Blame view
fs/eventpoll.c
49.4 KB
1da177e4c Linux-2.6.12-rc2 |
1 |
/* |
5071f97ec epoll: fix epoll'... |
2 3 |
* fs/eventpoll.c (Efficient event retrieval implementation) * Copyright (C) 2001,...,2009 Davide Libenzi |
1da177e4c Linux-2.6.12-rc2 |
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 Linux-2.6.12-rc2 |
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 Linux-2.6.12-rc2 |
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 Linux-2.6.12-rc2 |
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 [PATCH] sem2mutex... |
33 |
#include <linux/mutex.h> |
da66f7cb0 epoll: use anonym... |
34 |
#include <linux/anon_inodes.h> |
1da177e4c Linux-2.6.12-rc2 |
35 36 37 38 |
#include <asm/uaccess.h> #include <asm/system.h> #include <asm/io.h> #include <asm/mman.h> |
60063497a atomic: use <linu... |
39 |
#include <linux/atomic.h> |
1da177e4c Linux-2.6.12-rc2 |
40 |
|
1da177e4c Linux-2.6.12-rc2 |
41 42 43 44 |
/* * LOCKING: * There are three level of locking required by epoll : * |
144efe3e3 [PATCH] sem2mutex... |
45 |
* 1) epmutex (mutex) |
c7ea76302 epoll locks chang... |
46 47 |
* 2) ep->mtx (mutex) * 3) ep->lock (spinlock) |
1da177e4c Linux-2.6.12-rc2 |
48 49 50 51 52 53 54 55 56 |
* * 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 fix epoll single ... |
57 58 59 60 61 |
* 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 Linux-2.6.12-rc2 |
62 63 |
* 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 fs/eventpoll.c: f... |
64 |
* close()d without a previous call to epoll_ctl(EPOLL_CTL_DEL). |
22bacca48 epoll: prevent cr... |
65 66 67 68 69 70 71 |
* 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 epoll: fix spurio... |
72 73 74 75 76 77 78 79 80 |
* 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 fix epoll single ... |
81 82 83 |
* 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 [PATCH] sem2mutex... |
84 |
* Events that require holding "epmutex" are very rare, while for |
d47de16c7 fix epoll single ... |
85 86 |
* normal operations the epoll private "ep->mtx" will guarantee * a better scalability. |
1da177e4c Linux-2.6.12-rc2 |
87 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
88 89 |
/* Epoll private bits inside the event mask */ #define EP_PRIVATE_BITS (EPOLLONESHOT | EPOLLET) |
5071f97ec epoll: fix epoll'... |
90 91 |
/* Maximum number of nesting allowed inside epoll sets */ #define EP_MAX_NESTS 4 |
1da177e4c Linux-2.6.12-rc2 |
92 |
|
b611967de [PATCH] epoll_pwa... |
93 |
#define EP_MAX_EVENTS (INT_MAX / sizeof(struct epoll_event)) |
d47de16c7 fix epoll single ... |
94 |
#define EP_UNACTIVE_PTR ((void *) -1L) |
7ef9964e6 epoll: introduce ... |
95 |
#define EP_ITEM_COST (sizeof(struct epitem) + sizeof(struct eppoll_entry)) |
1da177e4c Linux-2.6.12-rc2 |
96 97 98 99 100 101 |
struct epoll_filefd { struct file *file; int fd; }; /* |
5071f97ec epoll: fix epoll'... |
102 103 |
* Structure used to track possible nested calls, for too deep recursions * and loop cycles. |
1da177e4c Linux-2.6.12-rc2 |
104 |
*/ |
5071f97ec epoll: fix epoll'... |
105 |
struct nested_call_node { |
1da177e4c Linux-2.6.12-rc2 |
106 |
struct list_head llink; |
5071f97ec epoll: fix epoll'... |
107 |
void *cookie; |
3fe4a975d epoll: fix nested... |
108 |
void *ctx; |
1da177e4c Linux-2.6.12-rc2 |
109 110 111 |
}; /* |
5071f97ec epoll: fix epoll'... |
112 113 |
* This structure is used as collector for nested calls, to check for * maximum recursion dept and loop cycles. |
1da177e4c Linux-2.6.12-rc2 |
114 |
*/ |
5071f97ec epoll: fix epoll'... |
115 116 |
struct nested_calls { struct list_head tasks_call_list; |
1da177e4c Linux-2.6.12-rc2 |
117 118 119 120 |
spinlock_t lock; }; /* |
d47de16c7 fix epoll single ... |
121 122 123 124 |
* Each file descriptor added to the eventpoll interface will * have an entry of this type linked to the "rbr" RB tree. */ struct epitem { |
67647d0fb epoll: fix some c... |
125 |
/* RB tree node used to link this structure to the eventpoll RB tree */ |
d47de16c7 fix epoll single ... |
126 127 128 129 |
struct rb_node rbn; /* List header used to link this structure to the eventpoll ready list */ struct list_head rdllink; |
c7ea76302 epoll locks chang... |
130 131 132 133 134 |
/* * Works together "struct eventpoll"->ovflist in keeping the * single linked chain of items. */ struct epitem *next; |
d47de16c7 fix epoll single ... |
135 136 137 138 139 140 141 142 143 144 145 |
/* 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 fix epoll single ... |
146 147 |
/* List header used to link this item to the "struct file" items list */ struct list_head fllink; |
c7ea76302 epoll locks chang... |
148 149 |
/* The structure that describe the interested events and the source fd */ struct epoll_event event; |
d47de16c7 fix epoll single ... |
150 151 152 |
}; /* |
1da177e4c Linux-2.6.12-rc2 |
153 |
* This structure is stored inside the "private_data" member of the file |
bf6a41db7 fs/eventpoll.c: f... |
154 |
* structure and represents the main data structure for the eventpoll |
1da177e4c Linux-2.6.12-rc2 |
155 156 157 |
* interface. */ struct eventpoll { |
bf6a41db7 fs/eventpoll.c: f... |
158 |
/* Protect the access to this structure */ |
c7ea76302 epoll locks chang... |
159 |
spinlock_t lock; |
1da177e4c Linux-2.6.12-rc2 |
160 161 |
/* |
d47de16c7 fix epoll single ... |
162 163 164 165 |
* 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 Linux-2.6.12-rc2 |
166 |
*/ |
d47de16c7 fix epoll single ... |
167 |
struct mutex mtx; |
1da177e4c Linux-2.6.12-rc2 |
168 169 170 171 172 173 174 175 176 |
/* 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 epoll: fix some c... |
177 |
/* RB tree root used to store monitored fd structs */ |
1da177e4c Linux-2.6.12-rc2 |
178 |
struct rb_root rbr; |
d47de16c7 fix epoll single ... |
179 180 181 |
/* * This is a single linked list that chains all the "struct epitem" that |
25985edce Fix common misspe... |
182 |
* happened while transferring ready events to userspace w/out |
d47de16c7 fix epoll single ... |
183 184 185 |
* holding ->lock. */ struct epitem *ovflist; |
7ef9964e6 epoll: introduce ... |
186 187 188 |
/* The user that created the eventpoll descriptor */ struct user_struct *user; |
28d82dc1c epoll: limit paths |
189 190 191 192 193 194 |
struct file *file; /* used to optimize loop detection check */ int visited; struct list_head visited_list_link; |
1da177e4c Linux-2.6.12-rc2 |
195 196 197 198 199 200 201 202 |
}; /* 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 epoll: use real t... |
203 |
struct epitem *base; |
1da177e4c Linux-2.6.12-rc2 |
204 205 206 207 208 209 210 211 212 213 |
/* * 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 Linux-2.6.12-rc2 |
214 215 216 217 218 |
/* Wrapper struct used by poll queueing */ struct ep_pqueue { poll_table pt; struct epitem *epi; }; |
5071f97ec epoll: fix epoll'... |
219 220 221 222 223 |
/* Used by the ep_send_events() function as callback private data */ struct ep_send_events_data { int maxevents; struct epoll_event __user *events; }; |
1da177e4c Linux-2.6.12-rc2 |
224 |
/* |
7ef9964e6 epoll: introduce ... |
225 226 |
* Configuration options available inside /proc/sys/fs/epoll/ */ |
7ef9964e6 epoll: introduce ... |
227 |
/* Maximum number of epoll watched descriptors, per user */ |
52bd19f76 epoll: convert ma... |
228 |
static long max_user_watches __read_mostly; |
7ef9964e6 epoll: introduce ... |
229 230 |
/* |
d47de16c7 fix epoll single ... |
231 |
* This mutex is used to serialize ep_free() and eventpoll_release_file(). |
1da177e4c Linux-2.6.12-rc2 |
232 |
*/ |
7ef9964e6 epoll: introduce ... |
233 |
static DEFINE_MUTEX(epmutex); |
1da177e4c Linux-2.6.12-rc2 |
234 |
|
22bacca48 epoll: prevent cr... |
235 236 |
/* Used to check for epoll file descriptor inclusion loops */ static struct nested_calls poll_loop_ncalls; |
5071f97ec epoll: fix epoll'... |
237 238 239 240 241 |
/* 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 Linux-2.6.12-rc2 |
242 243 |
/* Slab cache used to allocate "struct epitem" */ |
e18b890bb [PATCH] slab: rem... |
244 |
static struct kmem_cache *epi_cache __read_mostly; |
1da177e4c Linux-2.6.12-rc2 |
245 246 |
/* Slab cache used to allocate "struct eppoll_entry" */ |
e18b890bb [PATCH] slab: rem... |
247 |
static struct kmem_cache *pwq_cache __read_mostly; |
1da177e4c Linux-2.6.12-rc2 |
248 |
|
28d82dc1c epoll: limit paths |
249 250 251 252 253 254 255 256 |
/* 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 epoll: introduce ... |
257 258 259 |
#ifdef CONFIG_SYSCTL #include <linux/sysctl.h> |
52bd19f76 epoll: convert ma... |
260 261 |
static long zero; static long long_max = LONG_MAX; |
7ef9964e6 epoll: introduce ... |
262 263 264 |
ctl_table epoll_table[] = { { |
7ef9964e6 epoll: introduce ... |
265 266 |
.procname = "max_user_watches", .data = &max_user_watches, |
52bd19f76 epoll: convert ma... |
267 |
.maxlen = sizeof(max_user_watches), |
7ef9964e6 epoll: introduce ... |
268 |
.mode = 0644, |
52bd19f76 epoll: convert ma... |
269 |
.proc_handler = proc_doulongvec_minmax, |
7ef9964e6 epoll: introduce ... |
270 |
.extra1 = &zero, |
52bd19f76 epoll: convert ma... |
271 |
.extra2 = &long_max, |
7ef9964e6 epoll: introduce ... |
272 |
}, |
ab09203e3 sysctl fs: Remove... |
273 |
{ } |
7ef9964e6 epoll: introduce ... |
274 275 |
}; #endif /* CONFIG_SYSCTL */ |
28d82dc1c epoll: limit paths |
276 277 278 279 280 281 |
static const struct file_operations eventpoll_fops; static inline int is_file_epoll(struct file *f) { return f->f_op == &eventpoll_fops; } |
b030a4dd6 [PATCH] Remove ev... |
282 |
|
67647d0fb epoll: fix some c... |
283 |
/* Setup the structure that is used as key for the RB tree */ |
b030a4dd6 [PATCH] Remove ev... |
284 285 286 287 288 289 |
static inline void ep_set_ffd(struct epoll_filefd *ffd, struct file *file, int fd) { ffd->file = file; ffd->fd = fd; } |
67647d0fb epoll: fix some c... |
290 |
/* Compare RB tree keys */ |
b030a4dd6 [PATCH] Remove ev... |
291 292 293 294 295 296 |
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 [PATCH] Remove ev... |
297 298 299 300 301 302 303 |
/* Tells us if the item is currently linked */ static inline int ep_is_linked(struct list_head *p) { return !list_empty(p); } /* Get the "struct epitem" from a wait queue pointer */ |
cdac75e6f epoll: avoid kmem... |
304 |
static inline struct epitem *ep_item_from_wait(wait_queue_t *p) |
b030a4dd6 [PATCH] Remove ev... |
305 306 307 308 309 |
{ return container_of(p, struct eppoll_entry, wait)->base; } /* Get the "struct epitem" from an epoll queue wrapper */ |
cdac75e6f epoll: avoid kmem... |
310 |
static inline struct epitem *ep_item_from_epqueue(poll_table *p) |
b030a4dd6 [PATCH] Remove ev... |
311 312 313 314 315 |
{ return container_of(p, struct ep_pqueue, pt)->epi; } /* Tells if the epoll_ctl(2) operation needs an event copy from userspace */ |
6192bd536 epoll: optimizati... |
316 |
static inline int ep_op_has_event(int op) |
b030a4dd6 [PATCH] Remove ev... |
317 318 319 |
{ return op != EPOLL_CTL_DEL; } |
1da177e4c Linux-2.6.12-rc2 |
320 |
/* Initialize the poll safe wake up structure */ |
5071f97ec epoll: fix epoll'... |
321 |
static void ep_nested_calls_init(struct nested_calls *ncalls) |
1da177e4c Linux-2.6.12-rc2 |
322 |
{ |
5071f97ec epoll: fix epoll'... |
323 324 |
INIT_LIST_HEAD(&ncalls->tasks_call_list); spin_lock_init(&ncalls->lock); |
1da177e4c Linux-2.6.12-rc2 |
325 |
} |
5071f97ec epoll: fix epoll'... |
326 |
/** |
3fb0e584a epoll: move ready... |
327 328 329 330 331 332 333 334 335 336 337 338 339 |
* 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 epoll: fix epoll'... |
340 341 342 343 344 345 346 347 348 349 |
* 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 epoll: fix nested... |
350 |
* @ctx: This instance context. |
5071f97ec epoll: fix epoll'... |
351 352 353 |
* * Returns: Returns the code returned by the @nproc callback, or -1 if * the maximum recursion limit has been exceeded. |
1da177e4c Linux-2.6.12-rc2 |
354 |
*/ |
5071f97ec epoll: fix epoll'... |
355 356 |
static int ep_call_nested(struct nested_calls *ncalls, int max_nests, int (*nproc)(void *, void *, int), void *priv, |
3fe4a975d epoll: fix nested... |
357 |
void *cookie, void *ctx) |
1da177e4c Linux-2.6.12-rc2 |
358 |
{ |
5071f97ec epoll: fix epoll'... |
359 |
int error, call_nests = 0; |
1da177e4c Linux-2.6.12-rc2 |
360 |
unsigned long flags; |
5071f97ec epoll: fix epoll'... |
361 362 363 |
struct list_head *lsthead = &ncalls->tasks_call_list; struct nested_call_node *tncur; struct nested_call_node tnode; |
1da177e4c Linux-2.6.12-rc2 |
364 |
|
5071f97ec epoll: fix epoll'... |
365 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
366 |
|
5071f97ec epoll: fix epoll'... |
367 368 369 370 371 |
/* * 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 fs/eventpoll.c: u... |
372 |
list_for_each_entry(tncur, lsthead, llink) { |
3fe4a975d epoll: fix nested... |
373 |
if (tncur->ctx == ctx && |
5071f97ec epoll: fix epoll'... |
374 |
(tncur->cookie == cookie || ++call_nests > max_nests)) { |
1da177e4c Linux-2.6.12-rc2 |
375 376 377 378 |
/* * Ops ... loop detected or maximum nest level reached. * We abort this wake by breaking the cycle itself. */ |
abff55cee epoll: don't use ... |
379 380 |
error = -1; goto out_unlock; |
1da177e4c Linux-2.6.12-rc2 |
381 382 |
} } |
5071f97ec epoll: fix epoll'... |
383 |
/* Add the current task and cookie to the list */ |
3fe4a975d epoll: fix nested... |
384 |
tnode.ctx = ctx; |
5071f97ec epoll: fix epoll'... |
385 |
tnode.cookie = cookie; |
1da177e4c Linux-2.6.12-rc2 |
386 |
list_add(&tnode.llink, lsthead); |
5071f97ec epoll: fix epoll'... |
387 |
spin_unlock_irqrestore(&ncalls->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
388 |
|
5071f97ec epoll: fix epoll'... |
389 390 |
/* Call the nested function */ error = (*nproc)(priv, cookie, call_nests); |
1da177e4c Linux-2.6.12-rc2 |
391 392 |
/* Remove the current task from the list */ |
5071f97ec epoll: fix epoll'... |
393 |
spin_lock_irqsave(&ncalls->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
394 |
list_del(&tnode.llink); |
3fe4a975d epoll: fix nested... |
395 |
out_unlock: |
5071f97ec epoll: fix epoll'... |
396 397 398 399 |
spin_unlock_irqrestore(&ncalls->lock, flags); return error; } |
2dfa4eeab epoll keyed wakeu... |
400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 |
#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 epoll: fix epoll'... |
417 418 |
static int ep_poll_wakeup_proc(void *priv, void *cookie, int call_nests) { |
2dfa4eeab epoll keyed wakeu... |
419 420 |
ep_wake_up_nested((wait_queue_head_t *) cookie, POLLIN, 1 + call_nests); |
5071f97ec epoll: fix epoll'... |
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 |
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 epoll: fix nested... |
436 |
int this_cpu = get_cpu(); |
5071f97ec epoll: fix epoll'... |
437 |
ep_call_nested(&poll_safewake_ncalls, EP_MAX_NESTS, |
3fe4a975d epoll: fix nested... |
438 439 440 |
ep_poll_wakeup_proc, NULL, wq, (void *) (long) this_cpu); put_cpu(); |
1da177e4c Linux-2.6.12-rc2 |
441 |
} |
1da177e4c Linux-2.6.12-rc2 |
442 |
/* |
d1bc90dd5 epoll: remove unn... |
443 444 445 |
* This function unregisters poll callbacks from the associated file * descriptor. Must be called with "mtx" held (or "epmutex" if called from * ep_free). |
1da177e4c Linux-2.6.12-rc2 |
446 |
*/ |
7699acd13 epoll cleanups: e... |
447 |
static void ep_unregister_pollwait(struct eventpoll *ep, struct epitem *epi) |
1da177e4c Linux-2.6.12-rc2 |
448 |
{ |
7699acd13 epoll cleanups: e... |
449 450 |
struct list_head *lsthead = &epi->pwqlist; struct eppoll_entry *pwq; |
1da177e4c Linux-2.6.12-rc2 |
451 |
|
d1bc90dd5 epoll: remove unn... |
452 453 |
while (!list_empty(lsthead)) { pwq = list_first_entry(lsthead, struct eppoll_entry, llink); |
1da177e4c Linux-2.6.12-rc2 |
454 |
|
d1bc90dd5 epoll: remove unn... |
455 456 457 |
list_del(&pwq->llink); remove_wait_queue(pwq->whead, &pwq->wait); kmem_cache_free(pwq_cache, pwq); |
1da177e4c Linux-2.6.12-rc2 |
458 |
} |
1da177e4c Linux-2.6.12-rc2 |
459 |
} |
5071f97ec epoll: fix epoll'... |
460 461 462 463 464 465 466 467 |
/** * 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 epoll: fix spurio... |
468 |
* @depth: The current depth of recursive f_op->poll calls. |
5071f97ec epoll: fix epoll'... |
469 470 471 472 473 474 |
* * 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 *), |
d8805e633 epoll: fix spurio... |
475 476 |
void *priv, int depth) |
5071f97ec epoll: fix epoll'... |
477 478 479 480 |
{ int error, pwake = 0; unsigned long flags; struct epitem *epi, *nepi; |
296e236e9 epoll: fix epoll'... |
481 |
LIST_HEAD(txlist); |
5071f97ec epoll: fix epoll'... |
482 483 484 |
/* * We need to lock this because we could be hit by |
e057e15ff epoll: clean up e... |
485 |
* eventpoll_release_file() and epoll_ctl(). |
5071f97ec epoll: fix epoll'... |
486 |
*/ |
d8805e633 epoll: fix spurio... |
487 |
mutex_lock_nested(&ep->mtx, depth); |
5071f97ec epoll: fix epoll'... |
488 489 490 491 492 493 494 495 496 497 |
/* * 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 epoll: fix epoll'... |
498 |
list_splice_init(&ep->rdllist, &txlist); |
5071f97ec epoll: fix epoll'... |
499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 |
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. */ if (!ep_is_linked(&epi->rdllink)) list_add_tail(&epi->rdllink, &ep->rdllist); } /* * 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); if (!list_empty(&ep->rdllist)) { /* |
296e236e9 epoll: fix epoll'... |
538 539 |
* Wake up (if active) both the eventpoll wait list and * the ->poll() wait list (delayed after we release the lock). |
5071f97ec epoll: fix epoll'... |
540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 |
*/ if (waitqueue_active(&ep->wq)) wake_up_locked(&ep->wq); if (waitqueue_active(&ep->poll_wait)) pwake++; } spin_unlock_irqrestore(&ep->lock, flags); mutex_unlock(&ep->mtx); /* We have to call this outside the lock */ if (pwake) ep_poll_safewake(&ep->poll_wait); return error; } |
1da177e4c Linux-2.6.12-rc2 |
556 |
/* |
7699acd13 epoll cleanups: e... |
557 |
* Removes a "struct epitem" from the eventpoll RB tree and deallocates |
c7ea76302 epoll locks chang... |
558 |
* all the associated resources. Must be called with "mtx" held. |
7699acd13 epoll cleanups: e... |
559 560 561 |
*/ static int ep_remove(struct eventpoll *ep, struct epitem *epi) { |
7699acd13 epoll cleanups: e... |
562 563 |
unsigned long flags; struct file *file = epi->ffd.file; |
1da177e4c Linux-2.6.12-rc2 |
564 565 |
/* |
7699acd13 epoll cleanups: e... |
566 567 568 569 570 571 |
* 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 Linux-2.6.12-rc2 |
572 |
*/ |
7699acd13 epoll cleanups: e... |
573 |
ep_unregister_pollwait(ep, epi); |
1da177e4c Linux-2.6.12-rc2 |
574 |
|
7699acd13 epoll cleanups: e... |
575 |
/* Remove the current item from the list of epoll hooks */ |
684999149 Rename struct fil... |
576 |
spin_lock(&file->f_lock); |
7699acd13 epoll cleanups: e... |
577 578 |
if (ep_is_linked(&epi->fllink)) list_del_init(&epi->fllink); |
684999149 Rename struct fil... |
579 |
spin_unlock(&file->f_lock); |
1da177e4c Linux-2.6.12-rc2 |
580 |
|
cdac75e6f epoll: avoid kmem... |
581 |
rb_erase(&epi->rbn, &ep->rbr); |
1da177e4c Linux-2.6.12-rc2 |
582 |
|
c7ea76302 epoll locks chang... |
583 584 585 586 |
spin_lock_irqsave(&ep->lock, flags); if (ep_is_linked(&epi->rdllink)) list_del_init(&epi->rdllink); spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
587 |
|
7699acd13 epoll cleanups: e... |
588 |
/* At this point it is safe to free the eventpoll item */ |
c7ea76302 epoll locks chang... |
589 |
kmem_cache_free(epi_cache, epi); |
1da177e4c Linux-2.6.12-rc2 |
590 |
|
52bd19f76 epoll: convert ma... |
591 |
atomic_long_dec(&ep->user->epoll_watches); |
7ef9964e6 epoll: introduce ... |
592 |
|
c7ea76302 epoll locks chang... |
593 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
594 |
} |
7699acd13 epoll cleanups: e... |
595 |
static void ep_free(struct eventpoll *ep) |
1da177e4c Linux-2.6.12-rc2 |
596 |
{ |
7699acd13 epoll cleanups: e... |
597 598 |
struct rb_node *rbp; struct epitem *epi; |
1da177e4c Linux-2.6.12-rc2 |
599 |
|
7699acd13 epoll cleanups: e... |
600 601 |
/* We need to release all tasks waiting for these file */ if (waitqueue_active(&ep->poll_wait)) |
5071f97ec epoll: fix epoll'... |
602 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c Linux-2.6.12-rc2 |
603 |
|
7699acd13 epoll cleanups: e... |
604 605 606 |
/* * We need to lock this because we could be hit by * eventpoll_release_file() while we're freeing the "struct eventpoll". |
d47de16c7 fix epoll single ... |
607 |
* We do not need to hold "ep->mtx" here because the epoll file |
7699acd13 epoll cleanups: e... |
608 609 |
* 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 Fix common misspe... |
610 |
* holding "epmutex" is sufficient here. |
7699acd13 epoll cleanups: e... |
611 612 |
*/ mutex_lock(&epmutex); |
1da177e4c Linux-2.6.12-rc2 |
613 614 |
/* |
7699acd13 epoll cleanups: e... |
615 |
* Walks through the whole tree by unregistering poll callbacks. |
1da177e4c Linux-2.6.12-rc2 |
616 |
*/ |
7699acd13 epoll cleanups: e... |
617 618 619 620 621 |
for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) { epi = rb_entry(rbp, struct epitem, rbn); ep_unregister_pollwait(ep, epi); } |
1da177e4c Linux-2.6.12-rc2 |
622 623 |
/* |
7699acd13 epoll cleanups: e... |
624 625 |
* 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 fix epoll single ... |
626 |
* holding "epmutex" we can be sure that no file cleanup code will hit |
7699acd13 epoll cleanups: e... |
627 |
* us during this operation. So we can avoid the lock on "ep->lock". |
1da177e4c Linux-2.6.12-rc2 |
628 |
*/ |
c80544dc0 sparse pointer us... |
629 |
while ((rbp = rb_first(&ep->rbr)) != NULL) { |
7699acd13 epoll cleanups: e... |
630 631 632 |
epi = rb_entry(rbp, struct epitem, rbn); ep_remove(ep, epi); } |
1da177e4c Linux-2.6.12-rc2 |
633 |
|
7699acd13 epoll cleanups: e... |
634 |
mutex_unlock(&epmutex); |
d47de16c7 fix epoll single ... |
635 |
mutex_destroy(&ep->mtx); |
7ef9964e6 epoll: introduce ... |
636 |
free_uid(ep->user); |
f0ee9aabb epoll: move kfree... |
637 |
kfree(ep); |
7699acd13 epoll cleanups: e... |
638 |
} |
1da177e4c Linux-2.6.12-rc2 |
639 |
|
7699acd13 epoll cleanups: e... |
640 641 642 |
static int ep_eventpoll_release(struct inode *inode, struct file *file) { struct eventpoll *ep = file->private_data; |
1da177e4c Linux-2.6.12-rc2 |
643 |
|
f0ee9aabb epoll: move kfree... |
644 |
if (ep) |
7699acd13 epoll cleanups: e... |
645 |
ep_free(ep); |
7699acd13 epoll cleanups: e... |
646 |
|
7699acd13 epoll cleanups: e... |
647 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
648 |
} |
296e236e9 epoll: fix epoll'... |
649 650 |
static int ep_read_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
5071f97ec epoll: fix epoll'... |
651 652 653 654 655 656 657 |
{ struct epitem *epi, *tmp; list_for_each_entry_safe(epi, tmp, head, rdllink) { if (epi->ffd.file->f_op->poll(epi->ffd.file, NULL) & epi->event.events) return POLLIN | POLLRDNORM; |
296e236e9 epoll: fix epoll'... |
658 |
else { |
5071f97ec epoll: fix epoll'... |
659 660 661 662 663 664 |
/* * 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. */ list_del_init(&epi->rdllink); |
296e236e9 epoll: fix epoll'... |
665 |
} |
5071f97ec epoll: fix epoll'... |
666 667 668 669 670 671 672 |
} return 0; } static int ep_poll_readyevents_proc(void *priv, void *cookie, int call_nests) { |
d8805e633 epoll: fix spurio... |
673 |
return ep_scan_ready_list(priv, ep_read_events_proc, NULL, call_nests + 1); |
5071f97ec epoll: fix epoll'... |
674 |
} |
7699acd13 epoll cleanups: e... |
675 676 |
static unsigned int ep_eventpoll_poll(struct file *file, poll_table *wait) { |
5071f97ec epoll: fix epoll'... |
677 |
int pollflags; |
7699acd13 epoll cleanups: e... |
678 |
struct eventpoll *ep = file->private_data; |
1da177e4c Linux-2.6.12-rc2 |
679 |
|
7699acd13 epoll cleanups: e... |
680 681 |
/* Insert inside our poll wait queue */ poll_wait(file, &ep->poll_wait, wait); |
5071f97ec epoll: fix epoll'... |
682 683 684 685 686 687 688 |
/* * 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, |
3fe4a975d epoll: fix nested... |
689 |
ep_poll_readyevents_proc, ep, ep, current); |
7699acd13 epoll cleanups: e... |
690 |
|
296e236e9 epoll: fix epoll'... |
691 |
return pollflags != -1 ? pollflags : 0; |
7699acd13 epoll cleanups: e... |
692 693 694 695 696 |
} /* File callbacks that implement the eventpoll file behaviour */ static const struct file_operations eventpoll_fops = { .release = ep_eventpoll_release, |
6038f373a llseek: automatic... |
697 698 |
.poll = ep_eventpoll_poll, .llseek = noop_llseek, |
7699acd13 epoll cleanups: e... |
699 |
}; |
b611967de [PATCH] epoll_pwa... |
700 |
/* |
7699acd13 epoll cleanups: e... |
701 702 703 |
* 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 [PATCH] epoll_pwa... |
704 |
*/ |
7699acd13 epoll cleanups: e... |
705 |
void eventpoll_release_file(struct file *file) |
b611967de [PATCH] epoll_pwa... |
706 |
{ |
7699acd13 epoll cleanups: e... |
707 708 709 |
struct list_head *lsthead = &file->f_ep_links; struct eventpoll *ep; struct epitem *epi; |
b611967de [PATCH] epoll_pwa... |
710 711 |
/* |
684999149 Rename struct fil... |
712 |
* We don't want to get "file->f_lock" because it is not |
7699acd13 epoll cleanups: e... |
713 |
* necessary. It is not necessary because we're in the "struct file" |
25985edce Fix common misspe... |
714 |
* cleanup path, and this means that no one is using this file anymore. |
5071f97ec epoll: fix epoll'... |
715 |
* So, for example, epoll_ctl() cannot hit here since if we reach this |
67647d0fb epoll: fix some c... |
716 |
* point, the file counter already went to zero and fget() would fail. |
d47de16c7 fix epoll single ... |
717 |
* The only hit might come from ep_free() but by holding the mutex |
7699acd13 epoll cleanups: e... |
718 |
* will correctly serialize the operation. We do need to acquire |
d47de16c7 fix epoll single ... |
719 |
* "ep->mtx" after "epmutex" because ep_remove() requires it when called |
7699acd13 epoll cleanups: e... |
720 |
* from anywhere but ep_free(). |
684999149 Rename struct fil... |
721 722 |
* * Besides, ep_remove() acquires the lock, so we can't hold it here. |
b611967de [PATCH] epoll_pwa... |
723 |
*/ |
7699acd13 epoll cleanups: e... |
724 |
mutex_lock(&epmutex); |
b611967de [PATCH] epoll_pwa... |
725 |
|
7699acd13 epoll cleanups: e... |
726 727 |
while (!list_empty(lsthead)) { epi = list_first_entry(lsthead, struct epitem, fllink); |
b611967de [PATCH] epoll_pwa... |
728 |
|
7699acd13 epoll cleanups: e... |
729 730 |
ep = epi->ep; list_del_init(&epi->fllink); |
d8805e633 epoll: fix spurio... |
731 |
mutex_lock_nested(&ep->mtx, 0); |
7699acd13 epoll cleanups: e... |
732 |
ep_remove(ep, epi); |
d47de16c7 fix epoll single ... |
733 |
mutex_unlock(&ep->mtx); |
b611967de [PATCH] epoll_pwa... |
734 |
} |
7699acd13 epoll cleanups: e... |
735 |
mutex_unlock(&epmutex); |
b611967de [PATCH] epoll_pwa... |
736 |
} |
53d2be79d [PATCH] epoll: fi... |
737 |
static int ep_alloc(struct eventpoll **pep) |
1da177e4c Linux-2.6.12-rc2 |
738 |
{ |
7ef9964e6 epoll: introduce ... |
739 740 741 |
int error; struct user_struct *user; struct eventpoll *ep; |
1da177e4c Linux-2.6.12-rc2 |
742 |
|
7ef9964e6 epoll: introduce ... |
743 |
user = get_current_user(); |
7ef9964e6 epoll: introduce ... |
744 745 746 747 |
error = -ENOMEM; ep = kzalloc(sizeof(*ep), GFP_KERNEL); if (unlikely(!ep)) goto free_uid; |
1da177e4c Linux-2.6.12-rc2 |
748 |
|
c7ea76302 epoll locks chang... |
749 |
spin_lock_init(&ep->lock); |
d47de16c7 fix epoll single ... |
750 |
mutex_init(&ep->mtx); |
1da177e4c Linux-2.6.12-rc2 |
751 752 753 754 |
init_waitqueue_head(&ep->wq); init_waitqueue_head(&ep->poll_wait); INIT_LIST_HEAD(&ep->rdllist); ep->rbr = RB_ROOT; |
d47de16c7 fix epoll single ... |
755 |
ep->ovflist = EP_UNACTIVE_PTR; |
7ef9964e6 epoll: introduce ... |
756 |
ep->user = user; |
1da177e4c Linux-2.6.12-rc2 |
757 |
|
53d2be79d [PATCH] epoll: fi... |
758 |
*pep = ep; |
1da177e4c Linux-2.6.12-rc2 |
759 |
|
1da177e4c Linux-2.6.12-rc2 |
760 |
return 0; |
7ef9964e6 epoll: introduce ... |
761 762 763 764 |
free_uid: free_uid(user); return error; |
1da177e4c Linux-2.6.12-rc2 |
765 |
} |
1da177e4c Linux-2.6.12-rc2 |
766 |
/* |
c7ea76302 epoll locks chang... |
767 768 769 |
* 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 Linux-2.6.12-rc2 |
770 771 772 773 |
*/ static struct epitem *ep_find(struct eventpoll *ep, struct file *file, int fd) { int kcmp; |
1da177e4c Linux-2.6.12-rc2 |
774 775 776 |
struct rb_node *rbp; struct epitem *epi, *epir = NULL; struct epoll_filefd ffd; |
b030a4dd6 [PATCH] Remove ev... |
777 |
ep_set_ffd(&ffd, file, fd); |
1da177e4c Linux-2.6.12-rc2 |
778 779 |
for (rbp = ep->rbr.rb_node; rbp; ) { epi = rb_entry(rbp, struct epitem, rbn); |
b030a4dd6 [PATCH] Remove ev... |
780 |
kcmp = ep_cmp_ffd(&ffd, &epi->ffd); |
1da177e4c Linux-2.6.12-rc2 |
781 782 783 784 785 |
if (kcmp > 0) rbp = rbp->rb_right; else if (kcmp < 0) rbp = rbp->rb_left; else { |
1da177e4c Linux-2.6.12-rc2 |
786 787 788 789 |
epir = epi; break; } } |
1da177e4c Linux-2.6.12-rc2 |
790 |
|
1da177e4c Linux-2.6.12-rc2 |
791 792 |
return epir; } |
1da177e4c Linux-2.6.12-rc2 |
793 |
/* |
7699acd13 epoll cleanups: e... |
794 |
* This is the callback that is passed to the wait queue wakeup |
bf6a41db7 fs/eventpoll.c: f... |
795 |
* mechanism. It is called by the stored file descriptors when they |
7699acd13 epoll cleanups: e... |
796 |
* have events to report. |
1da177e4c Linux-2.6.12-rc2 |
797 |
*/ |
7699acd13 epoll cleanups: e... |
798 |
static int ep_poll_callback(wait_queue_t *wait, unsigned mode, int sync, void *key) |
1da177e4c Linux-2.6.12-rc2 |
799 |
{ |
7699acd13 epoll cleanups: e... |
800 801 802 803 |
int pwake = 0; unsigned long flags; struct epitem *epi = ep_item_from_wait(wait); struct eventpoll *ep = epi->ep; |
1da177e4c Linux-2.6.12-rc2 |
804 |
|
c7ea76302 epoll locks chang... |
805 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
806 |
|
7699acd13 epoll cleanups: e... |
807 808 809 810 811 812 813 |
/* * 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 fix epoll single ... |
814 815 816 |
goto out_unlock; /* |
2dfa4eeab epoll keyed wakeu... |
817 818 819 820 821 822 823 824 825 |
* 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 fs/eventpoll.c: f... |
826 |
* If we are transferring events to userspace, we can hold no locks |
d47de16c7 fix epoll single ... |
827 |
* (because we're accessing user memory, and because of linux f_op->poll() |
bf6a41db7 fs/eventpoll.c: f... |
828 |
* semantics). All the events that happen during that period of time are |
d47de16c7 fix epoll single ... |
829 830 831 832 833 834 835 836 837 |
* 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; } goto out_unlock; } |
1da177e4c Linux-2.6.12-rc2 |
838 |
|
7699acd13 epoll cleanups: e... |
839 |
/* If this file is already in the ready list we exit soon */ |
5071f97ec epoll: fix epoll'... |
840 841 |
if (!ep_is_linked(&epi->rdllink)) list_add_tail(&epi->rdllink, &ep->rdllist); |
7699acd13 epoll cleanups: e... |
842 |
|
7699acd13 epoll cleanups: e... |
843 844 845 846 847 |
/* * Wake up ( if active ) both the eventpoll wait list and the ->poll() * wait list. */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce Use wake_up_locke... |
848 |
wake_up_locked(&ep->wq); |
7699acd13 epoll cleanups: e... |
849 850 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
d47de16c7 fix epoll single ... |
851 |
out_unlock: |
c7ea76302 epoll locks chang... |
852 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
853 |
|
7699acd13 epoll cleanups: e... |
854 855 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec epoll: fix epoll'... |
856 |
ep_poll_safewake(&ep->poll_wait); |
7699acd13 epoll cleanups: e... |
857 858 859 |
return 1; } |
1da177e4c Linux-2.6.12-rc2 |
860 861 862 863 864 865 866 867 |
/* * 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 [PATCH] Remove ev... |
868 |
struct epitem *epi = ep_item_from_epqueue(pt); |
1da177e4c Linux-2.6.12-rc2 |
869 |
struct eppoll_entry *pwq; |
e94b17660 [PATCH] slab: rem... |
870 |
if (epi->nwait >= 0 && (pwq = kmem_cache_alloc(pwq_cache, GFP_KERNEL))) { |
1da177e4c Linux-2.6.12-rc2 |
871 872 873 874 875 876 |
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 epoll: fix epoll'... |
877 |
} else { |
1da177e4c Linux-2.6.12-rc2 |
878 879 |
/* We have to signal that an error occurred */ epi->nwait = -1; |
296e236e9 epoll: fix epoll'... |
880 |
} |
1da177e4c Linux-2.6.12-rc2 |
881 |
} |
1da177e4c Linux-2.6.12-rc2 |
882 883 884 885 886 887 888 889 890 |
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 [PATCH] Remove ev... |
891 |
kcmp = ep_cmp_ffd(&epi->ffd, &epic->ffd); |
1da177e4c Linux-2.6.12-rc2 |
892 893 894 895 896 897 898 899 |
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); } |
28d82dc1c epoll: limit paths |
900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 |
#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) { 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; list_for_each_entry(epi, &file->f_ep_links, fllink) { 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! "); } } 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) { int length = 0; 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) { length++; 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; } |
c7ea76302 epoll locks chang... |
993 994 995 |
/* * Must be called with "mtx" held. */ |
1da177e4c Linux-2.6.12-rc2 |
996 997 998 999 1000 |
static int ep_insert(struct eventpoll *ep, struct epoll_event *event, struct file *tfile, int fd) { int error, revents, pwake = 0; unsigned long flags; |
52bd19f76 epoll: convert ma... |
1001 |
long user_watches; |
1da177e4c Linux-2.6.12-rc2 |
1002 1003 |
struct epitem *epi; struct ep_pqueue epq; |
52bd19f76 epoll: convert ma... |
1004 1005 |
user_watches = atomic_long_read(&ep->user->epoll_watches); if (unlikely(user_watches >= max_user_watches)) |
7ef9964e6 epoll: introduce ... |
1006 |
return -ENOSPC; |
e94b17660 [PATCH] slab: rem... |
1007 |
if (!(epi = kmem_cache_alloc(epi_cache, GFP_KERNEL))) |
7ef9964e6 epoll: introduce ... |
1008 |
return -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
1009 1010 |
/* Item initialization follow here ... */ |
1da177e4c Linux-2.6.12-rc2 |
1011 1012 |
INIT_LIST_HEAD(&epi->rdllink); INIT_LIST_HEAD(&epi->fllink); |
1da177e4c Linux-2.6.12-rc2 |
1013 1014 |
INIT_LIST_HEAD(&epi->pwqlist); epi->ep = ep; |
b030a4dd6 [PATCH] Remove ev... |
1015 |
ep_set_ffd(&epi->ffd, tfile, fd); |
1da177e4c Linux-2.6.12-rc2 |
1016 |
epi->event = *event; |
1da177e4c Linux-2.6.12-rc2 |
1017 |
epi->nwait = 0; |
d47de16c7 fix epoll single ... |
1018 |
epi->next = EP_UNACTIVE_PTR; |
1da177e4c Linux-2.6.12-rc2 |
1019 1020 1021 1022 1023 1024 1025 1026 |
/* 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 epoll locks chang... |
1027 1028 1029 |
* been increased by the caller of this function. Note that after * this operation completes, the poll callback can start hitting * the new item. |
1da177e4c Linux-2.6.12-rc2 |
1030 1031 1032 1033 1034 1035 1036 1037 |
*/ revents = tfile->f_op->poll(tfile, &epq.pt); /* * 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 epoll: introduce ... |
1038 |
error = -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
1039 |
if (epi->nwait < 0) |
7699acd13 epoll cleanups: e... |
1040 |
goto error_unregister; |
1da177e4c Linux-2.6.12-rc2 |
1041 1042 |
/* Add the current item to the list of active epoll hook for this file */ |
684999149 Rename struct fil... |
1043 |
spin_lock(&tfile->f_lock); |
1da177e4c Linux-2.6.12-rc2 |
1044 |
list_add_tail(&epi->fllink, &tfile->f_ep_links); |
684999149 Rename struct fil... |
1045 |
spin_unlock(&tfile->f_lock); |
1da177e4c Linux-2.6.12-rc2 |
1046 |
|
c7ea76302 epoll locks chang... |
1047 1048 1049 1050 |
/* * 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 Linux-2.6.12-rc2 |
1051 |
ep_rbtree_insert(ep, epi); |
28d82dc1c epoll: limit paths |
1052 1053 1054 1055 |
/* now check if we've created too many backpaths */ error = -EINVAL; if (reverse_path_check()) goto error_remove_epi; |
c7ea76302 epoll locks chang... |
1056 1057 |
/* We have to drop the new item inside our item list to keep track of it */ spin_lock_irqsave(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1058 |
/* If the file is already "ready" we drop it inside the ready list */ |
b030a4dd6 [PATCH] Remove ev... |
1059 |
if ((revents & event->events) && !ep_is_linked(&epi->rdllink)) { |
1da177e4c Linux-2.6.12-rc2 |
1060 1061 1062 1063 |
list_add_tail(&epi->rdllink, &ep->rdllist); /* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce Use wake_up_locke... |
1064 |
wake_up_locked(&ep->wq); |
1da177e4c Linux-2.6.12-rc2 |
1065 1066 1067 |
if (waitqueue_active(&ep->poll_wait)) pwake++; } |
c7ea76302 epoll locks chang... |
1068 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1069 |
|
52bd19f76 epoll: convert ma... |
1070 |
atomic_long_inc(&ep->user->epoll_watches); |
7ef9964e6 epoll: introduce ... |
1071 |
|
1da177e4c Linux-2.6.12-rc2 |
1072 1073 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec epoll: fix epoll'... |
1074 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c Linux-2.6.12-rc2 |
1075 |
|
1da177e4c Linux-2.6.12-rc2 |
1076 |
return 0; |
28d82dc1c epoll: limit paths |
1077 1078 1079 1080 1081 1082 1083 |
error_remove_epi: spin_lock(&tfile->f_lock); if (ep_is_linked(&epi->fllink)) list_del_init(&epi->fllink); spin_unlock(&tfile->f_lock); rb_erase(&epi->rbn, &ep->rbr); |
7699acd13 epoll cleanups: e... |
1084 |
error_unregister: |
1da177e4c Linux-2.6.12-rc2 |
1085 1086 1087 1088 |
ep_unregister_pollwait(ep, epi); /* * We need to do this because an event could have been arrived on some |
67647d0fb epoll: fix some c... |
1089 1090 1091 |
* 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 Linux-2.6.12-rc2 |
1092 |
*/ |
c7ea76302 epoll locks chang... |
1093 |
spin_lock_irqsave(&ep->lock, flags); |
b030a4dd6 [PATCH] Remove ev... |
1094 |
if (ep_is_linked(&epi->rdllink)) |
6192bd536 epoll: optimizati... |
1095 |
list_del_init(&epi->rdllink); |
c7ea76302 epoll locks chang... |
1096 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1097 |
|
b030a4dd6 [PATCH] Remove ev... |
1098 |
kmem_cache_free(epi_cache, epi); |
7ef9964e6 epoll: introduce ... |
1099 |
|
1da177e4c Linux-2.6.12-rc2 |
1100 1101 |
return error; } |
1da177e4c Linux-2.6.12-rc2 |
1102 1103 |
/* * Modify the interest event mask by dropping an event if the new mask |
c7ea76302 epoll locks chang... |
1104 |
* has a match in the current file status. Must be called with "mtx" held. |
1da177e4c Linux-2.6.12-rc2 |
1105 1106 1107 1108 1109 |
*/ static int ep_modify(struct eventpoll *ep, struct epitem *epi, struct epoll_event *event) { int pwake = 0; unsigned int revents; |
1da177e4c Linux-2.6.12-rc2 |
1110 1111 |
/* |
e057e15ff epoll: clean up e... |
1112 1113 1114 |
* 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 Linux-2.6.12-rc2 |
1115 1116 |
*/ epi->event.events = event->events; |
e057e15ff epoll: clean up e... |
1117 |
epi->event.data = event->data; /* protected by mtx */ |
1da177e4c Linux-2.6.12-rc2 |
1118 1119 1120 1121 1122 1123 |
/* * Get current event bits. We can safely use the file* here because * its usage count has been increased by the caller of this function. */ revents = epi->ffd.file->f_op->poll(epi->ffd.file, NULL); |
1da177e4c Linux-2.6.12-rc2 |
1124 |
/* |
c7ea76302 epoll locks chang... |
1125 |
* If the item is "hot" and it is not registered inside the ready |
67647d0fb epoll: fix some c... |
1126 |
* list, push it inside. |
1da177e4c Linux-2.6.12-rc2 |
1127 |
*/ |
c7ea76302 epoll locks chang... |
1128 |
if (revents & event->events) { |
e057e15ff epoll: clean up e... |
1129 |
spin_lock_irq(&ep->lock); |
c7ea76302 epoll locks chang... |
1130 1131 1132 1133 1134 |
if (!ep_is_linked(&epi->rdllink)) { list_add_tail(&epi->rdllink, &ep->rdllist); /* Notify waiting tasks that events are available */ if (waitqueue_active(&ep->wq)) |
4a6e9e2ce Use wake_up_locke... |
1135 |
wake_up_locked(&ep->wq); |
c7ea76302 epoll locks chang... |
1136 1137 |
if (waitqueue_active(&ep->poll_wait)) pwake++; |
7699acd13 epoll cleanups: e... |
1138 |
} |
e057e15ff epoll: clean up e... |
1139 |
spin_unlock_irq(&ep->lock); |
7699acd13 epoll cleanups: e... |
1140 |
} |
1da177e4c Linux-2.6.12-rc2 |
1141 |
|
7699acd13 epoll cleanups: e... |
1142 1143 |
/* We have to call this outside the lock */ if (pwake) |
5071f97ec epoll: fix epoll'... |
1144 |
ep_poll_safewake(&ep->poll_wait); |
1da177e4c Linux-2.6.12-rc2 |
1145 |
|
7699acd13 epoll cleanups: e... |
1146 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1147 |
} |
296e236e9 epoll: fix epoll'... |
1148 1149 |
static int ep_send_events_proc(struct eventpoll *ep, struct list_head *head, void *priv) |
1da177e4c Linux-2.6.12-rc2 |
1150 |
{ |
5071f97ec epoll: fix epoll'... |
1151 1152 |
struct ep_send_events_data *esed = priv; int eventcnt; |
296e236e9 epoll: fix epoll'... |
1153 |
unsigned int revents; |
5071f97ec epoll: fix epoll'... |
1154 1155 |
struct epitem *epi; struct epoll_event __user *uevent; |
1da177e4c Linux-2.6.12-rc2 |
1156 |
|
296e236e9 epoll: fix epoll'... |
1157 |
/* |
5071f97ec epoll: fix epoll'... |
1158 1159 1160 |
* 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 epoll: fix epoll'... |
1161 |
*/ |
5071f97ec epoll: fix epoll'... |
1162 1163 1164 |
for (eventcnt = 0, uevent = esed->events; !list_empty(head) && eventcnt < esed->maxevents;) { epi = list_first_entry(head, struct epitem, rdllink); |
d47de16c7 fix epoll single ... |
1165 1166 |
list_del_init(&epi->rdllink); |
1da177e4c Linux-2.6.12-rc2 |
1167 |
|
296e236e9 epoll: fix epoll'... |
1168 1169 |
revents = epi->ffd.file->f_op->poll(epi->ffd.file, NULL) & epi->event.events; |
5071f97ec epoll: fix epoll'... |
1170 |
|
296e236e9 epoll: fix epoll'... |
1171 |
/* |
5071f97ec epoll: fix epoll'... |
1172 1173 1174 1175 |
* 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 epoll: fix epoll'... |
1176 1177 |
*/ if (revents) { |
5071f97ec epoll: fix epoll'... |
1178 |
if (__put_user(revents, &uevent->events) || |
d03058828 epoll: remember t... |
1179 1180 |
__put_user(epi->event.data, &uevent->data)) { list_add(&epi->rdllink, head); |
296e236e9 epoll: fix epoll'... |
1181 |
return eventcnt ? eventcnt : -EFAULT; |
d03058828 epoll: remember t... |
1182 |
} |
296e236e9 epoll: fix epoll'... |
1183 |
eventcnt++; |
5071f97ec epoll: fix epoll'... |
1184 |
uevent++; |
296e236e9 epoll: fix epoll'... |
1185 1186 1187 1188 1189 1190 1191 1192 |
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 Fix common misspe... |
1193 |
* availability. At this point, no one can insert |
296e236e9 epoll: fix epoll'... |
1194 1195 1196 1197 1198 1199 1200 1201 1202 |
* 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); } } } |
5071f97ec epoll: fix epoll'... |
1203 1204 1205 |
return eventcnt; } |
d47de16c7 fix epoll single ... |
1206 |
|
296e236e9 epoll: fix epoll'... |
1207 1208 |
static int ep_send_events(struct eventpoll *ep, struct epoll_event __user *events, int maxevents) |
5071f97ec epoll: fix epoll'... |
1209 1210 |
{ struct ep_send_events_data esed; |
1da177e4c Linux-2.6.12-rc2 |
1211 |
|
5071f97ec epoll: fix epoll'... |
1212 1213 |
esed.maxevents = maxevents; esed.events = events; |
6192bd536 epoll: optimizati... |
1214 |
|
d8805e633 epoll: fix spurio... |
1215 |
return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0); |
1da177e4c Linux-2.6.12-rc2 |
1216 |
} |
0781b909b epoll: epoll_wait... |
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 |
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 epoll: fix compil... |
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 |
/** * 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 Linux-2.6.12-rc2 |
1244 1245 1246 |
static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events, int maxevents, long timeout) { |
f4d93ad74 epoll: fix compil... |
1247 |
int res = 0, eavail, timed_out = 0; |
1da177e4c Linux-2.6.12-rc2 |
1248 |
unsigned long flags; |
f4d93ad74 epoll: fix compil... |
1249 |
long slack = 0; |
1da177e4c Linux-2.6.12-rc2 |
1250 |
wait_queue_t wait; |
95aac7b1c epoll: make epoll... |
1251 1252 1253 |
ktime_t expires, *to = NULL; if (timeout > 0) { |
0781b909b epoll: epoll_wait... |
1254 |
struct timespec end_time = ep_set_mstimeout(timeout); |
95aac7b1c epoll: make epoll... |
1255 1256 1257 1258 |
slack = select_estimate_accuracy(&end_time); to = &expires; *to = timespec_to_ktime(end_time); } else if (timeout == 0) { |
f4d93ad74 epoll: fix compil... |
1259 1260 1261 1262 |
/* * Avoid the unnecessary trip to the wait queue loop, if the * caller specified a non blocking operation. */ |
95aac7b1c epoll: make epoll... |
1263 |
timed_out = 1; |
f4d93ad74 epoll: fix compil... |
1264 1265 |
spin_lock_irqsave(&ep->lock, flags); goto check_events; |
95aac7b1c epoll: make epoll... |
1266 |
} |
1da177e4c Linux-2.6.12-rc2 |
1267 |
|
f4d93ad74 epoll: fix compil... |
1268 |
fetch_events: |
c7ea76302 epoll locks chang... |
1269 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1270 |
|
3fb0e584a epoll: move ready... |
1271 |
if (!ep_events_available(ep)) { |
1da177e4c Linux-2.6.12-rc2 |
1272 1273 1274 1275 1276 1277 |
/* * 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 sched, wait: Use ... |
1278 |
__add_wait_queue_exclusive(&ep->wq, &wait); |
1da177e4c Linux-2.6.12-rc2 |
1279 1280 1281 1282 1283 1284 1285 1286 |
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 epoll: move ready... |
1287 |
if (ep_events_available(ep) || timed_out) |
1da177e4c Linux-2.6.12-rc2 |
1288 1289 1290 1291 1292 |
break; if (signal_pending(current)) { res = -EINTR; break; } |
c7ea76302 epoll locks chang... |
1293 |
spin_unlock_irqrestore(&ep->lock, flags); |
95aac7b1c epoll: make epoll... |
1294 1295 |
if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS)) timed_out = 1; |
c7ea76302 epoll locks chang... |
1296 |
spin_lock_irqsave(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1297 |
} |
3419b23a9 [PATCH] epoll: us... |
1298 |
__remove_wait_queue(&ep->wq, &wait); |
1da177e4c Linux-2.6.12-rc2 |
1299 1300 1301 |
set_current_state(TASK_RUNNING); } |
f4d93ad74 epoll: fix compil... |
1302 |
check_events: |
1da177e4c Linux-2.6.12-rc2 |
1303 |
/* Is it worth to try to dig for events ? */ |
3fb0e584a epoll: move ready... |
1304 |
eavail = ep_events_available(ep); |
1da177e4c Linux-2.6.12-rc2 |
1305 |
|
c7ea76302 epoll locks chang... |
1306 |
spin_unlock_irqrestore(&ep->lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1307 1308 1309 1310 1311 1312 1313 |
/* * 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 epoll: make epoll... |
1314 |
!(res = ep_send_events(ep, events, maxevents)) && !timed_out) |
f4d93ad74 epoll: fix compil... |
1315 |
goto fetch_events; |
1da177e4c Linux-2.6.12-rc2 |
1316 1317 1318 |
return res; } |
22bacca48 epoll: prevent cr... |
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 |
/** * 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 epoll: limit paths |
1339 |
struct eventpoll *ep_tovisit; |
22bacca48 epoll: prevent cr... |
1340 1341 |
struct rb_node *rbp; struct epitem *epi; |
d8805e633 epoll: fix spurio... |
1342 |
mutex_lock_nested(&ep->mtx, call_nests + 1); |
28d82dc1c epoll: limit paths |
1343 1344 |
ep->visited = 1; list_add(&ep->visited_list_link, &visited_list); |
22bacca48 epoll: prevent cr... |
1345 1346 1347 |
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 epoll: limit paths |
1348 1349 1350 |
ep_tovisit = epi->ffd.file->private_data; if (ep_tovisit->visited) continue; |
22bacca48 epoll: prevent cr... |
1351 |
error = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
28d82dc1c epoll: limit paths |
1352 1353 |
ep_loop_check_proc, epi->ffd.file, ep_tovisit, current); |
22bacca48 epoll: prevent cr... |
1354 1355 |
if (error != 0) break; |
28d82dc1c epoll: limit paths |
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 |
} 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 epoll: prevent cr... |
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 |
} } 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 epoll: limit paths |
1388 1389 1390 1391 |
int ret; struct eventpoll *ep_cur, *ep_next; ret = ep_call_nested(&poll_loop_ncalls, EP_MAX_NESTS, |
22bacca48 epoll: prevent cr... |
1392 |
ep_loop_check_proc, file, ep, current); |
28d82dc1c epoll: limit paths |
1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 |
/* 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 epoll: prevent cr... |
1413 |
} |
7699acd13 epoll cleanups: e... |
1414 |
/* |
523723bb5 fs/eventpoll.c: f... |
1415 |
* Open an eventpoll file descriptor. |
7699acd13 epoll cleanups: e... |
1416 |
*/ |
5a8a82b1d [CVE-2009-0029] S... |
1417 |
SYSCALL_DEFINE1(epoll_create1, int, flags) |
7699acd13 epoll cleanups: e... |
1418 |
{ |
28d82dc1c epoll: limit paths |
1419 |
int error, fd; |
bb57c3edc epoll: remove deb... |
1420 |
struct eventpoll *ep = NULL; |
28d82dc1c epoll: limit paths |
1421 |
struct file *file; |
7699acd13 epoll cleanups: e... |
1422 |
|
e38b36f32 flag parameters: ... |
1423 1424 |
/* Check the EPOLL_* constant for consistency. */ BUILD_BUG_ON(EPOLL_CLOEXEC != O_CLOEXEC); |
296e236e9 epoll: fix epoll'... |
1425 1426 |
if (flags & ~EPOLL_CLOEXEC) return -EINVAL; |
7699acd13 epoll cleanups: e... |
1427 |
/* |
bb57c3edc epoll: remove deb... |
1428 |
* Create the internal data structure ("struct eventpoll"). |
7699acd13 epoll cleanups: e... |
1429 |
*/ |
9fe5ad9c8 flag parameters a... |
1430 |
error = ep_alloc(&ep); |
bb57c3edc epoll: remove deb... |
1431 1432 |
if (error < 0) return error; |
7699acd13 epoll cleanups: e... |
1433 1434 |
/* * Creates all the items needed to setup an eventpoll file. That is, |
2030a42ce [PATCH] sanitize ... |
1435 |
* a file structure and a free file descriptor. |
7699acd13 epoll cleanups: e... |
1436 |
*/ |
28d82dc1c epoll: limit paths |
1437 1438 1439 1440 1441 1442 |
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 anonfd: Allow mak... |
1443 |
O_RDWR | (flags & O_CLOEXEC)); |
28d82dc1c epoll: limit paths |
1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 |
if (IS_ERR(file)) { error = PTR_ERR(file); goto out_free_fd; } fd_install(fd, file); ep->file = file; return fd; out_free_fd: put_unused_fd(fd); out_free_ep: ep_free(ep); |
bb57c3edc epoll: remove deb... |
1456 |
return error; |
7699acd13 epoll cleanups: e... |
1457 |
} |
5a8a82b1d [CVE-2009-0029] S... |
1458 |
SYSCALL_DEFINE1(epoll_create, int, size) |
a0998b50c flag parameters: ... |
1459 |
{ |
bfe3891a5 epoll: fix size c... |
1460 |
if (size <= 0) |
9fe5ad9c8 flag parameters a... |
1461 1462 1463 |
return -EINVAL; return sys_epoll_create1(0); |
a0998b50c flag parameters: ... |
1464 |
} |
7699acd13 epoll cleanups: e... |
1465 1466 1467 |
/* * The following function implements the controller interface for * the eventpoll file that enables the insertion/removal/change of |
67647d0fb epoll: fix some c... |
1468 |
* file descriptors inside the interest set. |
7699acd13 epoll cleanups: e... |
1469 |
*/ |
5a8a82b1d [CVE-2009-0029] S... |
1470 1471 |
SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd, struct epoll_event __user *, event) |
7699acd13 epoll cleanups: e... |
1472 1473 |
{ int error; |
22bacca48 epoll: prevent cr... |
1474 |
int did_lock_epmutex = 0; |
7699acd13 epoll cleanups: e... |
1475 1476 1477 1478 |
struct file *file, *tfile; struct eventpoll *ep; struct epitem *epi; struct epoll_event epds; |
7699acd13 epoll cleanups: e... |
1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 |
error = -EFAULT; if (ep_op_has_event(op) && copy_from_user(&epds, event, sizeof(struct epoll_event))) goto error_return; /* Get the "struct file *" for the eventpoll file */ error = -EBADF; file = fget(epfd); if (!file) goto error_return; /* Get the "struct file *" for the target file */ tfile = fget(fd); if (!tfile) goto error_fput; /* The target file descriptor must support poll */ error = -EPERM; if (!tfile->f_op || !tfile->f_op->poll) goto error_tgt_fput; /* * 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; if (file == tfile || !is_file_epoll(file)) goto error_tgt_fput; /* * At this point it is safe to assume that the "private_data" contains * our own data structure. */ ep = file->private_data; |
22bacca48 epoll: prevent cr... |
1514 1515 1516 |
/* * When we insert an epoll file descriptor, inside another epoll file * descriptor, there is the change of creating closed loops, which are |
28d82dc1c epoll: limit paths |
1517 1518 1519 1520 |
* 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 epoll: prevent cr... |
1521 |
* |
28d82dc1c epoll: limit paths |
1522 1523 1524 |
* We need to hold the epmutex across both ep_insert and ep_remove * b/c we want to make sure we are looking at a coherent view of * epoll network. |
22bacca48 epoll: prevent cr... |
1525 |
*/ |
28d82dc1c epoll: limit paths |
1526 |
if (op == EPOLL_CTL_ADD || op == EPOLL_CTL_DEL) { |
22bacca48 epoll: prevent cr... |
1527 1528 |
mutex_lock(&epmutex); did_lock_epmutex = 1; |
22bacca48 epoll: prevent cr... |
1529 |
} |
28d82dc1c epoll: limit paths |
1530 1531 1532 1533 1534 1535 1536 1537 |
if (op == EPOLL_CTL_ADD) { if (is_file_epoll(tfile)) { error = -ELOOP; if (ep_loop_check(ep, tfile) != 0) goto error_tgt_fput; } else list_add(&tfile->f_tfile_llink, &tfile_check_list); } |
22bacca48 epoll: prevent cr... |
1538 |
|
d8805e633 epoll: fix spurio... |
1539 |
mutex_lock_nested(&ep->mtx, 0); |
7699acd13 epoll cleanups: e... |
1540 |
|
67647d0fb epoll: fix some c... |
1541 1542 1543 1544 1545 |
/* * 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. */ |
7699acd13 epoll cleanups: e... |
1546 1547 1548 1549 1550 1551 1552 |
epi = ep_find(ep, tfile, fd); error = -EINVAL; switch (op) { case EPOLL_CTL_ADD: if (!epi) { epds.events |= POLLERR | POLLHUP; |
7699acd13 epoll cleanups: e... |
1553 1554 1555 |
error = ep_insert(ep, &epds, tfile, fd); } else error = -EEXIST; |
28d82dc1c epoll: limit paths |
1556 |
clear_tfile_check_list(); |
7699acd13 epoll cleanups: e... |
1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 |
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; } |
d47de16c7 fix epoll single ... |
1572 |
mutex_unlock(&ep->mtx); |
7699acd13 epoll cleanups: e... |
1573 1574 |
error_tgt_fput: |
28d82dc1c epoll: limit paths |
1575 |
if (did_lock_epmutex) |
22bacca48 epoll: prevent cr... |
1576 |
mutex_unlock(&epmutex); |
7699acd13 epoll cleanups: e... |
1577 1578 1579 1580 |
fput(tfile); error_fput: fput(file); error_return: |
7699acd13 epoll cleanups: e... |
1581 1582 1583 1584 1585 1586 1587 1588 |
return error; } /* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_wait(2). */ |
5a8a82b1d [CVE-2009-0029] S... |
1589 1590 |
SYSCALL_DEFINE4(epoll_wait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout) |
7699acd13 epoll cleanups: e... |
1591 1592 1593 1594 |
{ int error; struct file *file; struct eventpoll *ep; |
7699acd13 epoll cleanups: e... |
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 |
/* 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 */ if (!access_ok(VERIFY_WRITE, events, maxevents * sizeof(struct epoll_event))) { error = -EFAULT; goto error_return; } /* Get the "struct file *" for the eventpoll file */ error = -EBADF; file = fget(epfd); if (!file) goto error_return; /* * We have to check that the file structure underneath the fd * the user passed to us _is_ an eventpoll file. */ error = -EINVAL; if (!is_file_epoll(file)) goto error_fput; /* * At this point it is safe to assume that the "private_data" contains * our own data structure. */ ep = file->private_data; /* Time to fish for events ... */ error = ep_poll(ep, events, maxevents, timeout); error_fput: fput(file); error_return: |
7699acd13 epoll cleanups: e... |
1631 1632 1633 |
return error; } |
f3de272b8 signals: use HAVE... |
1634 |
#ifdef HAVE_SET_RESTORE_SIGMASK |
7699acd13 epoll cleanups: e... |
1635 1636 1637 1638 1639 |
/* * Implement the event wait interface for the eventpoll file. It is the kernel * part of the user space epoll_pwait(2). */ |
5a8a82b1d [CVE-2009-0029] S... |
1640 1641 1642 |
SYSCALL_DEFINE6(epoll_pwait, int, epfd, struct epoll_event __user *, events, int, maxevents, int, timeout, const sigset_t __user *, sigmask, size_t, sigsetsize) |
7699acd13 epoll cleanups: e... |
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 |
{ 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; sigdelsetmask(&ksigmask, sigmask(SIGKILL) | sigmask(SIGSTOP)); sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved); } 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 epoll locks chang... |
1671 |
sizeof(sigsaved)); |
4e4c22c71 signals: add set_... |
1672 |
set_restore_sigmask(); |
7699acd13 epoll cleanups: e... |
1673 1674 1675 1676 1677 1678 |
} else sigprocmask(SIG_SETMASK, &sigsaved, NULL); } return error; } |
f3de272b8 signals: use HAVE... |
1679 |
#endif /* HAVE_SET_RESTORE_SIGMASK */ |
7699acd13 epoll cleanups: e... |
1680 |
|
1da177e4c Linux-2.6.12-rc2 |
1681 1682 |
static int __init eventpoll_init(void) { |
7ef9964e6 epoll: introduce ... |
1683 1684 1685 |
struct sysinfo si; si_meminfo(&si); |
9df04e1f2 epoll: drop max_u... |
1686 1687 1688 1689 |
/* * Allows top 4% of lomem to be allocated for epoll watches (per user). */ max_user_watches = (((si.totalram - si.totalhigh) / 25) << PAGE_SHIFT) / |
7ef9964e6 epoll: introduce ... |
1690 |
EP_ITEM_COST; |
52bd19f76 epoll: convert ma... |
1691 |
BUG_ON(max_user_watches < 0); |
1da177e4c Linux-2.6.12-rc2 |
1692 |
|
22bacca48 epoll: prevent cr... |
1693 1694 1695 1696 1697 |
/* * Initialize the structure used to perform epoll file descriptor * inclusion loops checks. */ ep_nested_calls_init(&poll_loop_ncalls); |
1da177e4c Linux-2.6.12-rc2 |
1698 |
/* Initialize the structure used to perform safe poll wait head wake ups */ |
5071f97ec epoll: fix epoll'... |
1699 1700 1701 1702 |
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 Linux-2.6.12-rc2 |
1703 1704 1705 |
/* Allocates slab cache used to allocate "struct epitem" items */ epi_cache = kmem_cache_create("eventpoll_epi", sizeof(struct epitem), |
bb57c3edc epoll: remove deb... |
1706 |
0, SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL); |
1da177e4c Linux-2.6.12-rc2 |
1707 1708 1709 |
/* Allocates slab cache used to allocate "struct eppoll_entry" */ pwq_cache = kmem_cache_create("eventpoll_pwq", |
bb57c3edc epoll: remove deb... |
1710 |
sizeof(struct eppoll_entry), 0, SLAB_PANIC, NULL); |
1da177e4c Linux-2.6.12-rc2 |
1711 |
|
1da177e4c Linux-2.6.12-rc2 |
1712 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1713 |
} |
cea692418 epoll cleanups: e... |
1714 |
fs_initcall(eventpoll_init); |