Blame view
fs/eventpoll.c
59 KB
1da177e4c
|
1 |
/* |
5071f97ec
|
2 3 |
* fs/eventpoll.c (Efficient event retrieval implementation) * Copyright (C) 2001,...,2009 Davide Libenzi |
1da177e4c
|
4 5 6 7 8 9 10 11 12 |
* * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * Davide Libenzi <davidel@xmailserver.org> * */ |
1da177e4c
|
13 14 15 16 17 18 19 20 21 22 |
#include <linux/init.h> #include <linux/kernel.h> #include <linux/sched.h> #include <linux/fs.h> #include <linux/file.h> #include <linux/signal.h> #include <linux/errno.h> #include <linux/mm.h> #include <linux/slab.h> #include <linux/poll.h> |
1da177e4c
|
23 24 25 26 27 |
#include <linux/string.h> #include <linux/list.h> #include <linux/hash.h> #include <linux/spinlock.h> #include <linux/syscalls.h> |
1da177e4c
|
28 29 30 31 32 |
#include <linux/rbtree.h> #include <linux/wait.h> #include <linux/eventpoll.h> #include <linux/mount.h> #include <linux/bitops.h> |
144efe3e3
|
33 |
#include <linux/mutex.h> |
da66f7cb0
|
34 |
#include <linux/anon_inodes.h> |
4d7e30d98
|
35 |
#include <linux/device.h> |
1da177e4c
|
36 |
#include <asm/uaccess.h> |
1da177e4c
|
37 38 |
#include <asm/io.h> #include <asm/mman.h> |
60063497a
|
39 |
#include <linux/atomic.h> |
138d22b58
|
40 41 |
#include <linux/proc_fs.h> #include <linux/seq_file.h> |
35280bd4a
|
42 |
#include <linux/compat.h> |
ae10b2b4e
|
43 |
#include <linux/rculist.h> |
1da177e4c
|
44 |
|
1da177e4c
|
45 46 47 48 |
/* * LOCKING: * There are three level of locking required by epoll : * |
144efe3e3
|
49 |
* 1) epmutex (mutex) |
c7ea76302
|
50 51 |
* 2) ep->mtx (mutex) * 3) ep->lock (spinlock) |
1da177e4c
|
52 53 54 55 56 57 58 59 60 |
* * The acquire order is the one listed above, from 1 to 3. * We need a spinlock (ep->lock) because we manipulate objects * from inside the poll callback, that might be triggered from * a wake_up() that in turn might be called from IRQ context. * So we can't sleep inside the poll callback and hence we need * a spinlock. During the event transfer loop (from kernel to * user space) we could end up sleeping due a copy_to_user(), so * we need a lock that will allow us to sleep. This lock is a |
d47de16c7
|
61 62 63 64 65 |
* mutex (ep->mtx). It is acquired during the event transfer loop, * during epoll_ctl(EPOLL_CTL_DEL) and during eventpoll_release_file(). * Then we also need a global mutex to serialize eventpoll_release_file() * and ep_free(). * This mutex is acquired by ep_free() during the epoll file |
1da177e4c
|
66 67 |
* cleanup path and it is also acquired by eventpoll_release_file() * if a file has been pushed inside an epoll set and it is then |
bf6a41db7
|
68 |
* close()d without a previous call to epoll_ctl(EPOLL_CTL_DEL). |
22bacca48
|
69 70 71 72 73 74 75 |
* It is also acquired when inserting an epoll fd onto another epoll * fd. We do this so that we walk the epoll tree and ensure that this * insertion does not create a cycle of epoll file descriptors, which * could lead to deadlock. We need a global mutex to prevent two * simultaneous inserts (A into B and B into A) from racing and * constructing a cycle without either insert observing that it is * going to. |
d8805e633
|
76 77 78 79 80 81 82 83 84 |
* It is necessary to acquire multiple "ep->mtx"es at once in the * case when one epoll fd is added to another. In this case, we * always acquire the locks in the order of nesting (i.e. after * epoll_ctl(e1, EPOLL_CTL_ADD, e2), e1->mtx will always be acquired * before e2->mtx). Since we disallow cycles of epoll file * descriptors, this ensures that the mutexes are well-ordered. In * order to communicate this nesting to lockdep, when walking a tree * of epoll file descriptors, we use the current recursion depth as * the lockdep subkey. |
d47de16c7
|
85 86 87 |
* It is possible to drop the "ep->mtx" and to use the global * mutex "epmutex" (together with "ep->lock") to have it working, * but having "ep->mtx" will make the interface more scalable. |
144efe3e3
|
88 |
* Events that require holding "epmutex" are very rare, while for |
d47de16c7
|
89 90 |
* normal operations the epoll private "ep->mtx" will guarantee * a better scalability. |
1da177e4c
|
91 |
*/ |
1da177e4c
|
92 |
/* Epoll private bits inside the event mask */ |
4d7e30d98
|
93 |
#define EP_PRIVATE_BITS (EPOLLWAKEUP | EPOLLONESHOT | EPOLLET) |
1da177e4c
|
94 |
|
5071f97ec
|
95 96 |
/* Maximum number of nesting allowed inside epoll sets */ #define EP_MAX_NESTS 4 |
1da177e4c
|
97 |
|
b611967de
|
98 |
#define EP_MAX_EVENTS (INT_MAX / sizeof(struct epoll_event)) |
d47de16c7
|
99 |
#define EP_UNACTIVE_PTR ((void *) -1L) |
7ef9964e6
|
100 |
#define EP_ITEM_COST (sizeof(struct epitem) + sizeof(struct eppoll_entry)) |
1da177e4c
|
101 102 103 |
struct epoll_filefd { struct file *file; int fd; |
39732ca5a
|
104 |
} __packed; |
1da177e4c
|
105 106 |
/* |
5071f97ec
|
107 108 |
* Structure used to track possible nested calls, for too deep recursions * and loop cycles. |
1da177e4c
|
109 |
*/ |
5071f97ec
|
110 |
struct nested_call_node { |
1da177e4c
|
111 |
struct list_head llink; |
5071f97ec
|
112 |
void *cookie; |
3fe4a975d
|
113 |
void *ctx; |
1da177e4c
|
114 115 116 |
}; /* |
5071f97ec
|
117 118 |
* This structure is used as collector for nested calls, to check for * maximum recursion dept and loop cycles. |
1da177e4c
|
119 |
*/ |
5071f97ec
|
120 121 |
struct nested_calls { struct list_head tasks_call_list; |
1da177e4c
|
122 123 124 125 |
spinlock_t lock; }; /* |
d47de16c7
|
126 127 |
* Each file descriptor added to the eventpoll interface will * have an entry of this type linked to the "rbr" RB tree. |
39732ca5a
|
128 129 |
* Avoid increasing the size of this struct, there can be many thousands * of these on a server and we do not want this to take another cache line. |
d47de16c7
|
130 131 |
*/ struct epitem { |
ae10b2b4e
|
132 133 134 135 136 137 |
union { /* RB tree node links this structure to the eventpoll RB tree */ struct rb_node rbn; /* Used to free the struct epitem */ struct rcu_head rcu; }; |
d47de16c7
|
138 139 140 |
/* List header used to link this structure to the eventpoll ready list */ struct list_head rdllink; |
c7ea76302
|
141 142 143 144 145 |
/* * Works together "struct eventpoll"->ovflist in keeping the * single linked chain of items. */ struct epitem *next; |
d47de16c7
|
146 147 148 149 150 151 152 153 154 155 156 |
/* The file descriptor information this item refers to */ struct epoll_filefd ffd; /* Number of active wait queue attached to poll operations */ int nwait; /* List containing poll wait queues */ struct list_head pwqlist; /* The "container" of this item */ struct eventpoll *ep; |
d47de16c7
|
157 158 |
/* List header used to link this item to the "struct file" items list */ struct list_head fllink; |
4d7e30d98
|
159 |
/* wakeup_source used when EPOLLWAKEUP is set */ |
eea1d5859
|
160 |
struct wakeup_source __rcu *ws; |
4d7e30d98
|
161 |
|
c7ea76302
|
162 163 |
/* The structure that describe the interested events and the source fd */ struct epoll_event event; |
d47de16c7
|
164 165 166 |
}; /* |
1da177e4c
|
167 |
* This structure is stored inside the "private_data" member of the file |
bf6a41db7
|
168 |
* structure and represents the main data structure for the eventpoll |
1da177e4c
|
169 170 171 |
* interface. */ struct eventpoll { |
bf6a41db7
|
172 |
/* Protect the access to this structure */ |
c7ea76302
|
173 |
spinlock_t lock; |
1da177e4c
|
174 175 |
/* |
d47de16c7
|
176 177 178 179 |
* This mutex is used to ensure that files are not removed * while epoll is using them. This is held during the event * collection loop, the file cleanup path, the epoll file exit * code and the ctl operations. |
1da177e4c
|
180 |
*/ |
d47de16c7
|
181 |
struct mutex mtx; |
1da177e4c
|
182 183 184 185 186 187 188 189 190 |
/* Wait queue used by sys_epoll_wait() */ wait_queue_head_t wq; /* Wait queue used by file->poll() */ wait_queue_head_t poll_wait; /* List of ready file descriptors */ struct list_head rdllist; |
67647d0fb
|
191 |
/* RB tree root used to store monitored fd structs */ |
1da177e4c
|
192 |
struct rb_root rbr; |
d47de16c7
|
193 194 195 |
/* * This is a single linked list that chains all the "struct epitem" that |
25985edce
|
196 |
* happened while transferring ready events to userspace w/out |
d47de16c7
|
197 198 199 |
* holding ->lock. */ struct epitem *ovflist; |
7ef9964e6
|
200 |
|
4d7e30d98
|
201 202 |
/* wakeup_source used when ep_scan_ready_list is running */ struct wakeup_source *ws; |
7ef9964e6
|
203 204 |
/* The user that created the eventpoll descriptor */ struct user_struct *user; |
28d82dc1c
|
205 206 207 208 209 210 |
struct file *file; /* used to optimize loop detection check */ int visited; struct list_head visited_list_link; |
1da177e4c
|
211 212 213 214 215 216 217 218 |
}; /* Wait structure used by the poll hooks */ struct eppoll_entry { /* List header used to link this structure to the "struct epitem" */ struct list_head llink; /* The "base" pointer is set to the container "struct epitem" */ |
4f0989dbf
|
219 |
struct epitem *base; |
1da177e4c
|
220 221 222 223 224 225 226 227 228 229 |
/* * Wait queue item that will be linked to the target file wait * queue head. */ wait_queue_t wait; /* The wait queue head that linked the "wait" wait queue item */ wait_queue_head_t *whead; }; |
1da177e4c
|
230 231 232 233 234 |
/* Wrapper struct used by poll queueing */ struct ep_pqueue { poll_table pt; struct epitem *epi; }; |
5071f97ec
|
235 236 237 238 239 |
/* Used by the ep_send_events() function as callback private data */ struct ep_send_events_data { int maxevents; struct epoll_event __user *events; }; |
1da177e4c
|
240 |
/* |
7ef9964e6
|
241 242 |
* Configuration options available inside /proc/sys/fs/epoll/ */ |
7ef9964e6
|
243 |
/* Maximum number of epoll watched descriptors, per user */ |
52bd19f76
|
244 |
static long max_user_watches __read_mostly; |
7ef9964e6
|
245 246 |
/* |
d47de16c7
|
247 |
* This mutex is used to serialize ep_free() and eventpoll_release_file(). |
1da177e4c
|
248 |
*/ |
7ef9964e6
|
249 |
static DEFINE_MUTEX(epmutex); |
1da177e4c
|
250 |
|
22bacca48
|
251 252 |
/* Used to check for epoll file descriptor inclusion loops */ static struct nested_calls poll_loop_ncalls; |
5071f97ec
|
253 254 255 256 257 |
/* Used for safe wake up implementation */ static struct nested_calls poll_safewake_ncalls; /* Used to call file's f_op->poll() under the nested calls boundaries */ static struct nested_calls poll_readywalk_ncalls; |
1da177e4c
|
258 259 |
/* Slab cache used to allocate "struct epitem" */ |
e18b890bb
|
260 |
static struct kmem_cache *epi_cache __read_mostly; |
1da177e4c
|
261 262 |
/* Slab cache used to allocate "struct eppoll_entry" */ |
e18b890bb
|
263 |
static struct kmem_cache *pwq_cache __read_mostly; |
1da177e4c
|
264 |
|
28d82dc1c
|
265 266 267 268 269 270 271 272 |
/* Visited nodes during ep_loop_check(), so we can unset them when we finish */ static LIST_HEAD(visited_list); /* * List of files with newly added links, where we may need to limit the number * of emanating paths. Protected by the epmutex. */ static LIST_HEAD(tfile_check_list); |
7ef9964e6
|
273 274 275 |
#ifdef CONFIG_SYSCTL #include <linux/sysctl.h> |
52bd19f76
|
276 277 |
static long zero; static long long_max = LONG_MAX; |
7ef9964e6
|
278 |
|
1f7e0616c
|
279 |
struct ctl_table epoll_table[] = { |
7ef9964e6
|
280 |
{ |
7ef9964e6
|
281 282 |
.procname = "max_user_watches", .data = &max_user_watches, |
52bd19f76
|
283 |
.maxlen = sizeof(max_user_watches), |
7ef9964e6
|
284 |
.mode = 0644, |
52bd19f76
|
285 |
.proc_handler = proc_doulongvec_minmax, |
7ef9964e6
|
286 |
.extra1 = &zero, |
52bd19f76
|
287 |
.extra2 = &long_max, |
7ef9964e6
|
288 |
}, |
ab09203e3
|
289 |
{ } |
7ef9964e6
|
290 291 |
}; #endif /* CONFIG_SYSCTL */ |
28d82dc1c
|
292 293 294 295 296 297 |
static const struct file_operations eventpoll_fops; static inline int is_file_epoll(struct file *f) { return f->f_op == &eventpoll_fops; } |
b030a4dd6
|
298 |
|
67647d0fb
|
299 |
/* Setup the structure that is used as key for the RB tree */ |
b030a4dd6
|
300 301 302 303 304 305 |
static inline void ep_set_ffd(struct epoll_filefd *ffd, struct file *file, int fd) { ffd->file = file; ffd->fd = fd; } |
67647d0fb
|
306 |
/* Compare RB tree keys */ |
b030a4dd6
|
307 308 309 310 311 312 |
static inline int ep_cmp_ffd(struct epoll_filefd *p1, struct epoll_filefd *p2) { return (p1->file > p2->file ? +1: (p1->file < p2->file ? -1 : p1->fd - p2->fd)); } |
b030a4dd6
|
313 314 315 316 317 |
/* Tells us if the item is currently linked */ static inline int ep_is_linked(struct list_head *p) { return !list_empty(p); } |
971316f05
|
318 319 320 321 |
static inline struct eppoll_entry *ep_pwq_from_wait(wait_queue_t *p) { return container_of(p, struct eppoll_entry, wait); } |
b030a4dd6
|
322 |
/* Get the "struct epitem" from a wait queue pointer */ |
cdac75e6f
|
323 |
static inline struct epitem *ep_item_from_wait(wait_queue_t *p) |
b030a4dd6
|
324 325 326 327 328 |
{ return container_of(p, struct eppoll_entry, wait)->base; } /* Get the "struct epitem" from an epoll queue wrapper */ |
cdac75e6f
|
329 |
static inline struct epitem *ep_item_from_epqueue(poll_table *p) |
b030a4dd6
|
330 331 332 333 334 |
{ return container_of(p, struct ep_pqueue, pt)->epi; } /* Tells if the epoll_ctl(2) operation needs an event copy from userspace */ |
6192bd536
|
335 |
static inline int ep_op_has_event(int op) |
b030a4dd6
|
336 |
{ |
a80a6b85b
|
337 |
return op != EPOLL_CTL_DEL; |
b030a4dd6
|
338 |
} |
1da177e4c
|
339 |
/* Initialize the poll safe wake up structure */ |
5071f97ec
|
340 |
static void ep_nested_calls_init(struct nested_calls *ncalls) |
1da177e4c
|
341 |
{ |
5071f97ec
|
342 343 |
INIT_LIST_HEAD(&ncalls->tasks_call_list); spin_lock_init(&ncalls->lock); |
1da177e4c
|
344 |
} |
5071f97ec
|
345 |
/** |
3fb0e584a
|
346 347 348 349 350 351 352 353 354 355 356 357 358 |
* ep_events_available - Checks if ready events might be available. * * @ep: Pointer to the eventpoll context. * * Returns: Returns a value different than zero if ready events are available, * or zero otherwise. */ static inline int ep_events_available(struct eventpoll *ep) { return !list_empty(&ep->rdllist) || ep->ovflist != EP_UNACTIVE_PTR; } /** |
5071f97ec
|
359 360 361 362 363 364 365 366 367 368 |
* ep_call_nested - Perform a bound (possibly) nested call, by checking * that the recursion limit is not exceeded, and that * the same nested call (by the meaning of same cookie) is * no re-entered. * * @ncalls: Pointer to the nested_calls structure to be used for this call. * @max_nests: Maximum number of allowed nesting calls. * @nproc: Nested call core function pointer. * @priv: Opaque data to be passed to the @nproc callback. * @cookie: Cookie to be used to identify this nested call. |
3fe4a975d
|
369 |
* @ctx: This instance context. |
5071f97ec
|
370 371 372 |
* * Returns: Returns the code returned by the @nproc callback, or -1 if * the maximum recursion limit has been exceeded. |
1da177e4c
|
373 |
*/ |
5071f97ec
|
374 375 |
static int ep_call_nested(struct nested_calls *ncalls, int max_nests, int (*nproc)(void *, void *, int), void *priv, |
3fe4a975d
|
376 |
void *cookie, void *ctx) |
1da177e4c
|
377 |
{ |
5071f97ec
|
378 |
int error, call_nests = 0; |
1da177e4c
|
379 |
unsigned long flags; |
5071f97ec
|
380 381 382 |
struct list_head *lsthead = &ncalls->tasks_call_list; struct nested_call_node *tncur; struct nested_call_node tnode; |
1da177e4c
|
383 |
|
5071f97ec
|
384 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c
|
385 |
|
5071f97ec
|
386 387 388 389 390 |
/* * Try to see if the current task is already inside this wakeup call. * We use a list here, since the population inside this set is always * very much limited. */ |
b70c39409
|
391 |
list_for_each_entry(tncur, lsthead, llink) { |
3fe4a975d
|
392 |
if (tncur->ctx == ctx && |
5071f97ec
|
393 |
(tncur->cookie == cookie || ++call_nests > max_nests)) { |
1da177e4c
|
394 395 396 397 |
/* * Ops ... loop detected or maximum nest level reached. * We abort this wake by breaking the cycle itself. */ |
abff55cee
|
398 399 |
error = -1; goto out_unlock; |
1da177e4c
|
400 401 |
} } |
5071f97ec
|
402 |
/* Add the current task and cookie to the list */ |
3fe4a975d
|
403 |
tnode.ctx = ctx; |
5071f97ec
|
404 |
tnode.cookie = cookie; |
1da177e4c
|
405 |
list_add(&tnode.llink, lsthead); |
5071f97ec
|
406 |
spin_unlock_irqrestore(&ncalls->lock, flags); |
1da177e4c
|
407 |
|
5071f97ec
|
408 409 |
/* Call the nested function */ error = (*nproc)(priv, cookie, call_nests); |
1da177e4c
|
410 411 |
/* Remove the current task from the list */ |
5071f97ec
|
412 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c
|
413 |
list_del(&tnode.llink); |
3fe4a975d
|
414 |
out_unlock: |
5071f97ec
|
415 416 417 418 |
spin_unlock_irqrestore(&ncalls->lock, flags); return error; } |
02edc6fc4
|
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 |
/* * As described in commit 0ccf831cb lockdep: annotate epoll * the use of wait queues used by epoll is done in a very controlled * manner. Wake ups can nest inside each other, but are never done * with the same locking. For example: * * dfd = socket(...); * efd1 = epoll_create(); * efd2 = epoll_create(); * epoll_ctl(efd1, EPOLL_CTL_ADD, dfd, ...); * epoll_ctl(efd2, EPOLL_CTL_ADD, efd1, ...); * * When a packet arrives to the device underneath "dfd", the net code will * issue a wake_up() on its poll wake list. Epoll (efd1) has installed a * callback wakeup entry on that queue, and the wake_up() performed by the * "dfd" net code will end up in ep_poll_callback(). At this point epoll * (efd1) notices that it may have some event ready, so it needs to wake up * the waiters on its poll wait list (efd2). So it calls ep_poll_safewake() * that ends up in another wake_up(), after having checked about the * recursion constraints. That are, no more than EP_MAX_POLLWAKE_NESTS, to * avoid stack blasting. * * When CONFIG_DEBUG_LOCK_ALLOC is enabled, make sure lockdep can handle * this special case of epoll. */ |
2dfa4eeab
|
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 |
#ifdef CONFIG_DEBUG_LOCK_ALLOC static inline void ep_wake_up_nested(wait_queue_head_t *wqueue, unsigned long events, int subclass) { unsigned long flags; spin_lock_irqsave_nested(&wqueue->lock, flags, subclass); wake_up_locked_poll(wqueue, events); spin_unlock_irqrestore(&wqueue->lock, flags); } #else static inline void ep_wake_up_nested(wait_queue_head_t *wqueue, unsigned long events, int subclass) { wake_up_poll(wqueue, events); } #endif |
5071f97ec
|
461 462 |
static int ep_poll_wakeup_proc(void *priv, void *cookie, int call_nests) { |
2dfa4eeab
|
463 464 |
ep_wake_up_nested((wait_queue_head_t *) cookie, POLLIN, 1 + call_nests); |
5071f97ec
|
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 |
return 0; } /* * Perform a safe wake up of the poll wait list. The problem is that * with the new callback'd wake up system, it is possible that the * poll callback is reentered from inside the call to wake_up() done * on the poll wait queue head. The rule is that we cannot reenter the * wake up code from the same task more than EP_MAX_NESTS times, * and we cannot reenter the same wait queue head at all. This will * enable to have a hierarchy of epoll file descriptor of no more than * EP_MAX_NESTS deep. */ static void ep_poll_safewake(wait_queue_head_t *wq) { |
3fe4a975d
|
480 |
int this_cpu = get_cpu(); |
5071f97ec
|
481 |
ep_call_nested(&poll_safewake_ncalls, EP_MAX_NESTS, |
3fe4a975d
|
482 483 484 |
ep_poll_wakeup_proc, NULL, wq, (void *) (long) this_cpu); put_cpu(); |
1da177e4c
|
485 |
} |
971316f05
|
486 487 488 489 490 491 492 493 494 495 496 |
static void ep_remove_wait_queue(struct eppoll_entry *pwq) { wait_queue_head_t *whead; rcu_read_lock(); /* If it is cleared by POLLFREE, it should be rcu-safe */ whead = rcu_dereference(pwq->whead); if (whead) remove_wait_queue(whead, &pwq->wait); rcu_read_unlock(); } |
1da177e4c
|
497 |
/* |
d1bc90dd5
|
498 499 500 |
* This function unregisters poll callbacks from the associated file * descriptor. Must be called with "mtx" held (or "epmutex" if called from * ep_free). |
1da177e4c
|
501 |
*/ |
7699acd13
|
502 |
static void ep_unregister_pollwait(struct eventpoll *ep, struct epitem *epi) |
1da177e4c
|
503 |
{ |
7699acd13
|
504 505 |
struct list_head *lsthead = &epi->pwqlist; struct eppoll_entry *pwq; |
1da177e4c
|
506 |
|
d1bc90dd5
|
507 508 |
while (!list_empty(lsthead)) { pwq = list_first_entry(lsthead, struct eppoll_entry, llink); |
1da177e4c
|
509 |
|
d1bc90dd5
|
510 |
list_del(&pwq->llink); |
971316f05
|
511 |
ep_remove_wait_queue(pwq); |
d1bc90dd5
|
512 |
kmem_cache_free(pwq_cache, pwq); |
1da177e4c
|
513 |
} |
1da177e4c
|
514 |
} |
eea1d5859
|
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 |
/* call only when ep->mtx is held */ static inline struct wakeup_source *ep_wakeup_source(struct epitem *epi) { return rcu_dereference_check(epi->ws, lockdep_is_held(&epi->ep->mtx)); } /* call only when ep->mtx is held */ static inline void ep_pm_stay_awake(struct epitem *epi) { struct wakeup_source *ws = ep_wakeup_source(epi); if (ws) __pm_stay_awake(ws); } static inline bool ep_has_wakeup_source(struct epitem *epi) { return rcu_access_pointer(epi->ws) ? true : false; } /* call when ep->mtx cannot be held (ep_poll_callback) */ static inline void ep_pm_stay_awake_rcu(struct epitem *epi) { struct wakeup_source *ws; rcu_read_lock(); ws = rcu_dereference(epi->ws); if (ws) __pm_stay_awake(ws); rcu_read_unlock(); } |
5071f97ec
|
546 547 548 549 550 551 552 553 |
/** * ep_scan_ready_list - Scans the ready list in a way that makes possible for * the scan code, to call f_op->poll(). Also allows for * O(NumReady) performance. * * @ep: Pointer to the epoll private data structure. * @sproc: Pointer to the scan callback. * @priv: Private opaque data passed to the @sproc callback. |
d8805e633
|
554 |
* @depth: The current depth of recursive f_op->poll calls. |
67347fe4e
|
555 |
* @ep_locked: caller already holds ep->mtx |
5071f97ec
|
556 557 558 559 560 561 |
* * Returns: The same integer error code returned by the @sproc callback. */ static int ep_scan_ready_list(struct eventpoll *ep, int (*sproc)(struct eventpoll *, struct list_head *, void *), |
67347fe4e
|
562 |
void *priv, int depth, bool ep_locked) |
5071f97ec
|
563 564 565 566 |
{ int error, pwake = 0; unsigned long flags; struct epitem *epi, *nepi; |
296e236e9
|
567 |
LIST_HEAD(txlist); |
5071f97ec
|
568 569 570 |
/* * We need to lock this because we could be hit by |
e057e15ff
|
571 |
* eventpoll_release_file() and epoll_ctl(). |
5071f97ec
|
572 |
*/ |
67347fe4e
|
573 574 575 |
if (!ep_locked) mutex_lock_nested(&ep->mtx, depth); |
5071f97ec
|
576 577 578 579 580 581 582 583 584 585 |
/* * Steal the ready list, and re-init the original one to the * empty list. Also, set ep->ovflist to NULL so that events * happening while looping w/out locks, are not lost. We cannot * have the poll callback to queue directly on ep->rdllist, * because we want the "sproc" callback to be able to do it * in a lockless way. */ spin_lock_irqsave(&ep->lock, flags); |
296e236e9
|
586 |
list_splice_init(&ep->rdllist, &txlist); |
5071f97ec
|
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 |
ep->ovflist = NULL; spin_unlock_irqrestore(&ep->lock, flags); /* * Now call the callback function. */ error = (*sproc)(ep, &txlist, priv); spin_lock_irqsave(&ep->lock, flags); /* * During the time we spent inside the "sproc" callback, some * other events might have been queued by the poll callback. * We re-insert them inside the main ready-list here. */ for (nepi = ep->ovflist; (epi = nepi) != NULL; nepi = epi->next, epi->next = EP_UNACTIVE_PTR) { /* * We need to check if the item is already in the list. * During the "sproc" callback execution time, items are * queued into ->ovflist but the "txlist" might already * contain them, and the list_splice() below takes care of them. */ |
4d7e30d98
|
609 |
if (!ep_is_linked(&epi->rdllink)) { |
5071f97ec
|
610 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
611 |
ep_pm_stay_awake(epi); |
4d7e30d98
|
612 |
} |
5071f97ec
|
613 614 615 616 617 618 619 620 621 622 623 624 |
} /* * We need to set back ep->ovflist to EP_UNACTIVE_PTR, so that after * releasing the lock, events will be queued in the normal way inside * ep->rdllist. */ ep->ovflist = EP_UNACTIVE_PTR; /* * Quickly re-inject items left on "txlist". */ list_splice(&txlist, &ep->rdllist); |
4d7e30d98
|
625 |
__pm_relax(ep->ws); |
5071f97ec
|
626 627 628 |
if (!list_empty(&ep->rdllist)) { /* |
296e236e9
|
629 630 |
* Wake up (if active) both the eventpoll wait list and * the ->poll() wait list (delayed after we release the lock). |
5071f97ec
|
631 632 633 634 635 636 637 |
*/ if (waitqueue_active(&ep->wq)) wake_up_locked(&ep->wq); if (waitqueue_active(&ep->poll_wait)) pwake++; } spin_unlock_irqrestore(&ep->lock, flags); |
67347fe4e
|
638 639 |
if (!ep_locked) mutex_unlock(&ep->mtx); |
5071f97ec
|
640 641 642 643 644 645 646 |
/* We have to call this outside the lock */ if (pwake) ep_poll_safewake(&ep->poll_wait); return error; } |
ae10b2b4e
|
647 648 649 650 651 |
static void epi_rcu_free(struct rcu_head *head) { struct epitem *epi = container_of(head, struct epitem, rcu); kmem_cache_free(epi_cache, epi); } |
1da177e4c
|
652 |
/* |
7699acd13
|
653 |
* Removes a "struct epitem" from the eventpoll RB tree and deallocates |
c7ea76302
|
654 |
* all the associated resources. Must be called with "mtx" held. |
7699acd13
|
655 656 657 |
*/ static int ep_remove(struct eventpoll *ep, struct epitem *epi) { |
7699acd13
|
658 659 |
unsigned long flags; struct file *file = epi->ffd.file; |
1da177e4c
|
660 661 |
/* |
7699acd13
|
662 663 664 665 666 667 |
* Removes poll wait queue hooks. We _have_ to do this without holding * the "ep->lock" otherwise a deadlock might occur. This because of the * sequence of the lock acquisition. Here we do "ep->lock" then the wait * queue head lock when unregistering the wait queue. The wakeup callback * will run by holding the wait queue head lock and will call our callback * that will try to get "ep->lock". |
1da177e4c
|
668 |
*/ |
7699acd13
|
669 |
ep_unregister_pollwait(ep, epi); |
1da177e4c
|
670 |
|
7699acd13
|
671 |
/* Remove the current item from the list of epoll hooks */ |
684999149
|
672 |
spin_lock(&file->f_lock); |
ae10b2b4e
|
673 |
list_del_rcu(&epi->fllink); |
684999149
|
674 |
spin_unlock(&file->f_lock); |
1da177e4c
|
675 |
|
cdac75e6f
|
676 |
rb_erase(&epi->rbn, &ep->rbr); |
1da177e4c
|
677 |
|
c7ea76302
|
678 679 680 681 |
spin_lock_irqsave(&ep->lock, flags); if (ep_is_linked(&epi->rdllink)) list_del_init(&epi->rdllink); spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
682 |
|
eea1d5859
|
683 |
wakeup_source_unregister(ep_wakeup_source(epi)); |
ae10b2b4e
|
684 685 686 687 688 689 690 691 |
/* * At this point it is safe to free the eventpoll item. Use the union * field epi->rcu, since we are trying to minimize the size of * 'struct epitem'. The 'rbn' field is no longer in use. Protected by * ep->mtx. The rcu read side, reverse_path_check_proc(), does not make * use of the rbn field. */ call_rcu(&epi->rcu, epi_rcu_free); |
1da177e4c
|
692 |
|
52bd19f76
|
693 |
atomic_long_dec(&ep->user->epoll_watches); |
7ef9964e6
|
694 |
|
c7ea76302
|
695 |
return 0; |
1da177e4c
|
696 |
} |
7699acd13
|
697 |
static void ep_free(struct eventpoll *ep) |
1da177e4c
|
698 |
{ |
7699acd13
|
699 700 |
struct rb_node *rbp; struct epitem *epi; |
1da177e4c
|
701 |
|
7699acd13
|
702 703 |
/* We need to release all tasks waiting for these file */ if (waitqueue_active(&ep->poll_wait)) |
5071f97ec
|
704 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
705 |
|
7699acd13
|
706 707 708 |
/* * We need to lock this because we could be hit by * eventpoll_release_file() while we're freeing the "struct eventpoll". |
d47de16c7
|
709 |
* We do not need to hold "ep->mtx" here because the epoll file |
7699acd13
|
710 711 |
* is on the way to be removed and no one has references to it * anymore. The only hit might come from eventpoll_release_file() but |
25985edce
|
712 |
* holding "epmutex" is sufficient here. |
7699acd13
|
713 714 |
*/ mutex_lock(&epmutex); |
1da177e4c
|
715 716 |
/* |
7699acd13
|
717 |
* Walks through the whole tree by unregistering poll callbacks. |
1da177e4c
|
718 |
*/ |
7699acd13
|
719 720 721 722 |
for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) { epi = rb_entry(rbp, struct epitem, rbn); ep_unregister_pollwait(ep, epi); |
91cf5ab60
|
723 |
cond_resched(); |
7699acd13
|
724 |
} |
1da177e4c
|
725 726 |
/* |
7699acd13
|
727 728 |
* Walks through the whole tree by freeing each "struct epitem". At this * point we are sure no poll callbacks will be lingering around, and also by |
d47de16c7
|
729 |
* holding "epmutex" we can be sure that no file cleanup code will hit |
7699acd13
|
730 |
* us during this operation. So we can avoid the lock on "ep->lock". |
ddf676c38
|
731 732 |
* We do not need to lock ep->mtx, either, we only do it to prevent * a lockdep warning. |
1da177e4c
|
733 |
*/ |
ddf676c38
|
734 |
mutex_lock(&ep->mtx); |
c80544dc0
|
735 |
while ((rbp = rb_first(&ep->rbr)) != NULL) { |
7699acd13
|
736 737 |
epi = rb_entry(rbp, struct epitem, rbn); ep_remove(ep, epi); |
91cf5ab60
|
738 |
cond_resched(); |
7699acd13
|
739 |
} |
ddf676c38
|
740 |
mutex_unlock(&ep->mtx); |
1da177e4c
|
741 |
|
7699acd13
|
742 |
mutex_unlock(&epmutex); |
d47de16c7
|
743 |
mutex_destroy(&ep->mtx); |
7ef9964e6
|
744 |
free_uid(ep->user); |
4d7e30d98
|
745 |
wakeup_source_unregister(ep->ws); |
f0ee9aabb
|
746 |
kfree(ep); |
7699acd13
|
747 |
} |
1da177e4c
|
748 |
|
7699acd13
|
749 750 751 |
static int ep_eventpoll_release(struct inode *inode, struct file *file) { struct eventpoll *ep = file->private_data; |
1da177e4c
|
752 |
|
f0ee9aabb
|
753 |
if (ep) |
7699acd13
|
754 |
ep_free(ep); |
7699acd13
|
755 |
|
7699acd13
|
756 |
return 0; |
1da177e4c
|
757 |
} |
450d89ec0
|
758 759 760 761 762 763 |
static inline unsigned int ep_item_poll(struct epitem *epi, poll_table *pt) { pt->_key = epi->event.events; return epi->ffd.file->f_op->poll(epi->ffd.file, pt) & epi->event.events; } |
296e236e9
|
764 765 |
static int ep_read_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
5071f97ec
|
766 767 |
{ struct epitem *epi, *tmp; |
626cf2366
|
768 |
poll_table pt; |
5071f97ec
|
769 |
|
626cf2366
|
770 |
init_poll_funcptr(&pt, NULL); |
450d89ec0
|
771 |
|
5071f97ec
|
772 |
list_for_each_entry_safe(epi, tmp, head, rdllink) { |
450d89ec0
|
773 |
if (ep_item_poll(epi, &pt)) |
5071f97ec
|
774 |
return POLLIN | POLLRDNORM; |
296e236e9
|
775 |
else { |
5071f97ec
|
776 777 778 779 780 |
/* * Item has been dropped into the ready list by the poll * callback, but it's not actually ready, as far as * caller requested events goes. We can remove it here. */ |
eea1d5859
|
781 |
__pm_relax(ep_wakeup_source(epi)); |
5071f97ec
|
782 |
list_del_init(&epi->rdllink); |
296e236e9
|
783 |
} |
5071f97ec
|
784 785 786 787 |
} return 0; } |
67347fe4e
|
788 789 790 791 792 793 794 |
static void ep_ptable_queue_proc(struct file *file, wait_queue_head_t *whead, poll_table *pt); struct readyevents_arg { struct eventpoll *ep; bool locked; }; |
5071f97ec
|
795 796 |
static int ep_poll_readyevents_proc(void *priv, void *cookie, int call_nests) { |
67347fe4e
|
797 798 799 800 |
struct readyevents_arg *arg = priv; return ep_scan_ready_list(arg->ep, ep_read_events_proc, NULL, call_nests + 1, arg->locked); |
5071f97ec
|
801 |
} |
7699acd13
|
802 803 |
static unsigned int ep_eventpoll_poll(struct file *file, poll_table *wait) { |
5071f97ec
|
804 |
int pollflags; |
7699acd13
|
805 |
struct eventpoll *ep = file->private_data; |
67347fe4e
|
806 807 808 809 810 811 812 813 |
struct readyevents_arg arg; /* * During ep_insert() we already hold the ep->mtx for the tfile. * Prevent re-aquisition. */ arg.locked = wait && (wait->_qproc == ep_ptable_queue_proc); arg.ep = ep; |
1da177e4c
|
814 |
|
7699acd13
|
815 816 |
/* Insert inside our poll wait queue */ poll_wait(file, &ep->poll_wait, wait); |
5071f97ec
|
817 818 819 820 821 822 823 |
/* * Proceed to find out if wanted events are really available inside * the ready list. This need to be done under ep_call_nested() * supervision, since the call to f_op->poll() done on listed files * could re-enter here. */ pollflags = ep_call_nested(&poll_readywalk_ncalls, EP_MAX_NESTS, |
67347fe4e
|
824 |
ep_poll_readyevents_proc, &arg, ep, current); |
7699acd13
|
825 |
|
296e236e9
|
826 |
return pollflags != -1 ? pollflags : 0; |
7699acd13
|
827 |
} |
138d22b58
|
828 |
#ifdef CONFIG_PROC_FS |
a3816ab0e
|
829 |
static void ep_show_fdinfo(struct seq_file *m, struct file *f) |
138d22b58
|
830 831 832 |
{ struct eventpoll *ep = f->private_data; struct rb_node *rbp; |
138d22b58
|
833 834 835 836 |
mutex_lock(&ep->mtx); for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) { struct epitem *epi = rb_entry(rbp, struct epitem, rbn); |
a3816ab0e
|
837 838 839 840 841 |
seq_printf(m, "tfd: %8d events: %8x data: %16llx ", epi->ffd.fd, epi->event.events, (long long)epi->event.data); if (seq_has_overflowed(m)) |
138d22b58
|
842 843 844 |
break; } mutex_unlock(&ep->mtx); |
138d22b58
|
845 846 |
} #endif |
7699acd13
|
847 848 |
/* File callbacks that implement the eventpoll file behaviour */ static const struct file_operations eventpoll_fops = { |
138d22b58
|
849 850 851 |
#ifdef CONFIG_PROC_FS .show_fdinfo = ep_show_fdinfo, #endif |
7699acd13
|
852 |
.release = ep_eventpoll_release, |
6038f373a
|
853 854 |
.poll = ep_eventpoll_poll, .llseek = noop_llseek, |
7699acd13
|
855 |
}; |
b611967de
|
856 |
/* |
7699acd13
|
857 858 859 |
* This is called from eventpoll_release() to unlink files from the eventpoll * interface. We need to have this facility to cleanup correctly files that are * closed without being removed from the eventpoll interface. |
b611967de
|
860 |
*/ |
7699acd13
|
861 |
void eventpoll_release_file(struct file *file) |
b611967de
|
862 |
{ |
7699acd13
|
863 |
struct eventpoll *ep; |
ebe06187b
|
864 |
struct epitem *epi, *next; |
b611967de
|
865 866 |
/* |
684999149
|
867 |
* We don't want to get "file->f_lock" because it is not |
7699acd13
|
868 |
* necessary. It is not necessary because we're in the "struct file" |
25985edce
|
869 |
* cleanup path, and this means that no one is using this file anymore. |
5071f97ec
|
870 |
* So, for example, epoll_ctl() cannot hit here since if we reach this |
67647d0fb
|
871 |
* point, the file counter already went to zero and fget() would fail. |
d47de16c7
|
872 |
* The only hit might come from ep_free() but by holding the mutex |
7699acd13
|
873 |
* will correctly serialize the operation. We do need to acquire |
d47de16c7
|
874 |
* "ep->mtx" after "epmutex" because ep_remove() requires it when called |
7699acd13
|
875 |
* from anywhere but ep_free(). |
684999149
|
876 877 |
* * Besides, ep_remove() acquires the lock, so we can't hold it here. |
b611967de
|
878 |
*/ |
7699acd13
|
879 |
mutex_lock(&epmutex); |
ebe06187b
|
880 |
list_for_each_entry_safe(epi, next, &file->f_ep_links, fllink) { |
7699acd13
|
881 |
ep = epi->ep; |
d8805e633
|
882 |
mutex_lock_nested(&ep->mtx, 0); |
7699acd13
|
883 |
ep_remove(ep, epi); |
d47de16c7
|
884 |
mutex_unlock(&ep->mtx); |
b611967de
|
885 |
} |
7699acd13
|
886 |
mutex_unlock(&epmutex); |
b611967de
|
887 |
} |
53d2be79d
|
888 |
static int ep_alloc(struct eventpoll **pep) |
1da177e4c
|
889 |
{ |
7ef9964e6
|
890 891 892 |
int error; struct user_struct *user; struct eventpoll *ep; |
1da177e4c
|
893 |
|
7ef9964e6
|
894 |
user = get_current_user(); |
7ef9964e6
|
895 896 897 898 |
error = -ENOMEM; ep = kzalloc(sizeof(*ep), GFP_KERNEL); if (unlikely(!ep)) goto free_uid; |
1da177e4c
|
899 |
|
c7ea76302
|
900 |
spin_lock_init(&ep->lock); |
d47de16c7
|
901 |
mutex_init(&ep->mtx); |
1da177e4c
|
902 903 904 905 |
init_waitqueue_head(&ep->wq); init_waitqueue_head(&ep->poll_wait); INIT_LIST_HEAD(&ep->rdllist); ep->rbr = RB_ROOT; |
d47de16c7
|
906 |
ep->ovflist = EP_UNACTIVE_PTR; |
7ef9964e6
|
907 |
ep->user = user; |
1da177e4c
|
908 |
|
53d2be79d
|
909 |
*pep = ep; |
1da177e4c
|
910 |
|
1da177e4c
|
911 |
return 0; |
7ef9964e6
|
912 913 914 915 |
free_uid: free_uid(user); return error; |
1da177e4c
|
916 |
} |
1da177e4c
|
917 |
/* |
c7ea76302
|
918 919 920 |
* Search the file inside the eventpoll tree. The RB tree operations * are protected by the "mtx" mutex, and ep_find() must be called with * "mtx" held. |
1da177e4c
|
921 922 923 924 |
*/ static struct epitem *ep_find(struct eventpoll *ep, struct file *file, int fd) { int kcmp; |
1da177e4c
|
925 926 927 |
struct rb_node *rbp; struct epitem *epi, *epir = NULL; struct epoll_filefd ffd; |
b030a4dd6
|
928 |
ep_set_ffd(&ffd, file, fd); |
1da177e4c
|
929 930 |
for (rbp = ep->rbr.rb_node; rbp; ) { epi = rb_entry(rbp, struct epitem, rbn); |
b030a4dd6
|
931 |
kcmp = ep_cmp_ffd(&ffd, &epi->ffd); |
1da177e4c
|
932 933 934 935 936 |
if (kcmp > 0) rbp = rbp->rb_right; else if (kcmp < 0) rbp = rbp->rb_left; else { |
1da177e4c
|
937 938 939 940 |
epir = epi; break; } } |
1da177e4c
|
941 |
|
1da177e4c
|
942 943 |
return epir; } |
1da177e4c
|
944 |
/* |
7699acd13
|
945 |
* This is the callback that is passed to the wait queue wakeup |
bf6a41db7
|
946 |
* mechanism. It is called by the stored file descriptors when they |
7699acd13
|
947 |
* have events to report. |
1da177e4c
|
948 |
*/ |
7699acd13
|
949 |
static int ep_poll_callback(wait_queue_t *wait, unsigned mode, int sync, void *key) |
1da177e4c
|
950 |
{ |
7699acd13
|
951 952 953 954 |
int pwake = 0; unsigned long flags; struct epitem *epi = ep_item_from_wait(wait); struct eventpoll *ep = epi->ep; |
1da177e4c
|
955 |
|
971316f05
|
956 957 958 959 960 961 962 963 |
if ((unsigned long)key & POLLFREE) { ep_pwq_from_wait(wait)->whead = NULL; /* * whead = NULL above can race with ep_remove_wait_queue() * which can do another remove_wait_queue() after us, so we * can't use __remove_wait_queue(). whead->lock is held by * the caller. */ |
d80e731ec
|
964 |
list_del_init(&wait->task_list); |
971316f05
|
965 |
} |
d80e731ec
|
966 |
|
c7ea76302
|
967 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
968 |
|
7699acd13
|
969 970 971 972 973 974 975 |
/* * If the event mask does not contain any poll(2) event, we consider the * descriptor to be disabled. This condition is likely the effect of the * EPOLLONESHOT bit that disables the descriptor when an event is received, * until the next EPOLL_CTL_MOD will be issued. */ if (!(epi->event.events & ~EP_PRIVATE_BITS)) |
d47de16c7
|
976 977 978 |
goto out_unlock; /* |
2dfa4eeab
|
979 980 981 982 983 984 985 986 987 |
* Check the events coming with the callback. At this stage, not * every device reports the events in the "key" parameter of the * callback. We need to be able to handle both cases here, hence the * test for "key" != NULL before the event match test. */ if (key && !((unsigned long) key & epi->event.events)) goto out_unlock; /* |
bf6a41db7
|
988 |
* If we are transferring events to userspace, we can hold no locks |
d47de16c7
|
989 |
* (because we're accessing user memory, and because of linux f_op->poll() |
bf6a41db7
|
990 |
* semantics). All the events that happen during that period of time are |
d47de16c7
|
991 992 993 994 995 996 |
* chained in ep->ovflist and requeued later on. */ if (unlikely(ep->ovflist != EP_UNACTIVE_PTR)) { if (epi->next == EP_UNACTIVE_PTR) { epi->next = ep->ovflist; ep->ovflist = epi; |
4d7e30d98
|
997 998 999 1000 1001 1002 1003 |
if (epi->ws) { /* * Activate ep->ws since epi->ws may get * deactivated at any time. */ __pm_stay_awake(ep->ws); } |
d47de16c7
|
1004 1005 1006 |
} goto out_unlock; } |
1da177e4c
|
1007 |
|
7699acd13
|
1008 |
/* If this file is already in the ready list we exit soon */ |
4d7e30d98
|
1009 |
if (!ep_is_linked(&epi->rdllink)) { |
5071f97ec
|
1010 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1011 |
ep_pm_stay_awake_rcu(epi); |
4d7e30d98
|
1012 |
} |
7699acd13
|
1013 |
|
7699acd13
|
1014 1015 1016 1017 1018 |
/* * Wake up ( if active ) both the eventpoll wait list and the ->poll() * wait list. */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce
|
1019 |
wake_up_locked(&ep->wq); |
7699acd13
|
1020 1021 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
d47de16c7
|
1022 |
out_unlock: |
c7ea76302
|
1023 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1024 |
|
7699acd13
|
1025 1026 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1027 |
ep_poll_safewake(&ep->poll_wait); |
7699acd13
|
1028 1029 1030 |
return 1; } |
1da177e4c
|
1031 1032 1033 1034 1035 1036 1037 1038 |
/* * This is the callback that is used to add our wait queue to the * target file wakeup lists. */ static void ep_ptable_queue_proc(struct file *file, wait_queue_head_t *whead, poll_table *pt) { |
b030a4dd6
|
1039 |
struct epitem *epi = ep_item_from_epqueue(pt); |
1da177e4c
|
1040 |
struct eppoll_entry *pwq; |
e94b17660
|
1041 |
if (epi->nwait >= 0 && (pwq = kmem_cache_alloc(pwq_cache, GFP_KERNEL))) { |
1da177e4c
|
1042 1043 1044 1045 1046 1047 |
init_waitqueue_func_entry(&pwq->wait, ep_poll_callback); pwq->whead = whead; pwq->base = epi; add_wait_queue(whead, &pwq->wait); list_add_tail(&pwq->llink, &epi->pwqlist); epi->nwait++; |
296e236e9
|
1048 |
} else { |
1da177e4c
|
1049 1050 |
/* We have to signal that an error occurred */ epi->nwait = -1; |
296e236e9
|
1051 |
} |
1da177e4c
|
1052 |
} |
1da177e4c
|
1053 1054 1055 1056 1057 1058 1059 1060 1061 |
static void ep_rbtree_insert(struct eventpoll *ep, struct epitem *epi) { int kcmp; struct rb_node **p = &ep->rbr.rb_node, *parent = NULL; struct epitem *epic; while (*p) { parent = *p; epic = rb_entry(parent, struct epitem, rbn); |
b030a4dd6
|
1062 |
kcmp = ep_cmp_ffd(&epi->ffd, &epic->ffd); |
1da177e4c
|
1063 1064 1065 1066 1067 1068 1069 1070 |
if (kcmp > 0) p = &parent->rb_right; else p = &parent->rb_left; } rb_link_node(&epi->rbn, parent, p); rb_insert_color(&epi->rbn, &ep->rbr); } |
a80a6b85b
|
1071 |
|
28d82dc1c
|
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 |
#define PATH_ARR_SIZE 5 /* * These are the number paths of length 1 to 5, that we are allowing to emanate * from a single file of interest. For example, we allow 1000 paths of length * 1, to emanate from each file of interest. This essentially represents the * potential wakeup paths, which need to be limited in order to avoid massive * uncontrolled wakeup storms. The common use case should be a single ep which * is connected to n file sources. In this case each file source has 1 path * of length 1. Thus, the numbers below should be more than sufficient. These * path limits are enforced during an EPOLL_CTL_ADD operation, since a modify * and delete can't add additional paths. Protected by the epmutex. */ static const int path_limits[PATH_ARR_SIZE] = { 1000, 500, 100, 50, 10 }; static int path_count[PATH_ARR_SIZE]; static int path_count_inc(int nests) { |
93dc6107a
|
1089 1090 1091 |
/* Allow an arbitrary number of depth 1 paths */ if (nests == 0) return 0; |
28d82dc1c
|
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 |
if (++path_count[nests] > path_limits[nests]) return -1; return 0; } static void path_count_init(void) { int i; for (i = 0; i < PATH_ARR_SIZE; i++) path_count[i] = 0; } static int reverse_path_check_proc(void *priv, void *cookie, int call_nests) { int error = 0; struct file *file = priv; struct file *child_file; struct epitem *epi; |
ae10b2b4e
|
1111 1112 1113 |
/* CTL_DEL can remove links here, but that can't increase our count */ rcu_read_lock(); list_for_each_entry_rcu(epi, &file->f_ep_links, fllink) { |
28d82dc1c
|
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 |
child_file = epi->ep->file; if (is_file_epoll(child_file)) { if (list_empty(&child_file->f_ep_links)) { if (path_count_inc(call_nests)) { error = -1; break; } } else { error = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, reverse_path_check_proc, child_file, child_file, current); } if (error != 0) break; } else { printk(KERN_ERR "reverse_path_check_proc: " "file is not an ep! "); } } |
ae10b2b4e
|
1136 |
rcu_read_unlock(); |
28d82dc1c
|
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 |
return error; } /** * reverse_path_check - The tfile_check_list is list of file *, which have * links that are proposed to be newly added. We need to * make sure that those added links don't add too many * paths such that we will spend all our time waking up * eventpoll objects. * * Returns: Returns zero if the proposed links don't create too many paths, * -1 otherwise. */ static int reverse_path_check(void) { |
28d82dc1c
|
1152 1153 1154 1155 1156 |
int error = 0; struct file *current_file; /* let's call this for all tfiles */ list_for_each_entry(current_file, &tfile_check_list, f_tfile_llink) { |
28d82dc1c
|
1157 1158 1159 1160 1161 1162 1163 1164 1165 |
path_count_init(); error = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, reverse_path_check_proc, current_file, current_file, current); if (error) break; } return error; } |
4d7e30d98
|
1166 1167 1168 |
static int ep_create_wakeup_source(struct epitem *epi) { const char *name; |
eea1d5859
|
1169 |
struct wakeup_source *ws; |
4d7e30d98
|
1170 1171 1172 1173 1174 1175 1176 1177 |
if (!epi->ep->ws) { epi->ep->ws = wakeup_source_register("eventpoll"); if (!epi->ep->ws) return -ENOMEM; } name = epi->ffd.file->f_path.dentry->d_name.name; |
eea1d5859
|
1178 1179 1180 |
ws = wakeup_source_register(name); if (!ws) |
4d7e30d98
|
1181 |
return -ENOMEM; |
eea1d5859
|
1182 |
rcu_assign_pointer(epi->ws, ws); |
4d7e30d98
|
1183 1184 1185 |
return 0; } |
eea1d5859
|
1186 1187 |
/* rare code path, only used when EPOLL_CTL_MOD removes a wakeup source */ static noinline void ep_destroy_wakeup_source(struct epitem *epi) |
4d7e30d98
|
1188 |
{ |
eea1d5859
|
1189 |
struct wakeup_source *ws = ep_wakeup_source(epi); |
d6d67e723
|
1190 |
RCU_INIT_POINTER(epi->ws, NULL); |
eea1d5859
|
1191 1192 1193 1194 1195 1196 1197 1198 |
/* * wait for ep_pm_stay_awake_rcu to finish, synchronize_rcu is * used internally by wakeup_source_remove, too (called by * wakeup_source_unregister), so we cannot use call_rcu */ synchronize_rcu(); wakeup_source_unregister(ws); |
4d7e30d98
|
1199 |
} |
c7ea76302
|
1200 1201 1202 |
/* * Must be called with "mtx" held. */ |
1da177e4c
|
1203 |
static int ep_insert(struct eventpoll *ep, struct epoll_event *event, |
67347fe4e
|
1204 |
struct file *tfile, int fd, int full_check) |
1da177e4c
|
1205 1206 1207 |
{ int error, revents, pwake = 0; unsigned long flags; |
52bd19f76
|
1208 |
long user_watches; |
1da177e4c
|
1209 1210 |
struct epitem *epi; struct ep_pqueue epq; |
52bd19f76
|
1211 1212 |
user_watches = atomic_long_read(&ep->user->epoll_watches); if (unlikely(user_watches >= max_user_watches)) |
7ef9964e6
|
1213 |
return -ENOSPC; |
e94b17660
|
1214 |
if (!(epi = kmem_cache_alloc(epi_cache, GFP_KERNEL))) |
7ef9964e6
|
1215 |
return -ENOMEM; |
1da177e4c
|
1216 1217 |
/* Item initialization follow here ... */ |
1da177e4c
|
1218 1219 |
INIT_LIST_HEAD(&epi->rdllink); INIT_LIST_HEAD(&epi->fllink); |
1da177e4c
|
1220 1221 |
INIT_LIST_HEAD(&epi->pwqlist); epi->ep = ep; |
b030a4dd6
|
1222 |
ep_set_ffd(&epi->ffd, tfile, fd); |
1da177e4c
|
1223 |
epi->event = *event; |
1da177e4c
|
1224 |
epi->nwait = 0; |
d47de16c7
|
1225 |
epi->next = EP_UNACTIVE_PTR; |
4d7e30d98
|
1226 1227 1228 1229 1230 |
if (epi->event.events & EPOLLWAKEUP) { error = ep_create_wakeup_source(epi); if (error) goto error_create_wakeup_source; } else { |
eea1d5859
|
1231 |
RCU_INIT_POINTER(epi->ws, NULL); |
4d7e30d98
|
1232 |
} |
1da177e4c
|
1233 1234 1235 1236 1237 1238 1239 1240 |
/* Initialize the poll table using the queue callback */ epq.epi = epi; init_poll_funcptr(&epq.pt, ep_ptable_queue_proc); /* * Attach the item to the poll hooks and get current event bits. * We can safely use the file* here because its usage count has |
c7ea76302
|
1241 1242 1243 |
* been increased by the caller of this function. Note that after * this operation completes, the poll callback can start hitting * the new item. |
1da177e4c
|
1244 |
*/ |
450d89ec0
|
1245 |
revents = ep_item_poll(epi, &epq.pt); |
1da177e4c
|
1246 1247 1248 1249 1250 1251 |
/* * We have to check if something went wrong during the poll wait queue * install process. Namely an allocation for a wait queue failed due * high memory pressure. */ |
7ef9964e6
|
1252 |
error = -ENOMEM; |
1da177e4c
|
1253 |
if (epi->nwait < 0) |
7699acd13
|
1254 |
goto error_unregister; |
1da177e4c
|
1255 1256 |
/* Add the current item to the list of active epoll hook for this file */ |
684999149
|
1257 |
spin_lock(&tfile->f_lock); |
ae10b2b4e
|
1258 |
list_add_tail_rcu(&epi->fllink, &tfile->f_ep_links); |
684999149
|
1259 |
spin_unlock(&tfile->f_lock); |
1da177e4c
|
1260 |
|
c7ea76302
|
1261 1262 1263 1264 |
/* * Add the current item to the RB tree. All RB tree operations are * protected by "mtx", and ep_insert() is called with "mtx" held. */ |
1da177e4c
|
1265 |
ep_rbtree_insert(ep, epi); |
28d82dc1c
|
1266 1267 |
/* now check if we've created too many backpaths */ error = -EINVAL; |
67347fe4e
|
1268 |
if (full_check && reverse_path_check()) |
28d82dc1c
|
1269 |
goto error_remove_epi; |
c7ea76302
|
1270 1271 |
/* We have to drop the new item inside our item list to keep track of it */ spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1272 |
/* If the file is already "ready" we drop it inside the ready list */ |
b030a4dd6
|
1273 |
if ((revents & event->events) && !ep_is_linked(&epi->rdllink)) { |
1da177e4c
|
1274 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1275 |
ep_pm_stay_awake(epi); |
1da177e4c
|
1276 1277 1278 |
/* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce
|
1279 |
wake_up_locked(&ep->wq); |
1da177e4c
|
1280 1281 1282 |
if (waitqueue_active(&ep->poll_wait)) pwake++; } |
c7ea76302
|
1283 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1284 |
|
52bd19f76
|
1285 |
atomic_long_inc(&ep->user->epoll_watches); |
7ef9964e6
|
1286 |
|
1da177e4c
|
1287 1288 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1289 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
1290 |
|
1da177e4c
|
1291 |
return 0; |
28d82dc1c
|
1292 1293 |
error_remove_epi: spin_lock(&tfile->f_lock); |
ae10b2b4e
|
1294 |
list_del_rcu(&epi->fllink); |
28d82dc1c
|
1295 1296 1297 |
spin_unlock(&tfile->f_lock); rb_erase(&epi->rbn, &ep->rbr); |
7699acd13
|
1298 |
error_unregister: |
1da177e4c
|
1299 1300 1301 1302 |
ep_unregister_pollwait(ep, epi); /* * We need to do this because an event could have been arrived on some |
67647d0fb
|
1303 1304 1305 |
* allocated wait queue. Note that we don't care about the ep->ovflist * list, since that is used/cleaned only inside a section bound by "mtx". * And ep_insert() is called with "mtx" held. |
1da177e4c
|
1306 |
*/ |
c7ea76302
|
1307 |
spin_lock_irqsave(&ep->lock, flags); |
b030a4dd6
|
1308 |
if (ep_is_linked(&epi->rdllink)) |
6192bd536
|
1309 |
list_del_init(&epi->rdllink); |
c7ea76302
|
1310 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1311 |
|
eea1d5859
|
1312 |
wakeup_source_unregister(ep_wakeup_source(epi)); |
4d7e30d98
|
1313 1314 |
error_create_wakeup_source: |
b030a4dd6
|
1315 |
kmem_cache_free(epi_cache, epi); |
7ef9964e6
|
1316 |
|
1da177e4c
|
1317 1318 |
return error; } |
1da177e4c
|
1319 1320 |
/* * Modify the interest event mask by dropping an event if the new mask |
c7ea76302
|
1321 |
* has a match in the current file status. Must be called with "mtx" held. |
1da177e4c
|
1322 1323 1324 1325 1326 |
*/ static int ep_modify(struct eventpoll *ep, struct epitem *epi, struct epoll_event *event) { int pwake = 0; unsigned int revents; |
626cf2366
|
1327 1328 1329 |
poll_table pt; init_poll_funcptr(&pt, NULL); |
1da177e4c
|
1330 1331 |
/* |
e057e15ff
|
1332 1333 1334 |
* Set the new event interest mask before calling f_op->poll(); * otherwise we might miss an event that happens between the * f_op->poll() call and the new event set registering. |
1da177e4c
|
1335 |
*/ |
128dd1759
|
1336 |
epi->event.events = event->events; /* need barrier below */ |
e057e15ff
|
1337 |
epi->event.data = event->data; /* protected by mtx */ |
4d7e30d98
|
1338 |
if (epi->event.events & EPOLLWAKEUP) { |
eea1d5859
|
1339 |
if (!ep_has_wakeup_source(epi)) |
4d7e30d98
|
1340 |
ep_create_wakeup_source(epi); |
eea1d5859
|
1341 |
} else if (ep_has_wakeup_source(epi)) { |
4d7e30d98
|
1342 1343 |
ep_destroy_wakeup_source(epi); } |
1da177e4c
|
1344 1345 |
/* |
128dd1759
|
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 |
* The following barrier has two effects: * * 1) Flush epi changes above to other CPUs. This ensures * we do not miss events from ep_poll_callback if an * event occurs immediately after we call f_op->poll(). * We need this because we did not take ep->lock while * changing epi above (but ep_poll_callback does take * ep->lock). * * 2) We also need to ensure we do not miss _past_ events * when calling f_op->poll(). This barrier also * pairs with the barrier in wq_has_sleeper (see * comments for wq_has_sleeper). * * This barrier will now guarantee ep_poll_callback or f_op->poll * (or both) will notice the readiness of an item. */ smp_mb(); /* |
1da177e4c
|
1366 1367 1368 |
* Get current event bits. We can safely use the file* here because * its usage count has been increased by the caller of this function. */ |
450d89ec0
|
1369 |
revents = ep_item_poll(epi, &pt); |
1da177e4c
|
1370 |
|
1da177e4c
|
1371 |
/* |
c7ea76302
|
1372 |
* If the item is "hot" and it is not registered inside the ready |
67647d0fb
|
1373 |
* list, push it inside. |
1da177e4c
|
1374 |
*/ |
c7ea76302
|
1375 |
if (revents & event->events) { |
e057e15ff
|
1376 |
spin_lock_irq(&ep->lock); |
c7ea76302
|
1377 1378 |
if (!ep_is_linked(&epi->rdllink)) { list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1379 |
ep_pm_stay_awake(epi); |
c7ea76302
|
1380 1381 1382 |
/* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce
|
1383 |
wake_up_locked(&ep->wq); |
c7ea76302
|
1384 1385 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
7699acd13
|
1386 |
} |
e057e15ff
|
1387 |
spin_unlock_irq(&ep->lock); |
7699acd13
|
1388 |
} |
1da177e4c
|
1389 |
|
7699acd13
|
1390 1391 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1392 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
1393 |
|
7699acd13
|
1394 |
return 0; |
1da177e4c
|
1395 |
} |
296e236e9
|
1396 1397 |
static int ep_send_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
1da177e4c
|
1398 |
{ |
5071f97ec
|
1399 1400 |
struct ep_send_events_data *esed = priv; int eventcnt; |
296e236e9
|
1401 |
unsigned int revents; |
5071f97ec
|
1402 1403 |
struct epitem *epi; struct epoll_event __user *uevent; |
eea1d5859
|
1404 |
struct wakeup_source *ws; |
626cf2366
|
1405 1406 1407 |
poll_table pt; init_poll_funcptr(&pt, NULL); |
1da177e4c
|
1408 |
|
296e236e9
|
1409 |
/* |
5071f97ec
|
1410 1411 1412 |
* We can loop without lock because we are passed a task private list. * Items cannot vanish during the loop because ep_scan_ready_list() is * holding "mtx" during this call. |
296e236e9
|
1413 |
*/ |
5071f97ec
|
1414 1415 1416 |
for (eventcnt = 0, uevent = esed->events; !list_empty(head) && eventcnt < esed->maxevents;) { epi = list_first_entry(head, struct epitem, rdllink); |
d47de16c7
|
1417 |
|
4d7e30d98
|
1418 1419 1420 1421 1422 1423 1424 1425 1426 |
/* * Activate ep->ws before deactivating epi->ws to prevent * triggering auto-suspend here (in case we reactive epi->ws * below). * * This could be rearranged to delay the deactivation of epi->ws * instead, but then epi->ws would temporarily be out of sync * with ep_is_linked(). */ |
eea1d5859
|
1427 1428 1429 1430 1431 1432 |
ws = ep_wakeup_source(epi); if (ws) { if (ws->active) __pm_stay_awake(ep->ws); __pm_relax(ws); } |
d47de16c7
|
1433 |
list_del_init(&epi->rdllink); |
1da177e4c
|
1434 |
|
450d89ec0
|
1435 |
revents = ep_item_poll(epi, &pt); |
5071f97ec
|
1436 |
|
296e236e9
|
1437 |
/* |
5071f97ec
|
1438 1439 1440 1441 |
* If the event mask intersect the caller-requested one, * deliver the event to userspace. Again, ep_scan_ready_list() * is holding "mtx", so no operations coming from userspace * can change the item. |
296e236e9
|
1442 1443 |
*/ if (revents) { |
5071f97ec
|
1444 |
if (__put_user(revents, &uevent->events) || |
d03058828
|
1445 1446 |
__put_user(epi->event.data, &uevent->data)) { list_add(&epi->rdllink, head); |
eea1d5859
|
1447 |
ep_pm_stay_awake(epi); |
296e236e9
|
1448 |
return eventcnt ? eventcnt : -EFAULT; |
d03058828
|
1449 |
} |
296e236e9
|
1450 |
eventcnt++; |
5071f97ec
|
1451 |
uevent++; |
296e236e9
|
1452 1453 1454 1455 1456 1457 1458 1459 |
if (epi->event.events & EPOLLONESHOT) epi->event.events &= EP_PRIVATE_BITS; else if (!(epi->event.events & EPOLLET)) { /* * If this file has been added with Level * Trigger mode, we need to insert back inside * the ready list, so that the next call to * epoll_wait() will check again the events |
25985edce
|
1460 |
* availability. At this point, no one can insert |
296e236e9
|
1461 1462 1463 1464 1465 1466 |
* into ep->rdllist besides us. The epoll_ctl() * callers are locked out by * ep_scan_ready_list() holding "mtx" and the * poll callback will queue them in ep->ovflist. */ list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1467 |
ep_pm_stay_awake(epi); |
296e236e9
|
1468 1469 1470 |
} } } |
5071f97ec
|
1471 1472 1473 |
return eventcnt; } |
d47de16c7
|
1474 |
|
296e236e9
|
1475 1476 |
static int ep_send_events(struct eventpoll *ep, struct epoll_event __user *events, int maxevents) |
5071f97ec
|
1477 1478 |
{ struct ep_send_events_data esed; |
1da177e4c
|
1479 |
|
5071f97ec
|
1480 1481 |
esed.maxevents = maxevents; esed.events = events; |
6192bd536
|
1482 |
|
67347fe4e
|
1483 |
return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false); |
1da177e4c
|
1484 |
} |
0781b909b
|
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 |
static inline struct timespec ep_set_mstimeout(long ms) { struct timespec now, ts = { .tv_sec = ms / MSEC_PER_SEC, .tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC), }; ktime_get_ts(&now); return timespec_add_safe(now, ts); } |
f4d93ad74
|
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 |
/** * ep_poll - Retrieves ready events, and delivers them to the caller supplied * event buffer. * * @ep: Pointer to the eventpoll context. * @events: Pointer to the userspace buffer where the ready events should be * stored. * @maxevents: Size (in terms of number of events) of the caller event buffer. * @timeout: Maximum timeout for the ready events fetch operation, in * milliseconds. If the @timeout is zero, the function will not block, * while if the @timeout is less than zero, the function will block * until at least one event has been retrieved (or an error * occurred). * * Returns: Returns the number of ready events which have been fetched, or an * error code, in case of error. */ |
1da177e4c
|
1512 1513 1514 |
static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events, int maxevents, long timeout) { |
f4d93ad74
|
1515 |
int res = 0, eavail, timed_out = 0; |
1da177e4c
|
1516 |
unsigned long flags; |
f4d93ad74
|
1517 |
long slack = 0; |
1da177e4c
|
1518 |
wait_queue_t wait; |
95aac7b1c
|
1519 1520 1521 |
ktime_t expires, *to = NULL; if (timeout > 0) { |
0781b909b
|
1522 |
struct timespec end_time = ep_set_mstimeout(timeout); |
95aac7b1c
|
1523 1524 1525 1526 |
slack = select_estimate_accuracy(&end_time); to = &expires; *to = timespec_to_ktime(end_time); } else if (timeout == 0) { |
f4d93ad74
|
1527 1528 1529 1530 |
/* * Avoid the unnecessary trip to the wait queue loop, if the * caller specified a non blocking operation. */ |
95aac7b1c
|
1531 |
timed_out = 1; |
f4d93ad74
|
1532 1533 |
spin_lock_irqsave(&ep->lock, flags); goto check_events; |
95aac7b1c
|
1534 |
} |
1da177e4c
|
1535 |
|
f4d93ad74
|
1536 |
fetch_events: |
c7ea76302
|
1537 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1538 |
|
3fb0e584a
|
1539 |
if (!ep_events_available(ep)) { |
1da177e4c
|
1540 1541 1542 1543 1544 1545 |
/* * We don't have any available event to return to the caller. * We need to sleep here, and we will be wake up by * ep_poll_callback() when events will become available. */ init_waitqueue_entry(&wait, current); |
a93d2f174
|
1546 |
__add_wait_queue_exclusive(&ep->wq, &wait); |
1da177e4c
|
1547 1548 1549 1550 1551 1552 1553 1554 |
for (;;) { /* * We don't want to sleep if the ep_poll_callback() sends us * a wakeup in between. That's why we set the task state * to TASK_INTERRUPTIBLE before doing the checks. */ set_current_state(TASK_INTERRUPTIBLE); |
3fb0e584a
|
1555 |
if (ep_events_available(ep) || timed_out) |
1da177e4c
|
1556 1557 1558 1559 1560 |
break; if (signal_pending(current)) { res = -EINTR; break; } |
c7ea76302
|
1561 |
spin_unlock_irqrestore(&ep->lock, flags); |
c511851de
|
1562 |
if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS)) |
95aac7b1c
|
1563 |
timed_out = 1; |
c7ea76302
|
1564 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1565 |
} |
3419b23a9
|
1566 |
__remove_wait_queue(&ep->wq, &wait); |
1da177e4c
|
1567 1568 1569 |
set_current_state(TASK_RUNNING); } |
f4d93ad74
|
1570 |
check_events: |
1da177e4c
|
1571 |
/* Is it worth to try to dig for events ? */ |
3fb0e584a
|
1572 |
eavail = ep_events_available(ep); |
1da177e4c
|
1573 |
|
c7ea76302
|
1574 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1575 1576 1577 1578 1579 1580 1581 |
/* * Try to transfer events to user space. In case we get 0 events and * there's still timeout left over, we go trying again in search of * more luck. */ if (!res && eavail && |
95aac7b1c
|
1582 |
!(res = ep_send_events(ep, events, maxevents)) && !timed_out) |
f4d93ad74
|
1583 |
goto fetch_events; |
1da177e4c
|
1584 1585 1586 |
return res; } |
22bacca48
|
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 |
/** * ep_loop_check_proc - Callback function to be passed to the @ep_call_nested() * API, to verify that adding an epoll file inside another * epoll structure, does not violate the constraints, in * terms of closed loops, or too deep chains (which can * result in excessive stack usage). * * @priv: Pointer to the epoll file to be currently checked. * @cookie: Original cookie for this call. This is the top-of-the-chain epoll * data structure pointer. * @call_nests: Current dept of the @ep_call_nested() call stack. * * Returns: Returns zero if adding the epoll @file inside current epoll * structure @ep does not violate the constraints, or -1 otherwise. */ static int ep_loop_check_proc(void *priv, void *cookie, int call_nests) { int error = 0; struct file *file = priv; struct eventpoll *ep = file->private_data; |
28d82dc1c
|
1607 |
struct eventpoll *ep_tovisit; |
22bacca48
|
1608 1609 |
struct rb_node *rbp; struct epitem *epi; |
d8805e633
|
1610 |
mutex_lock_nested(&ep->mtx, call_nests + 1); |
28d82dc1c
|
1611 1612 |
ep->visited = 1; list_add(&ep->visited_list_link, &visited_list); |
22bacca48
|
1613 1614 1615 |
for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) { epi = rb_entry(rbp, struct epitem, rbn); if (unlikely(is_file_epoll(epi->ffd.file))) { |
28d82dc1c
|
1616 1617 1618 |
ep_tovisit = epi->ffd.file->private_data; if (ep_tovisit->visited) continue; |
22bacca48
|
1619 |
error = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
28d82dc1c
|
1620 1621 |
ep_loop_check_proc, epi->ffd.file, ep_tovisit, current); |
22bacca48
|
1622 1623 |
if (error != 0) break; |
28d82dc1c
|
1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 |
} else { /* * If we've reached a file that is not associated with * an ep, then we need to check if the newly added * links are going to add too many wakeup paths. We do * this by adding it to the tfile_check_list, if it's * not already there, and calling reverse_path_check() * during ep_insert(). */ if (list_empty(&epi->ffd.file->f_tfile_llink)) list_add(&epi->ffd.file->f_tfile_llink, &tfile_check_list); |
22bacca48
|
1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 |
} } mutex_unlock(&ep->mtx); return error; } /** * ep_loop_check - Performs a check to verify that adding an epoll file (@file) * another epoll file (represented by @ep) does not create * closed loops or too deep chains. * * @ep: Pointer to the epoll private data structure. * @file: Pointer to the epoll file to be checked. * * Returns: Returns zero if adding the epoll @file inside current epoll * structure @ep does not violate the constraints, or -1 otherwise. */ static int ep_loop_check(struct eventpoll *ep, struct file *file) { |
28d82dc1c
|
1656 1657 1658 1659 |
int ret; struct eventpoll *ep_cur, *ep_next; ret = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
22bacca48
|
1660 |
ep_loop_check_proc, file, ep, current); |
28d82dc1c
|
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 |
/* clear visited list */ list_for_each_entry_safe(ep_cur, ep_next, &visited_list, visited_list_link) { ep_cur->visited = 0; list_del(&ep_cur->visited_list_link); } return ret; } static void clear_tfile_check_list(void) { struct file *file; /* first clear the tfile_check_list */ while (!list_empty(&tfile_check_list)) { file = list_first_entry(&tfile_check_list, struct file, f_tfile_llink); list_del_init(&file->f_tfile_llink); } INIT_LIST_HEAD(&tfile_check_list); |
22bacca48
|
1681 |
} |
7699acd13
|
1682 |
/* |
523723bb5
|
1683 |
* Open an eventpoll file descriptor. |
7699acd13
|
1684 |
*/ |
5a8a82b1d
|
1685 |
SYSCALL_DEFINE1(epoll_create1, int, flags) |
7699acd13
|
1686 |
{ |
28d82dc1c
|
1687 |
int error, fd; |
bb57c3edc
|
1688 |
struct eventpoll *ep = NULL; |
28d82dc1c
|
1689 |
struct file *file; |
7699acd13
|
1690 |
|
e38b36f32
|
1691 1692 |
/* Check the EPOLL_* constant for consistency. */ BUILD_BUG_ON(EPOLL_CLOEXEC != O_CLOEXEC); |
296e236e9
|
1693 1694 |
if (flags & ~EPOLL_CLOEXEC) return -EINVAL; |
7699acd13
|
1695 |
/* |
bb57c3edc
|
1696 |
* Create the internal data structure ("struct eventpoll"). |
7699acd13
|
1697 |
*/ |
9fe5ad9c8
|
1698 |
error = ep_alloc(&ep); |
bb57c3edc
|
1699 1700 |
if (error < 0) return error; |
7699acd13
|
1701 1702 |
/* * Creates all the items needed to setup an eventpoll file. That is, |
2030a42ce
|
1703 |
* a file structure and a free file descriptor. |
7699acd13
|
1704 |
*/ |
28d82dc1c
|
1705 1706 1707 1708 1709 1710 |
fd = get_unused_fd_flags(O_RDWR | (flags & O_CLOEXEC)); if (fd < 0) { error = fd; goto out_free_ep; } file = anon_inode_getfile("[eventpoll]", &eventpoll_fops, ep, |
628ff7c1d
|
1711 |
O_RDWR | (flags & O_CLOEXEC)); |
28d82dc1c
|
1712 1713 1714 1715 |
if (IS_ERR(file)) { error = PTR_ERR(file); goto out_free_fd; } |
28d82dc1c
|
1716 |
ep->file = file; |
98022748f
|
1717 |
fd_install(fd, file); |
28d82dc1c
|
1718 1719 1720 1721 1722 1723 |
return fd; out_free_fd: put_unused_fd(fd); out_free_ep: ep_free(ep); |
bb57c3edc
|
1724 |
return error; |
7699acd13
|
1725 |
} |
5a8a82b1d
|
1726 |
SYSCALL_DEFINE1(epoll_create, int, size) |
a0998b50c
|
1727 |
{ |
bfe3891a5
|
1728 |
if (size <= 0) |
9fe5ad9c8
|
1729 1730 1731 |
return -EINVAL; return sys_epoll_create1(0); |
a0998b50c
|
1732 |
} |
7699acd13
|
1733 1734 1735 |
/* * The following function implements the controller interface for * the eventpoll file that enables the insertion/removal/change of |
67647d0fb
|
1736 |
* file descriptors inside the interest set. |
7699acd13
|
1737 |
*/ |
5a8a82b1d
|
1738 1739 |
SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd, struct epoll_event __user *, event) |
7699acd13
|
1740 1741 |
{ int error; |
67347fe4e
|
1742 |
int full_check = 0; |
7e3fb5842
|
1743 |
struct fd f, tf; |
7699acd13
|
1744 1745 1746 |
struct eventpoll *ep; struct epitem *epi; struct epoll_event epds; |
67347fe4e
|
1747 |
struct eventpoll *tep = NULL; |
7699acd13
|
1748 |
|
7699acd13
|
1749 1750 1751 1752 |
error = -EFAULT; if (ep_op_has_event(op) && copy_from_user(&epds, event, sizeof(struct epoll_event))) goto error_return; |
7699acd13
|
1753 |
error = -EBADF; |
7e3fb5842
|
1754 1755 |
f = fdget(epfd); if (!f.file) |
7699acd13
|
1756 1757 1758 |
goto error_return; /* Get the "struct file *" for the target file */ |
7e3fb5842
|
1759 1760 |
tf = fdget(fd); if (!tf.file) |
7699acd13
|
1761 1762 1763 1764 |
goto error_fput; /* The target file descriptor must support poll */ error = -EPERM; |
72c2d5319
|
1765 |
if (!tf.file->f_op->poll) |
7699acd13
|
1766 |
goto error_tgt_fput; |
4d7e30d98
|
1767 |
/* Check if EPOLLWAKEUP is allowed */ |
c680e41b3
|
1768 1769 |
if (ep_op_has_event(op)) ep_take_care_of_epollwakeup(&epds); |
4d7e30d98
|
1770 |
|
7699acd13
|
1771 1772 1773 1774 1775 1776 |
/* * We have to check that the file structure underneath the file descriptor * the user passed to us _is_ an eventpoll file. And also we do not permit * adding an epoll file descriptor inside itself. */ error = -EINVAL; |
7e3fb5842
|
1777 |
if (f.file == tf.file || !is_file_epoll(f.file)) |
7699acd13
|
1778 1779 1780 1781 1782 1783 |
goto error_tgt_fput; /* * At this point it is safe to assume that the "private_data" contains * our own data structure. */ |
7e3fb5842
|
1784 |
ep = f.file->private_data; |
7699acd13
|
1785 |
|
22bacca48
|
1786 1787 1788 |
/* * When we insert an epoll file descriptor, inside another epoll file * descriptor, there is the change of creating closed loops, which are |
28d82dc1c
|
1789 1790 1791 1792 |
* better be handled here, than in more critical paths. While we are * checking for loops we also determine the list of files reachable * and hang them on the tfile_check_list, so we can check that we * haven't created too many possible wakeup paths. |
22bacca48
|
1793 |
* |
67347fe4e
|
1794 1795 1796 1797 1798 1799 |
* We do not need to take the global 'epumutex' on EPOLL_CTL_ADD when * the epoll file descriptor is attaching directly to a wakeup source, * unless the epoll file descriptor is nested. The purpose of taking the * 'epmutex' on add is to prevent complex toplogies such as loops and * deep wakeup paths from forming in parallel through multiple * EPOLL_CTL_ADD operations. |
22bacca48
|
1800 |
*/ |
67347fe4e
|
1801 |
mutex_lock_nested(&ep->mtx, 0); |
28d82dc1c
|
1802 |
if (op == EPOLL_CTL_ADD) { |
67347fe4e
|
1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 |
if (!list_empty(&f.file->f_ep_links) || is_file_epoll(tf.file)) { full_check = 1; mutex_unlock(&ep->mtx); mutex_lock(&epmutex); if (is_file_epoll(tf.file)) { error = -ELOOP; if (ep_loop_check(ep, tf.file) != 0) { clear_tfile_check_list(); goto error_tgt_fput; } } else list_add(&tf.file->f_tfile_llink, &tfile_check_list); mutex_lock_nested(&ep->mtx, 0); if (is_file_epoll(tf.file)) { tep = tf.file->private_data; mutex_lock_nested(&tep->mtx, 1); |
13d518074
|
1821 |
} |
67347fe4e
|
1822 1823 |
} } |
7699acd13
|
1824 |
|
67647d0fb
|
1825 1826 1827 1828 1829 |
/* * Try to lookup the file inside our RB tree, Since we grabbed "mtx" * above, we can be sure to be able to use the item looked up by * ep_find() till we release the mutex. */ |
7e3fb5842
|
1830 |
epi = ep_find(ep, tf.file, fd); |
7699acd13
|
1831 1832 1833 1834 1835 1836 |
error = -EINVAL; switch (op) { case EPOLL_CTL_ADD: if (!epi) { epds.events |= POLLERR | POLLHUP; |
67347fe4e
|
1837 |
error = ep_insert(ep, &epds, tf.file, fd, full_check); |
7699acd13
|
1838 1839 |
} else error = -EEXIST; |
67347fe4e
|
1840 1841 |
if (full_check) clear_tfile_check_list(); |
7699acd13
|
1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 |
break; case EPOLL_CTL_DEL: if (epi) error = ep_remove(ep, epi); else error = -ENOENT; break; case EPOLL_CTL_MOD: if (epi) { epds.events |= POLLERR | POLLHUP; error = ep_modify(ep, epi, &epds); } else error = -ENOENT; break; } |
67347fe4e
|
1857 1858 |
if (tep != NULL) mutex_unlock(&tep->mtx); |
d47de16c7
|
1859 |
mutex_unlock(&ep->mtx); |
7699acd13
|
1860 1861 |
error_tgt_fput: |
67347fe4e
|
1862 |
if (full_check) |
22bacca48
|
1863 |
mutex_unlock(&epmutex); |
7e3fb5842
|
1864 |
fdput(tf); |
7699acd13
|
1865 |
error_fput: |
7e3fb5842
|
1866 |
fdput(f); |
7699acd13
|
1867 |
error_return: |
7699acd13
|
1868 1869 1870 1871 1872 1873 1874 1875 |
return error; } /* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_wait(2). */ |
5a8a82b1d
|
1876 1877 |
SYSCALL_DEFINE4(epoll_wait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout) |
7699acd13
|
1878 |
{ |
2903ff019
|
1879 1880 |
int error; struct fd f; |
7699acd13
|
1881 |
struct eventpoll *ep; |
7699acd13
|
1882 1883 1884 1885 1886 |
/* The maximum number of event must be greater than zero */ if (maxevents <= 0 || maxevents > EP_MAX_EVENTS) return -EINVAL; /* Verify that the area passed by the user is writeable */ |
2903ff019
|
1887 1888 |
if (!access_ok(VERIFY_WRITE, events, maxevents * sizeof(struct epoll_event))) return -EFAULT; |
7699acd13
|
1889 1890 |
/* Get the "struct file *" for the eventpoll file */ |
2903ff019
|
1891 1892 1893 |
f = fdget(epfd); if (!f.file) return -EBADF; |
7699acd13
|
1894 1895 1896 1897 1898 1899 |
/* * We have to check that the file structure underneath the fd * the user passed to us _is_ an eventpoll file. */ error = -EINVAL; |
2903ff019
|
1900 |
if (!is_file_epoll(f.file)) |
7699acd13
|
1901 1902 1903 1904 1905 1906 |
goto error_fput; /* * At this point it is safe to assume that the "private_data" contains * our own data structure. */ |
2903ff019
|
1907 |
ep = f.file->private_data; |
7699acd13
|
1908 1909 1910 1911 1912 |
/* Time to fish for events ... */ error = ep_poll(ep, events, maxevents, timeout); error_fput: |
2903ff019
|
1913 |
fdput(f); |
7699acd13
|
1914 1915 |
return error; } |
7699acd13
|
1916 1917 1918 1919 |
/* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_pwait(2). */ |
5a8a82b1d
|
1920 1921 1922 |
SYSCALL_DEFINE6(epoll_pwait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout, const sigset_t __user *, sigmask, size_t, sigsetsize) |
7699acd13
|
1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 |
{ int error; sigset_t ksigmask, sigsaved; /* * If the caller wants a certain signal mask to be set during the wait, * we apply it here. */ if (sigmask) { if (sigsetsize != sizeof(sigset_t)) return -EINVAL; if (copy_from_user(&ksigmask, sigmask, sizeof(ksigmask))) return -EFAULT; |
77d559180
|
1936 1937 |
sigsaved = current->blocked; set_current_blocked(&ksigmask); |
7699acd13
|
1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 |
} error = sys_epoll_wait(epfd, events, maxevents, timeout); /* * If we changed the signal mask, we need to restore the original one. * In case we've got a signal while waiting, we do not restore the * signal mask yet, and we allow do_signal() to deliver the signal on * the way back to userspace, before the signal mask is restored. */ if (sigmask) { if (error == -EINTR) { memcpy(¤t->saved_sigmask, &sigsaved, |
c7ea76302
|
1951 |
sizeof(sigsaved)); |
4e4c22c71
|
1952 |
set_restore_sigmask(); |
7699acd13
|
1953 |
} else |
77d559180
|
1954 |
set_current_blocked(&sigsaved); |
7699acd13
|
1955 1956 1957 1958 |
} return error; } |
35280bd4a
|
1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 |
#ifdef CONFIG_COMPAT COMPAT_SYSCALL_DEFINE6(epoll_pwait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout, const compat_sigset_t __user *, sigmask, compat_size_t, sigsetsize) { long err; compat_sigset_t csigmask; sigset_t ksigmask, sigsaved; /* * If the caller wants a certain signal mask to be set during the wait, * we apply it here. */ if (sigmask) { if (sigsetsize != sizeof(compat_sigset_t)) return -EINVAL; if (copy_from_user(&csigmask, sigmask, sizeof(csigmask))) return -EFAULT; sigset_from_compat(&ksigmask, &csigmask); |
77d559180
|
1980 1981 |
sigsaved = current->blocked; set_current_blocked(&ksigmask); |
35280bd4a
|
1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 |
} err = sys_epoll_wait(epfd, events, maxevents, timeout); /* * If we changed the signal mask, we need to restore the original one. * In case we've got a signal while waiting, we do not restore the * signal mask yet, and we allow do_signal() to deliver the signal on * the way back to userspace, before the signal mask is restored. */ if (sigmask) { if (err == -EINTR) { memcpy(¤t->saved_sigmask, &sigsaved, sizeof(sigsaved)); set_restore_sigmask(); } else |
77d559180
|
1998 |
set_current_blocked(&sigsaved); |
35280bd4a
|
1999 2000 2001 2002 2003 |
} return err; } #endif |
1da177e4c
|
2004 2005 |
static int __init eventpoll_init(void) { |
7ef9964e6
|
2006 2007 2008 |
struct sysinfo si; si_meminfo(&si); |
9df04e1f2
|
2009 2010 2011 2012 |
/* * Allows top 4% of lomem to be allocated for epoll watches (per user). */ max_user_watches = (((si.totalram - si.totalhigh) / 25) << PAGE_SHIFT) / |
7ef9964e6
|
2013 |
EP_ITEM_COST; |
52bd19f76
|
2014 |
BUG_ON(max_user_watches < 0); |
1da177e4c
|
2015 |
|
22bacca48
|
2016 2017 2018 2019 2020 |
/* * Initialize the structure used to perform epoll file descriptor * inclusion loops checks. */ ep_nested_calls_init(&poll_loop_ncalls); |
1da177e4c
|
2021 |
/* Initialize the structure used to perform safe poll wait head wake ups */ |
5071f97ec
|
2022 2023 2024 2025 |
ep_nested_calls_init(&poll_safewake_ncalls); /* Initialize the structure used to perform file's f_op->poll() calls */ ep_nested_calls_init(&poll_readywalk_ncalls); |
1da177e4c
|
2026 |
|
39732ca5a
|
2027 2028 2029 2030 2031 |
/* * We can have many thousands of epitems, so prevent this from * using an extra cache line on 64-bit (and smaller) CPUs */ BUILD_BUG_ON(sizeof(void *) <= 8 && sizeof(struct epitem) > 128); |
1da177e4c
|
2032 2033 |
/* Allocates slab cache used to allocate "struct epitem" items */ epi_cache = kmem_cache_create("eventpoll_epi", sizeof(struct epitem), |
bb57c3edc
|
2034 |
0, SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL); |
1da177e4c
|
2035 2036 2037 |
/* Allocates slab cache used to allocate "struct eppoll_entry" */ pwq_cache = kmem_cache_create("eventpoll_pwq", |
bb57c3edc
|
2038 |
sizeof(struct eppoll_entry), 0, SLAB_PANIC, NULL); |
1da177e4c
|
2039 |
|
1da177e4c
|
2040 |
return 0; |
1da177e4c
|
2041 |
} |
cea692418
|
2042 |
fs_initcall(eventpoll_init); |