Blame view
fs/eventpoll.c
63.9 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 |
#include <linux/init.h> #include <linux/kernel.h> |
174cd4b1e
|
15 |
#include <linux/sched/signal.h> |
1da177e4c
|
16 17 18 19 20 21 22 |
#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> |
7c0f6ba68
|
36 |
#include <linux/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> |
bf3b9f637
|
44 |
#include <net/busy_poll.h> |
1da177e4c
|
45 |
|
1da177e4c
|
46 47 48 49 |
/* * LOCKING: * There are three level of locking required by epoll : * |
144efe3e3
|
50 |
* 1) epmutex (mutex) |
c7ea76302
|
51 52 |
* 2) ep->mtx (mutex) * 3) ep->lock (spinlock) |
1da177e4c
|
53 54 55 56 57 58 59 60 61 |
* * 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
|
62 63 64 65 66 |
* 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
|
67 68 |
* 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
|
69 |
* close()d without a previous call to epoll_ctl(EPOLL_CTL_DEL). |
22bacca48
|
70 71 72 73 74 75 76 |
* 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
|
77 78 79 80 81 82 83 84 85 |
* 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
|
86 87 88 |
* 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
|
89 |
* Events that require holding "epmutex" are very rare, while for |
d47de16c7
|
90 91 |
* normal operations the epoll private "ep->mtx" will guarantee * a better scalability. |
1da177e4c
|
92 |
*/ |
1da177e4c
|
93 |
/* Epoll private bits inside the event mask */ |
df0108c5d
|
94 |
#define EP_PRIVATE_BITS (EPOLLWAKEUP | EPOLLONESHOT | EPOLLET | EPOLLEXCLUSIVE) |
1da177e4c
|
95 |
|
b6a515c8a
|
96 97 98 99 |
#define EPOLLINOUT_BITS (POLLIN | POLLOUT) #define EPOLLEXCLUSIVE_OK_BITS (EPOLLINOUT_BITS | POLLERR | POLLHUP | \ EPOLLWAKEUP | EPOLLET | EPOLLEXCLUSIVE) |
5071f97ec
|
100 101 |
/* Maximum number of nesting allowed inside epoll sets */ #define EP_MAX_NESTS 4 |
1da177e4c
|
102 |
|
b611967de
|
103 |
#define EP_MAX_EVENTS (INT_MAX / sizeof(struct epoll_event)) |
d47de16c7
|
104 |
#define EP_UNACTIVE_PTR ((void *) -1L) |
7ef9964e6
|
105 |
#define EP_ITEM_COST (sizeof(struct epitem) + sizeof(struct eppoll_entry)) |
1da177e4c
|
106 107 108 |
struct epoll_filefd { struct file *file; int fd; |
39732ca5a
|
109 |
} __packed; |
1da177e4c
|
110 111 |
/* |
5071f97ec
|
112 113 |
* Structure used to track possible nested calls, for too deep recursions * and loop cycles. |
1da177e4c
|
114 |
*/ |
5071f97ec
|
115 |
struct nested_call_node { |
1da177e4c
|
116 |
struct list_head llink; |
5071f97ec
|
117 |
void *cookie; |
3fe4a975d
|
118 |
void *ctx; |
1da177e4c
|
119 120 121 |
}; /* |
5071f97ec
|
122 123 |
* This structure is used as collector for nested calls, to check for * maximum recursion dept and loop cycles. |
1da177e4c
|
124 |
*/ |
5071f97ec
|
125 126 |
struct nested_calls { struct list_head tasks_call_list; |
1da177e4c
|
127 128 129 130 |
spinlock_t lock; }; /* |
d47de16c7
|
131 132 |
* Each file descriptor added to the eventpoll interface will * have an entry of this type linked to the "rbr" RB tree. |
39732ca5a
|
133 134 |
* 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
|
135 136 |
*/ struct epitem { |
ae10b2b4e
|
137 138 139 140 141 142 |
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
|
143 144 145 |
/* List header used to link this structure to the eventpoll ready list */ struct list_head rdllink; |
c7ea76302
|
146 147 148 149 150 |
/* * Works together "struct eventpoll"->ovflist in keeping the * single linked chain of items. */ struct epitem *next; |
d47de16c7
|
151 152 153 154 155 156 157 158 159 160 161 |
/* 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
|
162 163 |
/* List header used to link this item to the "struct file" items list */ struct list_head fllink; |
4d7e30d98
|
164 |
/* wakeup_source used when EPOLLWAKEUP is set */ |
eea1d5859
|
165 |
struct wakeup_source __rcu *ws; |
4d7e30d98
|
166 |
|
c7ea76302
|
167 168 |
/* The structure that describe the interested events and the source fd */ struct epoll_event event; |
d47de16c7
|
169 170 171 |
}; /* |
1da177e4c
|
172 |
* This structure is stored inside the "private_data" member of the file |
bf6a41db7
|
173 |
* structure and represents the main data structure for the eventpoll |
1da177e4c
|
174 175 176 |
* interface. */ struct eventpoll { |
bf6a41db7
|
177 |
/* Protect the access to this structure */ |
c7ea76302
|
178 |
spinlock_t lock; |
1da177e4c
|
179 180 |
/* |
d47de16c7
|
181 182 183 184 |
* 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
|
185 |
*/ |
d47de16c7
|
186 |
struct mutex mtx; |
1da177e4c
|
187 188 189 190 191 192 193 194 195 |
/* 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
|
196 |
/* RB tree root used to store monitored fd structs */ |
b2ac2ea62
|
197 |
struct rb_root_cached rbr; |
d47de16c7
|
198 199 200 |
/* * This is a single linked list that chains all the "struct epitem" that |
25985edce
|
201 |
* happened while transferring ready events to userspace w/out |
d47de16c7
|
202 203 204 |
* holding ->lock. */ struct epitem *ovflist; |
7ef9964e6
|
205 |
|
4d7e30d98
|
206 207 |
/* wakeup_source used when ep_scan_ready_list is running */ struct wakeup_source *ws; |
7ef9964e6
|
208 209 |
/* The user that created the eventpoll descriptor */ struct user_struct *user; |
28d82dc1c
|
210 211 212 213 214 215 |
struct file *file; /* used to optimize loop detection check */ int visited; struct list_head visited_list_link; |
bf3b9f637
|
216 217 218 219 220 |
#ifdef CONFIG_NET_RX_BUSY_POLL /* used to track busy poll napi_id */ unsigned int napi_id; #endif |
1da177e4c
|
221 222 223 224 225 226 227 228 |
}; /* 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
|
229 |
struct epitem *base; |
1da177e4c
|
230 231 232 233 234 |
/* * Wait queue item that will be linked to the target file wait * queue head. */ |
ac6424b98
|
235 |
wait_queue_entry_t wait; |
1da177e4c
|
236 237 238 239 |
/* The wait queue head that linked the "wait" wait queue item */ wait_queue_head_t *whead; }; |
1da177e4c
|
240 241 242 243 244 |
/* Wrapper struct used by poll queueing */ struct ep_pqueue { poll_table pt; struct epitem *epi; }; |
5071f97ec
|
245 246 247 248 249 |
/* Used by the ep_send_events() function as callback private data */ struct ep_send_events_data { int maxevents; struct epoll_event __user *events; }; |
1da177e4c
|
250 |
/* |
7ef9964e6
|
251 252 |
* Configuration options available inside /proc/sys/fs/epoll/ */ |
7ef9964e6
|
253 |
/* Maximum number of epoll watched descriptors, per user */ |
52bd19f76
|
254 |
static long max_user_watches __read_mostly; |
7ef9964e6
|
255 256 |
/* |
d47de16c7
|
257 |
* This mutex is used to serialize ep_free() and eventpoll_release_file(). |
1da177e4c
|
258 |
*/ |
7ef9964e6
|
259 |
static DEFINE_MUTEX(epmutex); |
1da177e4c
|
260 |
|
22bacca48
|
261 262 |
/* Used to check for epoll file descriptor inclusion loops */ static struct nested_calls poll_loop_ncalls; |
5071f97ec
|
263 264 265 266 267 |
/* 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
|
268 269 |
/* Slab cache used to allocate "struct epitem" */ |
e18b890bb
|
270 |
static struct kmem_cache *epi_cache __read_mostly; |
1da177e4c
|
271 272 |
/* Slab cache used to allocate "struct eppoll_entry" */ |
e18b890bb
|
273 |
static struct kmem_cache *pwq_cache __read_mostly; |
1da177e4c
|
274 |
|
28d82dc1c
|
275 276 277 278 279 280 281 282 |
/* 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
|
283 284 285 |
#ifdef CONFIG_SYSCTL #include <linux/sysctl.h> |
52bd19f76
|
286 287 |
static long zero; static long long_max = LONG_MAX; |
7ef9964e6
|
288 |
|
1f7e0616c
|
289 |
struct ctl_table epoll_table[] = { |
7ef9964e6
|
290 |
{ |
7ef9964e6
|
291 292 |
.procname = "max_user_watches", .data = &max_user_watches, |
52bd19f76
|
293 |
.maxlen = sizeof(max_user_watches), |
7ef9964e6
|
294 |
.mode = 0644, |
52bd19f76
|
295 |
.proc_handler = proc_doulongvec_minmax, |
7ef9964e6
|
296 |
.extra1 = &zero, |
52bd19f76
|
297 |
.extra2 = &long_max, |
7ef9964e6
|
298 |
}, |
ab09203e3
|
299 |
{ } |
7ef9964e6
|
300 301 |
}; #endif /* CONFIG_SYSCTL */ |
28d82dc1c
|
302 303 304 305 306 307 |
static const struct file_operations eventpoll_fops; static inline int is_file_epoll(struct file *f) { return f->f_op == &eventpoll_fops; } |
b030a4dd6
|
308 |
|
67647d0fb
|
309 |
/* Setup the structure that is used as key for the RB tree */ |
b030a4dd6
|
310 311 312 313 314 315 |
static inline void ep_set_ffd(struct epoll_filefd *ffd, struct file *file, int fd) { ffd->file = file; ffd->fd = fd; } |
67647d0fb
|
316 |
/* Compare RB tree keys */ |
b030a4dd6
|
317 318 319 320 321 322 |
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
|
323 324 325 326 327 |
/* Tells us if the item is currently linked */ static inline int ep_is_linked(struct list_head *p) { return !list_empty(p); } |
ac6424b98
|
328 |
static inline struct eppoll_entry *ep_pwq_from_wait(wait_queue_entry_t *p) |
971316f05
|
329 330 331 |
{ return container_of(p, struct eppoll_entry, wait); } |
b030a4dd6
|
332 |
/* Get the "struct epitem" from a wait queue pointer */ |
ac6424b98
|
333 |
static inline struct epitem *ep_item_from_wait(wait_queue_entry_t *p) |
b030a4dd6
|
334 335 336 337 338 |
{ return container_of(p, struct eppoll_entry, wait)->base; } /* Get the "struct epitem" from an epoll queue wrapper */ |
cdac75e6f
|
339 |
static inline struct epitem *ep_item_from_epqueue(poll_table *p) |
b030a4dd6
|
340 341 342 343 344 |
{ return container_of(p, struct ep_pqueue, pt)->epi; } /* Tells if the epoll_ctl(2) operation needs an event copy from userspace */ |
6192bd536
|
345 |
static inline int ep_op_has_event(int op) |
b030a4dd6
|
346 |
{ |
a80a6b85b
|
347 |
return op != EPOLL_CTL_DEL; |
b030a4dd6
|
348 |
} |
1da177e4c
|
349 |
/* Initialize the poll safe wake up structure */ |
5071f97ec
|
350 |
static void ep_nested_calls_init(struct nested_calls *ncalls) |
1da177e4c
|
351 |
{ |
5071f97ec
|
352 353 |
INIT_LIST_HEAD(&ncalls->tasks_call_list); spin_lock_init(&ncalls->lock); |
1da177e4c
|
354 |
} |
5071f97ec
|
355 |
/** |
3fb0e584a
|
356 357 358 359 360 361 362 363 364 365 366 |
* 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; } |
bf3b9f637
|
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 |
#ifdef CONFIG_NET_RX_BUSY_POLL static bool ep_busy_loop_end(void *p, unsigned long start_time) { struct eventpoll *ep = p; return ep_events_available(ep) || busy_loop_timeout(start_time); } #endif /* CONFIG_NET_RX_BUSY_POLL */ /* * Busy poll if globally on and supporting sockets found && no events, * busy loop will return if need_resched or ep_events_available. * * we must do our busy polling with irqs enabled */ static void ep_busy_loop(struct eventpoll *ep, int nonblock) { #ifdef CONFIG_NET_RX_BUSY_POLL unsigned int napi_id = READ_ONCE(ep->napi_id); if ((napi_id >= MIN_NAPI_ID) && net_busy_loop_on()) napi_busy_loop(napi_id, nonblock ? NULL : ep_busy_loop_end, ep); #endif } static inline void ep_reset_busy_poll_napi_id(struct eventpoll *ep) { #ifdef CONFIG_NET_RX_BUSY_POLL if (ep->napi_id) ep->napi_id = 0; #endif } /* * Set epoll busy poll NAPI ID from sk. */ static inline void ep_set_busy_poll_napi_id(struct epitem *epi) { #ifdef CONFIG_NET_RX_BUSY_POLL struct eventpoll *ep; unsigned int napi_id; struct socket *sock; struct sock *sk; int err; if (!net_busy_loop_on()) return; sock = sock_from_file(epi->ffd.file, &err); if (!sock) return; sk = sock->sk; if (!sk) return; napi_id = READ_ONCE(sk->sk_napi_id); ep = epi->ep; /* Non-NAPI IDs can be rejected * or * Nothing to do if we already have this ID */ if (napi_id < MIN_NAPI_ID || napi_id == ep->napi_id) return; /* record NAPI ID for use in next busy poll */ ep->napi_id = napi_id; #endif } |
3fb0e584a
|
437 |
/** |
5071f97ec
|
438 439 440 441 442 443 444 445 446 447 |
* 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
|
448 |
* @ctx: This instance context. |
5071f97ec
|
449 450 451 |
* * Returns: Returns the code returned by the @nproc callback, or -1 if * the maximum recursion limit has been exceeded. |
1da177e4c
|
452 |
*/ |
5071f97ec
|
453 454 |
static int ep_call_nested(struct nested_calls *ncalls, int max_nests, int (*nproc)(void *, void *, int), void *priv, |
3fe4a975d
|
455 |
void *cookie, void *ctx) |
1da177e4c
|
456 |
{ |
5071f97ec
|
457 |
int error, call_nests = 0; |
1da177e4c
|
458 |
unsigned long flags; |
5071f97ec
|
459 460 461 |
struct list_head *lsthead = &ncalls->tasks_call_list; struct nested_call_node *tncur; struct nested_call_node tnode; |
1da177e4c
|
462 |
|
5071f97ec
|
463 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c
|
464 |
|
5071f97ec
|
465 466 467 468 469 |
/* * 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
|
470 |
list_for_each_entry(tncur, lsthead, llink) { |
3fe4a975d
|
471 |
if (tncur->ctx == ctx && |
5071f97ec
|
472 |
(tncur->cookie == cookie || ++call_nests > max_nests)) { |
1da177e4c
|
473 474 475 476 |
/* * Ops ... loop detected or maximum nest level reached. * We abort this wake by breaking the cycle itself. */ |
abff55cee
|
477 478 |
error = -1; goto out_unlock; |
1da177e4c
|
479 480 |
} } |
5071f97ec
|
481 |
/* Add the current task and cookie to the list */ |
3fe4a975d
|
482 |
tnode.ctx = ctx; |
5071f97ec
|
483 |
tnode.cookie = cookie; |
1da177e4c
|
484 |
list_add(&tnode.llink, lsthead); |
5071f97ec
|
485 |
spin_unlock_irqrestore(&ncalls->lock, flags); |
1da177e4c
|
486 |
|
5071f97ec
|
487 488 |
/* Call the nested function */ error = (*nproc)(priv, cookie, call_nests); |
1da177e4c
|
489 490 |
/* Remove the current task from the list */ |
5071f97ec
|
491 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c
|
492 |
list_del(&tnode.llink); |
3fe4a975d
|
493 |
out_unlock: |
5071f97ec
|
494 495 496 497 |
spin_unlock_irqrestore(&ncalls->lock, flags); return error; } |
02edc6fc4
|
498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 |
/* * 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
|
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 |
#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
|
540 541 |
static int ep_poll_wakeup_proc(void *priv, void *cookie, int call_nests) { |
2dfa4eeab
|
542 543 |
ep_wake_up_nested((wait_queue_head_t *) cookie, POLLIN, 1 + call_nests); |
5071f97ec
|
544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 |
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
|
559 |
int this_cpu = get_cpu(); |
5071f97ec
|
560 |
ep_call_nested(&poll_safewake_ncalls, EP_MAX_NESTS, |
3fe4a975d
|
561 562 563 |
ep_poll_wakeup_proc, NULL, wq, (void *) (long) this_cpu); put_cpu(); |
1da177e4c
|
564 |
} |
971316f05
|
565 566 567 568 569 |
static void ep_remove_wait_queue(struct eppoll_entry *pwq) { wait_queue_head_t *whead; rcu_read_lock(); |
138e4ad67
|
570 571 572 573 574 575 576 |
/* * If it is cleared by POLLFREE, it should be rcu-safe. * If we read NULL we need a barrier paired with * smp_store_release() in ep_poll_callback(), otherwise * we rely on whead->lock. */ whead = smp_load_acquire(&pwq->whead); |
971316f05
|
577 578 579 580 |
if (whead) remove_wait_queue(whead, &pwq->wait); rcu_read_unlock(); } |
1da177e4c
|
581 |
/* |
d1bc90dd5
|
582 583 584 |
* This function unregisters poll callbacks from the associated file * descriptor. Must be called with "mtx" held (or "epmutex" if called from * ep_free). |
1da177e4c
|
585 |
*/ |
7699acd13
|
586 |
static void ep_unregister_pollwait(struct eventpoll *ep, struct epitem *epi) |
1da177e4c
|
587 |
{ |
7699acd13
|
588 589 |
struct list_head *lsthead = &epi->pwqlist; struct eppoll_entry *pwq; |
1da177e4c
|
590 |
|
d1bc90dd5
|
591 592 |
while (!list_empty(lsthead)) { pwq = list_first_entry(lsthead, struct eppoll_entry, llink); |
1da177e4c
|
593 |
|
d1bc90dd5
|
594 |
list_del(&pwq->llink); |
971316f05
|
595 |
ep_remove_wait_queue(pwq); |
d1bc90dd5
|
596 |
kmem_cache_free(pwq_cache, pwq); |
1da177e4c
|
597 |
} |
1da177e4c
|
598 |
} |
eea1d5859
|
599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 |
/* 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
|
630 631 632 633 634 635 636 637 |
/** * 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
|
638 |
* @depth: The current depth of recursive f_op->poll calls. |
67347fe4e
|
639 |
* @ep_locked: caller already holds ep->mtx |
5071f97ec
|
640 641 642 643 644 645 |
* * 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
|
646 |
void *priv, int depth, bool ep_locked) |
5071f97ec
|
647 648 649 650 |
{ int error, pwake = 0; unsigned long flags; struct epitem *epi, *nepi; |
296e236e9
|
651 |
LIST_HEAD(txlist); |
5071f97ec
|
652 653 654 |
/* * We need to lock this because we could be hit by |
e057e15ff
|
655 |
* eventpoll_release_file() and epoll_ctl(). |
5071f97ec
|
656 |
*/ |
67347fe4e
|
657 658 659 |
if (!ep_locked) mutex_lock_nested(&ep->mtx, depth); |
5071f97ec
|
660 661 662 663 664 665 666 667 668 669 |
/* * 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
|
670 |
list_splice_init(&ep->rdllist, &txlist); |
5071f97ec
|
671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 |
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
|
693 |
if (!ep_is_linked(&epi->rdllink)) { |
5071f97ec
|
694 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
695 |
ep_pm_stay_awake(epi); |
4d7e30d98
|
696 |
} |
5071f97ec
|
697 698 699 700 701 702 703 704 705 706 707 708 |
} /* * 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
|
709 |
__pm_relax(ep->ws); |
5071f97ec
|
710 711 712 |
if (!list_empty(&ep->rdllist)) { /* |
296e236e9
|
713 714 |
* Wake up (if active) both the eventpoll wait list and * the ->poll() wait list (delayed after we release the lock). |
5071f97ec
|
715 716 717 718 719 720 721 |
*/ if (waitqueue_active(&ep->wq)) wake_up_locked(&ep->wq); if (waitqueue_active(&ep->poll_wait)) pwake++; } spin_unlock_irqrestore(&ep->lock, flags); |
67347fe4e
|
722 723 |
if (!ep_locked) mutex_unlock(&ep->mtx); |
5071f97ec
|
724 725 726 727 728 729 730 |
/* We have to call this outside the lock */ if (pwake) ep_poll_safewake(&ep->poll_wait); return error; } |
ae10b2b4e
|
731 732 733 734 735 |
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
|
736 |
/* |
7699acd13
|
737 |
* Removes a "struct epitem" from the eventpoll RB tree and deallocates |
c7ea76302
|
738 |
* all the associated resources. Must be called with "mtx" held. |
7699acd13
|
739 740 741 |
*/ static int ep_remove(struct eventpoll *ep, struct epitem *epi) { |
7699acd13
|
742 743 |
unsigned long flags; struct file *file = epi->ffd.file; |
1da177e4c
|
744 745 |
/* |
7699acd13
|
746 747 748 749 750 751 |
* 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
|
752 |
*/ |
7699acd13
|
753 |
ep_unregister_pollwait(ep, epi); |
1da177e4c
|
754 |
|
7699acd13
|
755 |
/* Remove the current item from the list of epoll hooks */ |
684999149
|
756 |
spin_lock(&file->f_lock); |
ae10b2b4e
|
757 |
list_del_rcu(&epi->fllink); |
684999149
|
758 |
spin_unlock(&file->f_lock); |
1da177e4c
|
759 |
|
b2ac2ea62
|
760 |
rb_erase_cached(&epi->rbn, &ep->rbr); |
1da177e4c
|
761 |
|
c7ea76302
|
762 763 764 765 |
spin_lock_irqsave(&ep->lock, flags); if (ep_is_linked(&epi->rdllink)) list_del_init(&epi->rdllink); spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
766 |
|
eea1d5859
|
767 |
wakeup_source_unregister(ep_wakeup_source(epi)); |
ae10b2b4e
|
768 769 770 771 772 773 774 775 |
/* * 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
|
776 |
|
52bd19f76
|
777 |
atomic_long_dec(&ep->user->epoll_watches); |
7ef9964e6
|
778 |
|
c7ea76302
|
779 |
return 0; |
1da177e4c
|
780 |
} |
7699acd13
|
781 |
static void ep_free(struct eventpoll *ep) |
1da177e4c
|
782 |
{ |
7699acd13
|
783 784 |
struct rb_node *rbp; struct epitem *epi; |
1da177e4c
|
785 |
|
7699acd13
|
786 787 |
/* We need to release all tasks waiting for these file */ if (waitqueue_active(&ep->poll_wait)) |
5071f97ec
|
788 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
789 |
|
7699acd13
|
790 791 792 |
/* * We need to lock this because we could be hit by * eventpoll_release_file() while we're freeing the "struct eventpoll". |
d47de16c7
|
793 |
* We do not need to hold "ep->mtx" here because the epoll file |
7699acd13
|
794 795 |
* 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
|
796 |
* holding "epmutex" is sufficient here. |
7699acd13
|
797 798 |
*/ mutex_lock(&epmutex); |
1da177e4c
|
799 800 |
/* |
7699acd13
|
801 |
* Walks through the whole tree by unregistering poll callbacks. |
1da177e4c
|
802 |
*/ |
b2ac2ea62
|
803 |
for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) { |
7699acd13
|
804 805 806 |
epi = rb_entry(rbp, struct epitem, rbn); ep_unregister_pollwait(ep, epi); |
91cf5ab60
|
807 |
cond_resched(); |
7699acd13
|
808 |
} |
1da177e4c
|
809 810 |
/* |
7699acd13
|
811 812 |
* 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
|
813 |
* holding "epmutex" we can be sure that no file cleanup code will hit |
7699acd13
|
814 |
* us during this operation. So we can avoid the lock on "ep->lock". |
ddf676c38
|
815 816 |
* We do not need to lock ep->mtx, either, we only do it to prevent * a lockdep warning. |
1da177e4c
|
817 |
*/ |
ddf676c38
|
818 |
mutex_lock(&ep->mtx); |
b2ac2ea62
|
819 |
while ((rbp = rb_first_cached(&ep->rbr)) != NULL) { |
7699acd13
|
820 821 |
epi = rb_entry(rbp, struct epitem, rbn); ep_remove(ep, epi); |
91cf5ab60
|
822 |
cond_resched(); |
7699acd13
|
823 |
} |
ddf676c38
|
824 |
mutex_unlock(&ep->mtx); |
1da177e4c
|
825 |
|
7699acd13
|
826 |
mutex_unlock(&epmutex); |
d47de16c7
|
827 |
mutex_destroy(&ep->mtx); |
7ef9964e6
|
828 |
free_uid(ep->user); |
4d7e30d98
|
829 |
wakeup_source_unregister(ep->ws); |
f0ee9aabb
|
830 |
kfree(ep); |
7699acd13
|
831 |
} |
1da177e4c
|
832 |
|
7699acd13
|
833 834 835 |
static int ep_eventpoll_release(struct inode *inode, struct file *file) { struct eventpoll *ep = file->private_data; |
1da177e4c
|
836 |
|
f0ee9aabb
|
837 |
if (ep) |
7699acd13
|
838 |
ep_free(ep); |
7699acd13
|
839 |
|
7699acd13
|
840 |
return 0; |
1da177e4c
|
841 |
} |
450d89ec0
|
842 843 844 845 846 847 |
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
|
848 849 |
static int ep_read_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
5071f97ec
|
850 851 |
{ struct epitem *epi, *tmp; |
626cf2366
|
852 |
poll_table pt; |
5071f97ec
|
853 |
|
626cf2366
|
854 |
init_poll_funcptr(&pt, NULL); |
450d89ec0
|
855 |
|
5071f97ec
|
856 |
list_for_each_entry_safe(epi, tmp, head, rdllink) { |
450d89ec0
|
857 |
if (ep_item_poll(epi, &pt)) |
5071f97ec
|
858 |
return POLLIN | POLLRDNORM; |
296e236e9
|
859 |
else { |
5071f97ec
|
860 861 862 863 864 |
/* * 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
|
865 |
__pm_relax(ep_wakeup_source(epi)); |
5071f97ec
|
866 |
list_del_init(&epi->rdllink); |
296e236e9
|
867 |
} |
5071f97ec
|
868 869 870 871 |
} return 0; } |
67347fe4e
|
872 873 874 875 876 877 878 |
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
|
879 880 |
static int ep_poll_readyevents_proc(void *priv, void *cookie, int call_nests) { |
67347fe4e
|
881 882 883 884 |
struct readyevents_arg *arg = priv; return ep_scan_ready_list(arg->ep, ep_read_events_proc, NULL, call_nests + 1, arg->locked); |
5071f97ec
|
885 |
} |
7699acd13
|
886 887 |
static unsigned int ep_eventpoll_poll(struct file *file, poll_table *wait) { |
5071f97ec
|
888 |
int pollflags; |
7699acd13
|
889 |
struct eventpoll *ep = file->private_data; |
67347fe4e
|
890 891 892 893 894 895 896 897 |
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
|
898 |
|
7699acd13
|
899 900 |
/* Insert inside our poll wait queue */ poll_wait(file, &ep->poll_wait, wait); |
5071f97ec
|
901 902 903 904 905 906 907 |
/* * 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
|
908 |
ep_poll_readyevents_proc, &arg, ep, current); |
7699acd13
|
909 |
|
296e236e9
|
910 |
return pollflags != -1 ? pollflags : 0; |
7699acd13
|
911 |
} |
138d22b58
|
912 |
#ifdef CONFIG_PROC_FS |
a3816ab0e
|
913 |
static void ep_show_fdinfo(struct seq_file *m, struct file *f) |
138d22b58
|
914 915 916 |
{ struct eventpoll *ep = f->private_data; struct rb_node *rbp; |
138d22b58
|
917 918 |
mutex_lock(&ep->mtx); |
b2ac2ea62
|
919 |
for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) { |
138d22b58
|
920 |
struct epitem *epi = rb_entry(rbp, struct epitem, rbn); |
77493f04b
|
921 |
struct inode *inode = file_inode(epi->ffd.file); |
138d22b58
|
922 |
|
77493f04b
|
923 924 925 |
seq_printf(m, "tfd: %8d events: %8x data: %16llx " " pos:%lli ino:%lx sdev:%x ", |
a3816ab0e
|
926 |
epi->ffd.fd, epi->event.events, |
77493f04b
|
927 928 929 |
(long long)epi->event.data, (long long)epi->ffd.file->f_pos, inode->i_ino, inode->i_sb->s_dev); |
a3816ab0e
|
930 |
if (seq_has_overflowed(m)) |
138d22b58
|
931 932 933 |
break; } mutex_unlock(&ep->mtx); |
138d22b58
|
934 935 |
} #endif |
7699acd13
|
936 937 |
/* File callbacks that implement the eventpoll file behaviour */ static const struct file_operations eventpoll_fops = { |
138d22b58
|
938 939 940 |
#ifdef CONFIG_PROC_FS .show_fdinfo = ep_show_fdinfo, #endif |
7699acd13
|
941 |
.release = ep_eventpoll_release, |
6038f373a
|
942 943 |
.poll = ep_eventpoll_poll, .llseek = noop_llseek, |
7699acd13
|
944 |
}; |
b611967de
|
945 |
/* |
7699acd13
|
946 947 948 |
* 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
|
949 |
*/ |
7699acd13
|
950 |
void eventpoll_release_file(struct file *file) |
b611967de
|
951 |
{ |
7699acd13
|
952 |
struct eventpoll *ep; |
ebe06187b
|
953 |
struct epitem *epi, *next; |
b611967de
|
954 955 |
/* |
684999149
|
956 |
* We don't want to get "file->f_lock" because it is not |
7699acd13
|
957 |
* necessary. It is not necessary because we're in the "struct file" |
25985edce
|
958 |
* cleanup path, and this means that no one is using this file anymore. |
5071f97ec
|
959 |
* So, for example, epoll_ctl() cannot hit here since if we reach this |
67647d0fb
|
960 |
* point, the file counter already went to zero and fget() would fail. |
d47de16c7
|
961 |
* The only hit might come from ep_free() but by holding the mutex |
7699acd13
|
962 |
* will correctly serialize the operation. We do need to acquire |
d47de16c7
|
963 |
* "ep->mtx" after "epmutex" because ep_remove() requires it when called |
7699acd13
|
964 |
* from anywhere but ep_free(). |
684999149
|
965 966 |
* * Besides, ep_remove() acquires the lock, so we can't hold it here. |
b611967de
|
967 |
*/ |
7699acd13
|
968 |
mutex_lock(&epmutex); |
ebe06187b
|
969 |
list_for_each_entry_safe(epi, next, &file->f_ep_links, fllink) { |
7699acd13
|
970 |
ep = epi->ep; |
d8805e633
|
971 |
mutex_lock_nested(&ep->mtx, 0); |
7699acd13
|
972 |
ep_remove(ep, epi); |
d47de16c7
|
973 |
mutex_unlock(&ep->mtx); |
b611967de
|
974 |
} |
7699acd13
|
975 |
mutex_unlock(&epmutex); |
b611967de
|
976 |
} |
53d2be79d
|
977 |
static int ep_alloc(struct eventpoll **pep) |
1da177e4c
|
978 |
{ |
7ef9964e6
|
979 980 981 |
int error; struct user_struct *user; struct eventpoll *ep; |
1da177e4c
|
982 |
|
7ef9964e6
|
983 |
user = get_current_user(); |
7ef9964e6
|
984 985 986 987 |
error = -ENOMEM; ep = kzalloc(sizeof(*ep), GFP_KERNEL); if (unlikely(!ep)) goto free_uid; |
1da177e4c
|
988 |
|
c7ea76302
|
989 |
spin_lock_init(&ep->lock); |
d47de16c7
|
990 |
mutex_init(&ep->mtx); |
1da177e4c
|
991 992 993 |
init_waitqueue_head(&ep->wq); init_waitqueue_head(&ep->poll_wait); INIT_LIST_HEAD(&ep->rdllist); |
b2ac2ea62
|
994 |
ep->rbr = RB_ROOT_CACHED; |
d47de16c7
|
995 |
ep->ovflist = EP_UNACTIVE_PTR; |
7ef9964e6
|
996 |
ep->user = user; |
1da177e4c
|
997 |
|
53d2be79d
|
998 |
*pep = ep; |
1da177e4c
|
999 |
|
1da177e4c
|
1000 |
return 0; |
7ef9964e6
|
1001 1002 1003 1004 |
free_uid: free_uid(user); return error; |
1da177e4c
|
1005 |
} |
1da177e4c
|
1006 |
/* |
c7ea76302
|
1007 1008 1009 |
* 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
|
1010 1011 1012 1013 |
*/ static struct epitem *ep_find(struct eventpoll *ep, struct file *file, int fd) { int kcmp; |
1da177e4c
|
1014 1015 1016 |
struct rb_node *rbp; struct epitem *epi, *epir = NULL; struct epoll_filefd ffd; |
b030a4dd6
|
1017 |
ep_set_ffd(&ffd, file, fd); |
b2ac2ea62
|
1018 |
for (rbp = ep->rbr.rb_root.rb_node; rbp; ) { |
1da177e4c
|
1019 |
epi = rb_entry(rbp, struct epitem, rbn); |
b030a4dd6
|
1020 |
kcmp = ep_cmp_ffd(&ffd, &epi->ffd); |
1da177e4c
|
1021 1022 1023 1024 1025 |
if (kcmp > 0) rbp = rbp->rb_right; else if (kcmp < 0) rbp = rbp->rb_left; else { |
1da177e4c
|
1026 1027 1028 1029 |
epir = epi; break; } } |
1da177e4c
|
1030 |
|
1da177e4c
|
1031 1032 |
return epir; } |
92ef6da3d
|
1033 |
#ifdef CONFIG_CHECKPOINT_RESTORE |
0791e3644
|
1034 1035 1036 1037 |
static struct epitem *ep_find_tfd(struct eventpoll *ep, int tfd, unsigned long toff) { struct rb_node *rbp; struct epitem *epi; |
b2ac2ea62
|
1038 |
for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) { |
0791e3644
|
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 |
epi = rb_entry(rbp, struct epitem, rbn); if (epi->ffd.fd == tfd) { if (toff == 0) return epi; else toff--; } cond_resched(); } return NULL; } struct file *get_epoll_tfile_raw_ptr(struct file *file, int tfd, unsigned long toff) { struct file *file_raw; struct eventpoll *ep; struct epitem *epi; if (!is_file_epoll(file)) return ERR_PTR(-EINVAL); ep = file->private_data; mutex_lock(&ep->mtx); epi = ep_find_tfd(ep, tfd, toff); if (epi) file_raw = epi->ffd.file; else file_raw = ERR_PTR(-ENOENT); mutex_unlock(&ep->mtx); return file_raw; } |
92ef6da3d
|
1074 |
#endif /* CONFIG_CHECKPOINT_RESTORE */ |
0791e3644
|
1075 |
|
1da177e4c
|
1076 |
/* |
7699acd13
|
1077 |
* This is the callback that is passed to the wait queue wakeup |
bf6a41db7
|
1078 |
* mechanism. It is called by the stored file descriptors when they |
7699acd13
|
1079 |
* have events to report. |
1da177e4c
|
1080 |
*/ |
ac6424b98
|
1081 |
static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, void *key) |
1da177e4c
|
1082 |
{ |
7699acd13
|
1083 1084 1085 1086 |
int pwake = 0; unsigned long flags; struct epitem *epi = ep_item_from_wait(wait); struct eventpoll *ep = epi->ep; |
df0108c5d
|
1087 |
int ewake = 0; |
1da177e4c
|
1088 |
|
c7ea76302
|
1089 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1090 |
|
bf3b9f637
|
1091 |
ep_set_busy_poll_napi_id(epi); |
7699acd13
|
1092 1093 1094 1095 1096 1097 1098 |
/* * 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
|
1099 1100 1101 |
goto out_unlock; /* |
2dfa4eeab
|
1102 1103 1104 1105 1106 1107 1108 1109 1110 |
* 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
|
1111 |
* If we are transferring events to userspace, we can hold no locks |
d47de16c7
|
1112 |
* (because we're accessing user memory, and because of linux f_op->poll() |
bf6a41db7
|
1113 |
* semantics). All the events that happen during that period of time are |
d47de16c7
|
1114 1115 1116 1117 1118 1119 |
* 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
|
1120 1121 1122 1123 1124 1125 1126 |
if (epi->ws) { /* * Activate ep->ws since epi->ws may get * deactivated at any time. */ __pm_stay_awake(ep->ws); } |
d47de16c7
|
1127 1128 1129 |
} goto out_unlock; } |
1da177e4c
|
1130 |
|
7699acd13
|
1131 |
/* If this file is already in the ready list we exit soon */ |
4d7e30d98
|
1132 |
if (!ep_is_linked(&epi->rdllink)) { |
5071f97ec
|
1133 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1134 |
ep_pm_stay_awake_rcu(epi); |
4d7e30d98
|
1135 |
} |
7699acd13
|
1136 |
|
7699acd13
|
1137 1138 1139 1140 |
/* * Wake up ( if active ) both the eventpoll wait list and the ->poll() * wait list. */ |
df0108c5d
|
1141 |
if (waitqueue_active(&ep->wq)) { |
b6a515c8a
|
1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 |
if ((epi->event.events & EPOLLEXCLUSIVE) && !((unsigned long)key & POLLFREE)) { switch ((unsigned long)key & EPOLLINOUT_BITS) { case POLLIN: if (epi->event.events & POLLIN) ewake = 1; break; case POLLOUT: if (epi->event.events & POLLOUT) ewake = 1; break; case 0: ewake = 1; break; } } |
4a6e9e2ce
|
1158 |
wake_up_locked(&ep->wq); |
df0108c5d
|
1159 |
} |
7699acd13
|
1160 1161 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
d47de16c7
|
1162 |
out_unlock: |
c7ea76302
|
1163 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1164 |
|
7699acd13
|
1165 1166 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1167 |
ep_poll_safewake(&ep->poll_wait); |
7699acd13
|
1168 |
|
138e4ad67
|
1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 |
if (!(epi->event.events & EPOLLEXCLUSIVE)) ewake = 1; if ((unsigned long)key & POLLFREE) { /* * If we race with ep_remove_wait_queue() it can miss * ->whead = NULL and do another remove_wait_queue() after * us, so we can't use __remove_wait_queue(). */ list_del_init(&wait->entry); /* * ->whead != NULL protects us from the race with ep_free() * or ep_remove(), ep_remove_wait_queue() takes whead->lock * held by the caller. Once we nullify it, nothing protects * ep/epi or even wait. */ smp_store_release(&ep_pwq_from_wait(wait)->whead, NULL); } |
df0108c5d
|
1187 |
|
138e4ad67
|
1188 |
return ewake; |
7699acd13
|
1189 |
} |
1da177e4c
|
1190 1191 1192 1193 1194 1195 1196 1197 |
/* * 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
|
1198 |
struct epitem *epi = ep_item_from_epqueue(pt); |
1da177e4c
|
1199 |
struct eppoll_entry *pwq; |
e94b17660
|
1200 |
if (epi->nwait >= 0 && (pwq = kmem_cache_alloc(pwq_cache, GFP_KERNEL))) { |
1da177e4c
|
1201 1202 1203 |
init_waitqueue_func_entry(&pwq->wait, ep_poll_callback); pwq->whead = whead; pwq->base = epi; |
df0108c5d
|
1204 1205 1206 1207 |
if (epi->event.events & EPOLLEXCLUSIVE) add_wait_queue_exclusive(whead, &pwq->wait); else add_wait_queue(whead, &pwq->wait); |
1da177e4c
|
1208 1209 |
list_add_tail(&pwq->llink, &epi->pwqlist); epi->nwait++; |
296e236e9
|
1210 |
} else { |
1da177e4c
|
1211 1212 |
/* We have to signal that an error occurred */ epi->nwait = -1; |
296e236e9
|
1213 |
} |
1da177e4c
|
1214 |
} |
1da177e4c
|
1215 1216 1217 |
static void ep_rbtree_insert(struct eventpoll *ep, struct epitem *epi) { int kcmp; |
b2ac2ea62
|
1218 |
struct rb_node **p = &ep->rbr.rb_root.rb_node, *parent = NULL; |
1da177e4c
|
1219 |
struct epitem *epic; |
b2ac2ea62
|
1220 |
bool leftmost = true; |
1da177e4c
|
1221 1222 1223 1224 |
while (*p) { parent = *p; epic = rb_entry(parent, struct epitem, rbn); |
b030a4dd6
|
1225 |
kcmp = ep_cmp_ffd(&epi->ffd, &epic->ffd); |
b2ac2ea62
|
1226 |
if (kcmp > 0) { |
1da177e4c
|
1227 |
p = &parent->rb_right; |
b2ac2ea62
|
1228 1229 |
leftmost = false; } else |
1da177e4c
|
1230 1231 1232 |
p = &parent->rb_left; } rb_link_node(&epi->rbn, parent, p); |
b2ac2ea62
|
1233 |
rb_insert_color_cached(&epi->rbn, &ep->rbr, leftmost); |
1da177e4c
|
1234 |
} |
a80a6b85b
|
1235 |
|
28d82dc1c
|
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 |
#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
|
1253 1254 1255 |
/* Allow an arbitrary number of depth 1 paths */ if (nests == 0) return 0; |
28d82dc1c
|
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 |
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
|
1275 1276 1277 |
/* 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
|
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 |
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
|
1300 |
rcu_read_unlock(); |
28d82dc1c
|
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 |
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
|
1316 1317 1318 1319 1320 |
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
|
1321 1322 1323 1324 1325 1326 1327 1328 1329 |
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
|
1330 1331 1332 |
static int ep_create_wakeup_source(struct epitem *epi) { const char *name; |
eea1d5859
|
1333 |
struct wakeup_source *ws; |
4d7e30d98
|
1334 1335 1336 1337 1338 1339 1340 1341 |
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
|
1342 1343 1344 |
ws = wakeup_source_register(name); if (!ws) |
4d7e30d98
|
1345 |
return -ENOMEM; |
eea1d5859
|
1346 |
rcu_assign_pointer(epi->ws, ws); |
4d7e30d98
|
1347 1348 1349 |
return 0; } |
eea1d5859
|
1350 1351 |
/* rare code path, only used when EPOLL_CTL_MOD removes a wakeup source */ static noinline void ep_destroy_wakeup_source(struct epitem *epi) |
4d7e30d98
|
1352 |
{ |
eea1d5859
|
1353 |
struct wakeup_source *ws = ep_wakeup_source(epi); |
d6d67e723
|
1354 |
RCU_INIT_POINTER(epi->ws, NULL); |
eea1d5859
|
1355 1356 1357 1358 1359 1360 1361 1362 |
/* * 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
|
1363 |
} |
c7ea76302
|
1364 1365 1366 |
/* * Must be called with "mtx" held. */ |
1da177e4c
|
1367 |
static int ep_insert(struct eventpoll *ep, struct epoll_event *event, |
67347fe4e
|
1368 |
struct file *tfile, int fd, int full_check) |
1da177e4c
|
1369 1370 1371 |
{ int error, revents, pwake = 0; unsigned long flags; |
52bd19f76
|
1372 |
long user_watches; |
1da177e4c
|
1373 1374 |
struct epitem *epi; struct ep_pqueue epq; |
52bd19f76
|
1375 1376 |
user_watches = atomic_long_read(&ep->user->epoll_watches); if (unlikely(user_watches >= max_user_watches)) |
7ef9964e6
|
1377 |
return -ENOSPC; |
e94b17660
|
1378 |
if (!(epi = kmem_cache_alloc(epi_cache, GFP_KERNEL))) |
7ef9964e6
|
1379 |
return -ENOMEM; |
1da177e4c
|
1380 1381 |
/* Item initialization follow here ... */ |
1da177e4c
|
1382 1383 |
INIT_LIST_HEAD(&epi->rdllink); INIT_LIST_HEAD(&epi->fllink); |
1da177e4c
|
1384 1385 |
INIT_LIST_HEAD(&epi->pwqlist); epi->ep = ep; |
b030a4dd6
|
1386 |
ep_set_ffd(&epi->ffd, tfile, fd); |
1da177e4c
|
1387 |
epi->event = *event; |
1da177e4c
|
1388 |
epi->nwait = 0; |
d47de16c7
|
1389 |
epi->next = EP_UNACTIVE_PTR; |
4d7e30d98
|
1390 1391 1392 1393 1394 |
if (epi->event.events & EPOLLWAKEUP) { error = ep_create_wakeup_source(epi); if (error) goto error_create_wakeup_source; } else { |
eea1d5859
|
1395 |
RCU_INIT_POINTER(epi->ws, NULL); |
4d7e30d98
|
1396 |
} |
1da177e4c
|
1397 1398 1399 1400 1401 1402 1403 1404 |
/* 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
|
1405 1406 1407 |
* been increased by the caller of this function. Note that after * this operation completes, the poll callback can start hitting * the new item. |
1da177e4c
|
1408 |
*/ |
450d89ec0
|
1409 |
revents = ep_item_poll(epi, &epq.pt); |
1da177e4c
|
1410 1411 1412 1413 1414 1415 |
/* * 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
|
1416 |
error = -ENOMEM; |
1da177e4c
|
1417 |
if (epi->nwait < 0) |
7699acd13
|
1418 |
goto error_unregister; |
1da177e4c
|
1419 1420 |
/* Add the current item to the list of active epoll hook for this file */ |
684999149
|
1421 |
spin_lock(&tfile->f_lock); |
ae10b2b4e
|
1422 |
list_add_tail_rcu(&epi->fllink, &tfile->f_ep_links); |
684999149
|
1423 |
spin_unlock(&tfile->f_lock); |
1da177e4c
|
1424 |
|
c7ea76302
|
1425 1426 1427 1428 |
/* * 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
|
1429 |
ep_rbtree_insert(ep, epi); |
28d82dc1c
|
1430 1431 |
/* now check if we've created too many backpaths */ error = -EINVAL; |
67347fe4e
|
1432 |
if (full_check && reverse_path_check()) |
28d82dc1c
|
1433 |
goto error_remove_epi; |
c7ea76302
|
1434 1435 |
/* We have to drop the new item inside our item list to keep track of it */ spin_lock_irqsave(&ep->lock, flags); |
bf3b9f637
|
1436 1437 |
/* record NAPI ID of new item if present */ ep_set_busy_poll_napi_id(epi); |
1da177e4c
|
1438 |
/* If the file is already "ready" we drop it inside the ready list */ |
b030a4dd6
|
1439 |
if ((revents & event->events) && !ep_is_linked(&epi->rdllink)) { |
1da177e4c
|
1440 |
list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1441 |
ep_pm_stay_awake(epi); |
1da177e4c
|
1442 1443 1444 |
/* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce
|
1445 |
wake_up_locked(&ep->wq); |
1da177e4c
|
1446 1447 1448 |
if (waitqueue_active(&ep->poll_wait)) pwake++; } |
c7ea76302
|
1449 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1450 |
|
52bd19f76
|
1451 |
atomic_long_inc(&ep->user->epoll_watches); |
7ef9964e6
|
1452 |
|
1da177e4c
|
1453 1454 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1455 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
1456 |
|
1da177e4c
|
1457 |
return 0; |
28d82dc1c
|
1458 1459 |
error_remove_epi: spin_lock(&tfile->f_lock); |
ae10b2b4e
|
1460 |
list_del_rcu(&epi->fllink); |
28d82dc1c
|
1461 |
spin_unlock(&tfile->f_lock); |
b2ac2ea62
|
1462 |
rb_erase_cached(&epi->rbn, &ep->rbr); |
28d82dc1c
|
1463 |
|
7699acd13
|
1464 |
error_unregister: |
1da177e4c
|
1465 1466 1467 1468 |
ep_unregister_pollwait(ep, epi); /* * We need to do this because an event could have been arrived on some |
67647d0fb
|
1469 1470 1471 |
* 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
|
1472 |
*/ |
c7ea76302
|
1473 |
spin_lock_irqsave(&ep->lock, flags); |
b030a4dd6
|
1474 |
if (ep_is_linked(&epi->rdllink)) |
6192bd536
|
1475 |
list_del_init(&epi->rdllink); |
c7ea76302
|
1476 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1477 |
|
eea1d5859
|
1478 |
wakeup_source_unregister(ep_wakeup_source(epi)); |
4d7e30d98
|
1479 1480 |
error_create_wakeup_source: |
b030a4dd6
|
1481 |
kmem_cache_free(epi_cache, epi); |
7ef9964e6
|
1482 |
|
1da177e4c
|
1483 1484 |
return error; } |
1da177e4c
|
1485 1486 |
/* * Modify the interest event mask by dropping an event if the new mask |
c7ea76302
|
1487 |
* has a match in the current file status. Must be called with "mtx" held. |
1da177e4c
|
1488 1489 1490 1491 1492 |
*/ static int ep_modify(struct eventpoll *ep, struct epitem *epi, struct epoll_event *event) { int pwake = 0; unsigned int revents; |
626cf2366
|
1493 1494 1495 |
poll_table pt; init_poll_funcptr(&pt, NULL); |
1da177e4c
|
1496 1497 |
/* |
e057e15ff
|
1498 1499 1500 |
* 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
|
1501 |
*/ |
128dd1759
|
1502 |
epi->event.events = event->events; /* need barrier below */ |
e057e15ff
|
1503 |
epi->event.data = event->data; /* protected by mtx */ |
4d7e30d98
|
1504 |
if (epi->event.events & EPOLLWAKEUP) { |
eea1d5859
|
1505 |
if (!ep_has_wakeup_source(epi)) |
4d7e30d98
|
1506 |
ep_create_wakeup_source(epi); |
eea1d5859
|
1507 |
} else if (ep_has_wakeup_source(epi)) { |
4d7e30d98
|
1508 1509 |
ep_destroy_wakeup_source(epi); } |
1da177e4c
|
1510 1511 |
/* |
128dd1759
|
1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 |
* 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
|
1532 1533 1534 |
* 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
|
1535 |
revents = ep_item_poll(epi, &pt); |
1da177e4c
|
1536 |
|
1da177e4c
|
1537 |
/* |
c7ea76302
|
1538 |
* If the item is "hot" and it is not registered inside the ready |
67647d0fb
|
1539 |
* list, push it inside. |
1da177e4c
|
1540 |
*/ |
c7ea76302
|
1541 |
if (revents & event->events) { |
e057e15ff
|
1542 |
spin_lock_irq(&ep->lock); |
c7ea76302
|
1543 1544 |
if (!ep_is_linked(&epi->rdllink)) { list_add_tail(&epi->rdllink, &ep->rdllist); |
eea1d5859
|
1545 |
ep_pm_stay_awake(epi); |
c7ea76302
|
1546 1547 1548 |
/* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce
|
1549 |
wake_up_locked(&ep->wq); |
c7ea76302
|
1550 1551 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
7699acd13
|
1552 |
} |
e057e15ff
|
1553 |
spin_unlock_irq(&ep->lock); |
7699acd13
|
1554 |
} |
1da177e4c
|
1555 |
|
7699acd13
|
1556 1557 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec
|
1558 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c
|
1559 |
|
7699acd13
|
1560 |
return 0; |
1da177e4c
|
1561 |
} |
296e236e9
|
1562 1563 |
static int ep_send_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
1da177e4c
|
1564 |
{ |
5071f97ec
|
1565 1566 |
struct ep_send_events_data *esed = priv; int eventcnt; |
296e236e9
|
1567 |
unsigned int revents; |
5071f97ec
|
1568 1569 |
struct epitem *epi; struct epoll_event __user *uevent; |
eea1d5859
|
1570 |
struct wakeup_source *ws; |
626cf2366
|
1571 1572 1573 |
poll_table pt; init_poll_funcptr(&pt, NULL); |
1da177e4c
|
1574 |
|
296e236e9
|
1575 |
/* |
5071f97ec
|
1576 1577 1578 |
* 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
|
1579 |
*/ |
5071f97ec
|
1580 1581 1582 |
for (eventcnt = 0, uevent = esed->events; !list_empty(head) && eventcnt < esed->maxevents;) { epi = list_first_entry(head, struct epitem, rdllink); |
d47de16c7
|
1583 |
|
4d7e30d98
|
1584 1585 1586 1587 1588 1589 1590 1591 1592 |
/* * 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
|
1593 1594 1595 1596 1597 1598 |
ws = ep_wakeup_source(epi); if (ws) { if (ws->active) __pm_stay_awake(ep->ws); __pm_relax(ws); } |
d47de16c7
|
1599 |
list_del_init(&epi->rdllink); |
1da177e4c
|
1600 |
|
450d89ec0
|
1601 |
revents = ep_item_poll(epi, &pt); |
5071f97ec
|
1602 |
|
296e236e9
|
1603 |
/* |
5071f97ec
|
1604 1605 1606 1607 |
* 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
|
1608 1609 |
*/ if (revents) { |
5071f97ec
|
1610 |
if (__put_user(revents, &uevent->events) || |
d03058828
|
1611 1612 |
__put_user(epi->event.data, &uevent->data)) { list_add(&epi->rdllink, head); |
eea1d5859
|
1613 |
ep_pm_stay_awake(epi); |
296e236e9
|
1614 |
return eventcnt ? eventcnt : -EFAULT; |
d03058828
|
1615 |
} |
296e236e9
|
1616 |
eventcnt++; |
5071f97ec
|
1617 |
uevent++; |
296e236e9
|
1618 1619 1620 1621 1622 1623 1624 1625 |
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
|
1626 |
* availability. At this point, no one can insert |
296e236e9
|
1627 1628 1629 1630 1631 1632 |
* 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
|
1633 |
ep_pm_stay_awake(epi); |
296e236e9
|
1634 1635 1636 |
} } } |
5071f97ec
|
1637 1638 1639 |
return eventcnt; } |
d47de16c7
|
1640 |
|
296e236e9
|
1641 1642 |
static int ep_send_events(struct eventpoll *ep, struct epoll_event __user *events, int maxevents) |
5071f97ec
|
1643 1644 |
{ struct ep_send_events_data esed; |
1da177e4c
|
1645 |
|
5071f97ec
|
1646 1647 |
esed.maxevents = maxevents; esed.events = events; |
6192bd536
|
1648 |
|
67347fe4e
|
1649 |
return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false); |
1da177e4c
|
1650 |
} |
766b9f928
|
1651 |
static inline struct timespec64 ep_set_mstimeout(long ms) |
0781b909b
|
1652 |
{ |
766b9f928
|
1653 |
struct timespec64 now, ts = { |
0781b909b
|
1654 1655 1656 |
.tv_sec = ms / MSEC_PER_SEC, .tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC), }; |
766b9f928
|
1657 1658 |
ktime_get_ts64(&now); return timespec64_add_safe(now, ts); |
0781b909b
|
1659 |
} |
f4d93ad74
|
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 |
/** * 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
|
1677 1678 1679 |
static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events, int maxevents, long timeout) { |
f4d93ad74
|
1680 |
int res = 0, eavail, timed_out = 0; |
1da177e4c
|
1681 |
unsigned long flags; |
da8b44d5a
|
1682 |
u64 slack = 0; |
ac6424b98
|
1683 |
wait_queue_entry_t wait; |
95aac7b1c
|
1684 1685 1686 |
ktime_t expires, *to = NULL; if (timeout > 0) { |
766b9f928
|
1687 |
struct timespec64 end_time = ep_set_mstimeout(timeout); |
0781b909b
|
1688 |
|
95aac7b1c
|
1689 1690 |
slack = select_estimate_accuracy(&end_time); to = &expires; |
766b9f928
|
1691 |
*to = timespec64_to_ktime(end_time); |
95aac7b1c
|
1692 |
} else if (timeout == 0) { |
f4d93ad74
|
1693 1694 1695 1696 |
/* * Avoid the unnecessary trip to the wait queue loop, if the * caller specified a non blocking operation. */ |
95aac7b1c
|
1697 |
timed_out = 1; |
f4d93ad74
|
1698 1699 |
spin_lock_irqsave(&ep->lock, flags); goto check_events; |
95aac7b1c
|
1700 |
} |
1da177e4c
|
1701 |
|
f4d93ad74
|
1702 |
fetch_events: |
bf3b9f637
|
1703 1704 1705 |
if (!ep_events_available(ep)) ep_busy_loop(ep, timed_out); |
c7ea76302
|
1706 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1707 |
|
3fb0e584a
|
1708 |
if (!ep_events_available(ep)) { |
1da177e4c
|
1709 |
/* |
bf3b9f637
|
1710 1711 1712 1713 1714 1715 1716 |
* Busy poll timed out. Drop NAPI ID for now, we can add * it back in when we have moved a socket with a valid NAPI * ID onto the ready list. */ ep_reset_busy_poll_napi_id(ep); /* |
1da177e4c
|
1717 1718 1719 1720 1721 |
* 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
|
1722 |
__add_wait_queue_exclusive(&ep->wq, &wait); |
1da177e4c
|
1723 1724 1725 1726 1727 1728 1729 1730 |
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); |
c257a340e
|
1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 |
/* * Always short-circuit for fatal signals to allow * threads to make a timely exit without the chance of * finding more events available and fetching * repeatedly. */ if (fatal_signal_pending(current)) { res = -EINTR; break; } |
3fb0e584a
|
1741 |
if (ep_events_available(ep) || timed_out) |
1da177e4c
|
1742 1743 1744 1745 1746 |
break; if (signal_pending(current)) { res = -EINTR; break; } |
c7ea76302
|
1747 |
spin_unlock_irqrestore(&ep->lock, flags); |
c511851de
|
1748 |
if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS)) |
95aac7b1c
|
1749 |
timed_out = 1; |
c7ea76302
|
1750 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c
|
1751 |
} |
1da177e4c
|
1752 |
|
4d5755b14
|
1753 1754 |
__remove_wait_queue(&ep->wq, &wait); __set_current_state(TASK_RUNNING); |
1da177e4c
|
1755 |
} |
f4d93ad74
|
1756 |
check_events: |
1da177e4c
|
1757 |
/* Is it worth to try to dig for events ? */ |
3fb0e584a
|
1758 |
eavail = ep_events_available(ep); |
1da177e4c
|
1759 |
|
c7ea76302
|
1760 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c
|
1761 1762 1763 1764 1765 1766 1767 |
/* * 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
|
1768 |
!(res = ep_send_events(ep, events, maxevents)) && !timed_out) |
f4d93ad74
|
1769 |
goto fetch_events; |
1da177e4c
|
1770 1771 1772 |
return res; } |
22bacca48
|
1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 |
/** * 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
|
1793 |
struct eventpoll *ep_tovisit; |
22bacca48
|
1794 1795 |
struct rb_node *rbp; struct epitem *epi; |
d8805e633
|
1796 |
mutex_lock_nested(&ep->mtx, call_nests + 1); |
28d82dc1c
|
1797 1798 |
ep->visited = 1; list_add(&ep->visited_list_link, &visited_list); |
b2ac2ea62
|
1799 |
for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) { |
22bacca48
|
1800 1801 |
epi = rb_entry(rbp, struct epitem, rbn); if (unlikely(is_file_epoll(epi->ffd.file))) { |
28d82dc1c
|
1802 1803 1804 |
ep_tovisit = epi->ffd.file->private_data; if (ep_tovisit->visited) continue; |
22bacca48
|
1805 |
error = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
28d82dc1c
|
1806 1807 |
ep_loop_check_proc, epi->ffd.file, ep_tovisit, current); |
22bacca48
|
1808 1809 |
if (error != 0) break; |
28d82dc1c
|
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 |
} 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
|
1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 |
} } 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
|
1842 1843 1844 1845 |
int ret; struct eventpoll *ep_cur, *ep_next; ret = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
22bacca48
|
1846 |
ep_loop_check_proc, file, ep, current); |
28d82dc1c
|
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 |
/* 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
|
1867 |
} |
7699acd13
|
1868 |
/* |
523723bb5
|
1869 |
* Open an eventpoll file descriptor. |
7699acd13
|
1870 |
*/ |
5a8a82b1d
|
1871 |
SYSCALL_DEFINE1(epoll_create1, int, flags) |
7699acd13
|
1872 |
{ |
28d82dc1c
|
1873 |
int error, fd; |
bb57c3edc
|
1874 |
struct eventpoll *ep = NULL; |
28d82dc1c
|
1875 |
struct file *file; |
7699acd13
|
1876 |
|
e38b36f32
|
1877 1878 |
/* Check the EPOLL_* constant for consistency. */ BUILD_BUG_ON(EPOLL_CLOEXEC != O_CLOEXEC); |
296e236e9
|
1879 1880 |
if (flags & ~EPOLL_CLOEXEC) return -EINVAL; |
7699acd13
|
1881 |
/* |
bb57c3edc
|
1882 |
* Create the internal data structure ("struct eventpoll"). |
7699acd13
|
1883 |
*/ |
9fe5ad9c8
|
1884 |
error = ep_alloc(&ep); |
bb57c3edc
|
1885 1886 |
if (error < 0) return error; |
7699acd13
|
1887 1888 |
/* * Creates all the items needed to setup an eventpoll file. That is, |
2030a42ce
|
1889 |
* a file structure and a free file descriptor. |
7699acd13
|
1890 |
*/ |
28d82dc1c
|
1891 1892 1893 1894 1895 1896 |
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
|
1897 |
O_RDWR | (flags & O_CLOEXEC)); |
28d82dc1c
|
1898 1899 1900 1901 |
if (IS_ERR(file)) { error = PTR_ERR(file); goto out_free_fd; } |
28d82dc1c
|
1902 |
ep->file = file; |
98022748f
|
1903 |
fd_install(fd, file); |
28d82dc1c
|
1904 1905 1906 1907 1908 1909 |
return fd; out_free_fd: put_unused_fd(fd); out_free_ep: ep_free(ep); |
bb57c3edc
|
1910 |
return error; |
7699acd13
|
1911 |
} |
5a8a82b1d
|
1912 |
SYSCALL_DEFINE1(epoll_create, int, size) |
a0998b50c
|
1913 |
{ |
bfe3891a5
|
1914 |
if (size <= 0) |
9fe5ad9c8
|
1915 1916 1917 |
return -EINVAL; return sys_epoll_create1(0); |
a0998b50c
|
1918 |
} |
7699acd13
|
1919 1920 1921 |
/* * The following function implements the controller interface for * the eventpoll file that enables the insertion/removal/change of |
67647d0fb
|
1922 |
* file descriptors inside the interest set. |
7699acd13
|
1923 |
*/ |
5a8a82b1d
|
1924 1925 |
SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd, struct epoll_event __user *, event) |
7699acd13
|
1926 1927 |
{ int error; |
67347fe4e
|
1928 |
int full_check = 0; |
7e3fb5842
|
1929 |
struct fd f, tf; |
7699acd13
|
1930 1931 1932 |
struct eventpoll *ep; struct epitem *epi; struct epoll_event epds; |
67347fe4e
|
1933 |
struct eventpoll *tep = NULL; |
7699acd13
|
1934 |
|
7699acd13
|
1935 1936 1937 1938 |
error = -EFAULT; if (ep_op_has_event(op) && copy_from_user(&epds, event, sizeof(struct epoll_event))) goto error_return; |
7699acd13
|
1939 |
error = -EBADF; |
7e3fb5842
|
1940 1941 |
f = fdget(epfd); if (!f.file) |
7699acd13
|
1942 1943 1944 |
goto error_return; /* Get the "struct file *" for the target file */ |
7e3fb5842
|
1945 1946 |
tf = fdget(fd); if (!tf.file) |
7699acd13
|
1947 1948 1949 1950 |
goto error_fput; /* The target file descriptor must support poll */ error = -EPERM; |
72c2d5319
|
1951 |
if (!tf.file->f_op->poll) |
7699acd13
|
1952 |
goto error_tgt_fput; |
4d7e30d98
|
1953 |
/* Check if EPOLLWAKEUP is allowed */ |
c680e41b3
|
1954 1955 |
if (ep_op_has_event(op)) ep_take_care_of_epollwakeup(&epds); |
4d7e30d98
|
1956 |
|
7699acd13
|
1957 1958 1959 1960 1961 1962 |
/* * 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
|
1963 |
if (f.file == tf.file || !is_file_epoll(f.file)) |
7699acd13
|
1964 1965 1966 |
goto error_tgt_fput; /* |
df0108c5d
|
1967 1968 1969 1970 |
* epoll adds to the wakeup queue at EPOLL_CTL_ADD time only, * so EPOLLEXCLUSIVE is not allowed for a EPOLL_CTL_MOD operation. * Also, we do not currently supported nested exclusive wakeups. */ |
c857ab640
|
1971 |
if (ep_op_has_event(op) && (epds.events & EPOLLEXCLUSIVE)) { |
b6a515c8a
|
1972 1973 1974 1975 1976 1977 |
if (op == EPOLL_CTL_MOD) goto error_tgt_fput; if (op == EPOLL_CTL_ADD && (is_file_epoll(tf.file) || (epds.events & ~EPOLLEXCLUSIVE_OK_BITS))) goto error_tgt_fput; } |
df0108c5d
|
1978 1979 |
/* |
7699acd13
|
1980 1981 1982 |
* At this point it is safe to assume that the "private_data" contains * our own data structure. */ |
7e3fb5842
|
1983 |
ep = f.file->private_data; |
7699acd13
|
1984 |
|
22bacca48
|
1985 1986 1987 |
/* * When we insert an epoll file descriptor, inside another epoll file * descriptor, there is the change of creating closed loops, which are |
28d82dc1c
|
1988 1989 1990 1991 |
* 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
|
1992 |
* |
67347fe4e
|
1993 1994 1995 1996 1997 1998 |
* 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
|
1999 |
*/ |
67347fe4e
|
2000 |
mutex_lock_nested(&ep->mtx, 0); |
28d82dc1c
|
2001 |
if (op == EPOLL_CTL_ADD) { |
67347fe4e
|
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 |
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
|
2020 |
} |
67347fe4e
|
2021 2022 |
} } |
7699acd13
|
2023 |
|
67647d0fb
|
2024 2025 2026 2027 2028 |
/* * 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
|
2029 |
epi = ep_find(ep, tf.file, fd); |
7699acd13
|
2030 2031 2032 2033 2034 2035 |
error = -EINVAL; switch (op) { case EPOLL_CTL_ADD: if (!epi) { epds.events |= POLLERR | POLLHUP; |
67347fe4e
|
2036 |
error = ep_insert(ep, &epds, tf.file, fd, full_check); |
7699acd13
|
2037 2038 |
} else error = -EEXIST; |
67347fe4e
|
2039 2040 |
if (full_check) clear_tfile_check_list(); |
7699acd13
|
2041 2042 2043 2044 2045 2046 2047 2048 2049 |
break; case EPOLL_CTL_DEL: if (epi) error = ep_remove(ep, epi); else error = -ENOENT; break; case EPOLL_CTL_MOD: if (epi) { |
b6a515c8a
|
2050 2051 2052 2053 |
if (!(epi->event.events & EPOLLEXCLUSIVE)) { epds.events |= POLLERR | POLLHUP; error = ep_modify(ep, epi, &epds); } |
7699acd13
|
2054 2055 2056 2057 |
} else error = -ENOENT; break; } |
67347fe4e
|
2058 2059 |
if (tep != NULL) mutex_unlock(&tep->mtx); |
d47de16c7
|
2060 |
mutex_unlock(&ep->mtx); |
7699acd13
|
2061 2062 |
error_tgt_fput: |
67347fe4e
|
2063 |
if (full_check) |
22bacca48
|
2064 |
mutex_unlock(&epmutex); |
7e3fb5842
|
2065 |
fdput(tf); |
7699acd13
|
2066 |
error_fput: |
7e3fb5842
|
2067 |
fdput(f); |
7699acd13
|
2068 |
error_return: |
7699acd13
|
2069 2070 2071 2072 2073 2074 2075 2076 |
return error; } /* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_wait(2). */ |
5a8a82b1d
|
2077 2078 |
SYSCALL_DEFINE4(epoll_wait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout) |
7699acd13
|
2079 |
{ |
2903ff019
|
2080 2081 |
int error; struct fd f; |
7699acd13
|
2082 |
struct eventpoll *ep; |
7699acd13
|
2083 2084 2085 2086 2087 |
/* 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
|
2088 2089 |
if (!access_ok(VERIFY_WRITE, events, maxevents * sizeof(struct epoll_event))) return -EFAULT; |
7699acd13
|
2090 2091 |
/* Get the "struct file *" for the eventpoll file */ |
2903ff019
|
2092 2093 2094 |
f = fdget(epfd); if (!f.file) return -EBADF; |
7699acd13
|
2095 2096 2097 2098 2099 2100 |
/* * We have to check that the file structure underneath the fd * the user passed to us _is_ an eventpoll file. */ error = -EINVAL; |
2903ff019
|
2101 |
if (!is_file_epoll(f.file)) |
7699acd13
|
2102 2103 2104 2105 2106 2107 |
goto error_fput; /* * At this point it is safe to assume that the "private_data" contains * our own data structure. */ |
2903ff019
|
2108 |
ep = f.file->private_data; |
7699acd13
|
2109 2110 2111 2112 2113 |
/* Time to fish for events ... */ error = ep_poll(ep, events, maxevents, timeout); error_fput: |
2903ff019
|
2114 |
fdput(f); |
7699acd13
|
2115 2116 |
return error; } |
7699acd13
|
2117 2118 2119 2120 |
/* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_pwait(2). */ |
5a8a82b1d
|
2121 2122 2123 |
SYSCALL_DEFINE6(epoll_pwait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout, const sigset_t __user *, sigmask, size_t, sigsetsize) |
7699acd13
|
2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 |
{ 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
|
2137 2138 |
sigsaved = current->blocked; set_current_blocked(&ksigmask); |
7699acd13
|
2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 |
} 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
|
2152 |
sizeof(sigsaved)); |
4e4c22c71
|
2153 |
set_restore_sigmask(); |
7699acd13
|
2154 |
} else |
77d559180
|
2155 |
set_current_blocked(&sigsaved); |
7699acd13
|
2156 2157 2158 2159 |
} return error; } |
35280bd4a
|
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 |
#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
|
2181 2182 |
sigsaved = current->blocked; set_current_blocked(&ksigmask); |
35280bd4a
|
2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 |
} 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
|
2199 |
set_current_blocked(&sigsaved); |
35280bd4a
|
2200 2201 2202 2203 2204 |
} return err; } #endif |
1da177e4c
|
2205 2206 |
static int __init eventpoll_init(void) { |
7ef9964e6
|
2207 2208 2209 |
struct sysinfo si; si_meminfo(&si); |
9df04e1f2
|
2210 2211 2212 2213 |
/* * Allows top 4% of lomem to be allocated for epoll watches (per user). */ max_user_watches = (((si.totalram - si.totalhigh) / 25) << PAGE_SHIFT) / |
7ef9964e6
|
2214 |
EP_ITEM_COST; |
52bd19f76
|
2215 |
BUG_ON(max_user_watches < 0); |
1da177e4c
|
2216 |
|
22bacca48
|
2217 2218 2219 2220 2221 |
/* * Initialize the structure used to perform epoll file descriptor * inclusion loops checks. */ ep_nested_calls_init(&poll_loop_ncalls); |
1da177e4c
|
2222 |
/* Initialize the structure used to perform safe poll wait head wake ups */ |
5071f97ec
|
2223 2224 2225 2226 |
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
|
2227 |
|
39732ca5a
|
2228 2229 2230 2231 2232 |
/* * 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
|
2233 2234 |
/* Allocates slab cache used to allocate "struct epitem" items */ epi_cache = kmem_cache_create("eventpoll_epi", sizeof(struct epitem), |
bb57c3edc
|
2235 |
0, SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL); |
1da177e4c
|
2236 2237 2238 |
/* Allocates slab cache used to allocate "struct eppoll_entry" */ pwq_cache = kmem_cache_create("eventpoll_pwq", |
bb57c3edc
|
2239 |
sizeof(struct eppoll_entry), 0, SLAB_PANIC, NULL); |
1da177e4c
|
2240 |
|
1da177e4c
|
2241 |
return 0; |
1da177e4c
|
2242 |
} |
cea692418
|
2243 |
fs_initcall(eventpoll_init); |