULL; * } */ |
3cdad4288
|
3135 |
SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size) |
1da177e4c
|
3136 |
{ |
552ce544e
|
3137 |
int error; |
6ac08c39a
|
3138 |
struct path pwd, root; |
3272c544d
|
3139 |
char *page = __getname(); |
1da177e4c
|
3140 3141 3142 |
if (!page) return -ENOMEM; |
8b19e3418
|
3143 3144 |
rcu_read_lock(); get_fs_root_and_pwd_rcu(current->fs, &root, &pwd); |
1da177e4c
|
3145 |
|
552ce544e
|
3146 |
error = -ENOENT; |
f3da392e9
|
3147 |
if (!d_unlinked(pwd.dentry)) { |
552ce544e
|
3148 |
unsigned long len; |
3272c544d
|
3149 3150 |
char *cwd = page + PATH_MAX; int buflen = PATH_MAX; |
1da177e4c
|
3151 |
|
8df9d1a41
|
3152 |
prepend(&cwd, &buflen, "\0", 1); |
02125a826
|
3153 |
error = prepend_path(&pwd, &root, &cwd, &buflen); |
ff812d724
|
3154 |
rcu_read_unlock(); |
552ce544e
|
3155 |
|
02125a826
|
3156 |
if (error < 0) |
552ce544e
|
3157 |
goto out; |
8df9d1a41
|
3158 |
/* Unreachable from current root */ |
02125a826
|
3159 |
if (error > 0) { |
8df9d1a41
|
3160 3161 3162 3163 |
error = prepend_unreachable(&cwd, &buflen); if (error) goto out; } |
552ce544e
|
3164 |
error = -ERANGE; |
3272c544d
|
3165 |
len = PATH_MAX + page - cwd; |
552ce544e
|
3166 3167 3168 3169 3170 |
if (len <= size) { error = len; if (copy_to_user(buf, cwd, len)) error = -EFAULT; } |
949854d02
|
3171 |
} else { |
ff812d724
|
3172 |
rcu_read_unlock(); |
949854d02
|
3173 |
} |
1da177e4c
|
3174 3175 |
out: |
3272c544d
|
3176 |
__putname(page); |
1da177e4c
|
3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 |
return error; } /* * Test whether new_dentry is a subdirectory of old_dentry. * * Trivially implemented using the dcache structure */ /** * is_subdir - is new dentry a subdirectory of old_dentry * @new_dentry: new dentry * @old_dentry: old dentry * * Returns 1 if new_dentry is a subdirectory of the parent (at any depth). * Returns 0 otherwise. * Caller must ensure that "new_dentry" is pinned before calling is_subdir() */ |
e2761a116
|
3196 |
int is_subdir(struct dentry *new_dentry, struct dentry *old_dentry) |
1da177e4c
|
3197 3198 |
{ int result; |
949854d02
|
3199 |
unsigned seq; |
1da177e4c
|
3200 |
|
e2761a116
|
3201 3202 |
if (new_dentry == old_dentry) return 1; |
e2761a116
|
3203 |
do { |
1da177e4c
|
3204 |
/* for restarting inner loop in case of seq retry */ |
1da177e4c
|
3205 |
seq = read_seqbegin(&rename_lock); |
949854d02
|
3206 3207 3208 3209 3210 |
/* * Need rcu_readlock to protect against the d_parent trashing * due to d_move */ rcu_read_lock(); |
e2761a116
|
3211 |
if (d_ancestor(old_dentry, new_dentry)) |
1da177e4c
|
3212 |
result = 1; |
e2761a116
|
3213 3214 |
else result = 0; |
949854d02
|
3215 |
rcu_read_unlock(); |
1da177e4c
|
3216 |
} while (read_seqretry(&rename_lock, seq)); |
1da177e4c
|
3217 3218 3219 |
return result; } |
db14fc3ab
|
3220 |
static enum d_walk_ret d_genocide_kill(void *data, struct dentry *dentry) |
1da177e4c
|
3221 |
{ |
db14fc3ab
|
3222 3223 3224 3225 |
struct dentry *root = data; if (dentry != root) { if (d_unhashed(dentry) || !dentry->d_inode) return D_WALK_SKIP; |
1da177e4c
|
3226 |
|
01ddc4ede
|
3227 3228 3229 3230 |
if (!(dentry->d_flags & DCACHE_GENOCIDE)) { dentry->d_flags |= DCACHE_GENOCIDE; dentry->d_lockref.count--; } |
1da177e4c
|
3231 |
} |
db14fc3ab
|
3232 3233 |
return D_WALK_CONTINUE; } |
58db63d08
|
3234 |
|
db14fc3ab
|
3235 3236 3237 |
void d_genocide(struct dentry *parent) { d_walk(parent, parent, d_genocide_kill, NULL); |
1da177e4c
|
3238 |
} |
60545d0d4
|
3239 |
void d_tmpfile(struct dentry *dentry, struct inode *inode) |
1da177e4c
|
3240 |
{ |
60545d0d4
|
3241 3242 |
inode_dec_link_count(inode); BUG_ON(dentry->d_name.name != dentry->d_iname || |
946e51f2b
|
3243 |
!hlist_unhashed(&dentry->d_u.d_alias) || |
60545d0d4
|
3244 3245 3246 3247 3248 3249 3250 3251 |
!d_unlinked(dentry)); spin_lock(&dentry->d_parent->d_lock); spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); dentry->d_name.len = sprintf(dentry->d_iname, "#%llu", (unsigned long long)inode->i_ino); spin_unlock(&dentry->d_lock); spin_unlock(&dentry->d_parent->d_lock); d_instantiate(dentry, inode); |
1da177e4c
|
3252 |
} |
60545d0d4
|
3253 |
EXPORT_SYMBOL(d_tmpfile); |
1da177e4c
|
3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 |
static __initdata unsigned long dhash_entries; static int __init set_dhash_entries(char *str) { if (!str) return 0; dhash_entries = simple_strtoul(str, &str, 0); return 1; } __setup("dhash_entries=", set_dhash_entries); static void __init dcache_init_early(void) { |
074b85175
|
3267 |
unsigned int loop; |
1da177e4c
|
3268 3269 3270 3271 3272 3273 3274 3275 3276 |
/* If hashes are distributed across NUMA nodes, defer * hash allocation until vmalloc space is available. */ if (hashdist) return; dentry_hashtable = alloc_large_system_hash("Dentry cache", |
b07ad9967
|
3277 |
sizeof(struct hlist_bl_head), |
1da177e4c
|
3278 3279 3280 3281 3282 |
dhash_entries, 13, HASH_EARLY, &d_hash_shift, &d_hash_mask, |
31fe62b95
|
3283 |
0, |
1da177e4c
|
3284 |
0); |
074b85175
|
3285 |
for (loop = 0; loop < (1U << d_hash_shift); loop++) |
b07ad9967
|
3286 |
INIT_HLIST_BL_HEAD(dentry_hashtable + loop); |
1da177e4c
|
3287 |
} |
74bf17cff
|
3288 |
static void __init dcache_init(void) |
1da177e4c
|
3289 |
{ |
074b85175
|
3290 |
unsigned int loop; |
1da177e4c
|
3291 3292 3293 3294 3295 3296 |
/* * A constructor could be added for stable state like the lists, * but it is probably not worth it because of the cache nature * of the dcache. */ |
0a31bd5f2
|
3297 3298 |
dentry_cache = KMEM_CACHE(dentry, SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD); |
1da177e4c
|
3299 3300 3301 3302 3303 3304 3305 |
/* Hash may have been set up in dcache_init_early */ if (!hashdist) return; dentry_hashtable = alloc_large_system_hash("Dentry cache", |
b07ad9967
|
3306 |
sizeof(struct hlist_bl_head), |
1da177e4c
|
3307 3308 3309 3310 3311 |
dhash_entries, 13, 0, &d_hash_shift, &d_hash_mask, |
31fe62b95
|
3312 |
0, |
1da177e4c
|
3313 |
0); |
074b85175
|
3314 |
for (loop = 0; loop < (1U << d_hash_shift); loop++) |
b07ad9967
|
3315 |
INIT_HLIST_BL_HEAD(dentry_hashtable + loop); |
1da177e4c
|
3316 3317 3318 |
} /* SLAB cache for __getname() consumers */ |
e18b890bb
|
3319 |
struct kmem_cache *names_cachep __read_mostly; |
ec4f86059
|
3320 |
EXPORT_SYMBOL(names_cachep); |
1da177e4c
|
3321 |
|
1da177e4c
|
3322 |
EXPORT_SYMBOL(d_genocide); |
1da177e4c
|
3323 3324 3325 3326 3327 |
void __init vfs_caches_init_early(void) { dcache_init_early(); inode_init_early(); } |
4248b0da4
|
3328 |
void __init vfs_caches_init(void) |
1da177e4c
|
3329 |
{ |
1da177e4c
|
3330 |
names_cachep = kmem_cache_create("names_cache", PATH_MAX, 0, |
20c2df83d
|
3331 |
SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); |
1da177e4c
|
3332 |
|
74bf17cff
|
3333 3334 |
dcache_init(); inode_init(); |
4248b0da4
|
3335 3336 |
files_init(); files_maxfiles_init(); |
74bf17cff
|
3337 |
mnt_init(); |
1da177e4c
|
3338 3339 3340 |
bdev_cache_init(); chrdev_init(); } |