Blame view
kernel/futex.c
113 KB
1a59d1b8e treewide: Replace... |
1 |
// SPDX-License-Identifier: GPL-2.0-or-later |
1da177e4c Linux-2.6.12-rc2 |
2 3 4 5 6 7 8 9 10 11 |
/* * Fast Userspace Mutexes (which I call "Futexes!"). * (C) Rusty Russell, IBM 2002 * * Generalized futexes, futex requeueing, misc fixes by Ingo Molnar * (C) Copyright 2003 Red Hat Inc, All Rights Reserved * * Removed page pinning, fix privately mapped COW pages and other cleanups * (C) Copyright 2003, 2004 Jamie Lokier * |
0771dfefc [PATCH] lightweig... |
12 13 14 15 |
* Robust futex support started by Ingo Molnar * (C) Copyright 2006 Red Hat Inc, All Rights Reserved * Thanks to Thomas Gleixner for suggestions, analysis and fixes. * |
c87e2837b [PATCH] pi-futex:... |
16 17 18 19 |
* PI-futex support started by Ingo Molnar and Thomas Gleixner * Copyright (C) 2006 Red Hat, Inc., Ingo Molnar <mingo@redhat.com> * Copyright (C) 2006 Timesys Corp., Thomas Gleixner <tglx@timesys.com> * |
34f01cc1f FUTEX: new PRIVAT... |
20 21 22 |
* PRIVATE futexes by Eric Dumazet * Copyright (C) 2007 Eric Dumazet <dada1@cosmosbay.com> * |
52400ba94 futex: add requeu... |
23 24 25 26 |
* Requeue-PI support by Darren Hart <dvhltc@us.ibm.com> * Copyright (C) IBM Corporation, 2009 * Thanks to Thomas Gleixner for conceptual design and careful reviews. * |
1da177e4c Linux-2.6.12-rc2 |
27 28 29 30 31 32 |
* Thanks to Ben LaHaise for yelling "hashed waitqueues" loudly * enough at me, Linus for the original (flawed) idea, Matthew * Kirkwood for proof-of-concept implementation. * * "The futexes are also cursed." * "But they come in a choice of three flavours!" |
1da177e4c Linux-2.6.12-rc2 |
33 |
*/ |
04e7712f4 y2038: futex: Mov... |
34 |
#include <linux/compat.h> |
1da177e4c Linux-2.6.12-rc2 |
35 36 37 38 39 40 41 42 43 44 |
#include <linux/slab.h> #include <linux/poll.h> #include <linux/fs.h> #include <linux/file.h> #include <linux/jhash.h> #include <linux/init.h> #include <linux/futex.h> #include <linux/mount.h> #include <linux/pagemap.h> #include <linux/syscalls.h> |
7ed20e1ad [PATCH] convert t... |
45 |
#include <linux/signal.h> |
9984de1a5 kernel: Map most ... |
46 |
#include <linux/export.h> |
fd5eea421 change inotifyfs ... |
47 |
#include <linux/magic.h> |
b488893a3 pid namespaces: c... |
48 49 |
#include <linux/pid.h> #include <linux/nsproxy.h> |
bdbb776f8 futex: Do not lea... |
50 |
#include <linux/ptrace.h> |
8bd75c77b sched/rt: Move rt... |
51 |
#include <linux/sched/rt.h> |
84f001e15 sched/headers: Pr... |
52 |
#include <linux/sched/wake_q.h> |
6e84f3152 sched/headers: Pr... |
53 |
#include <linux/sched/mm.h> |
13d60f4b6 futex: Take hugep... |
54 |
#include <linux/hugetlb.h> |
88c8004fd futex: Use freeza... |
55 |
#include <linux/freezer.h> |
57c8a661d mm: remove includ... |
56 |
#include <linux/memblock.h> |
ab51fbab3 futex: Fault/erro... |
57 |
#include <linux/fault-inject.h> |
49262de22 futex: Convert fu... |
58 |
#include <linux/refcount.h> |
b488893a3 pid namespaces: c... |
59 |
|
4732efbeb [PATCH] FUTEX_WAK... |
60 |
#include <asm/futex.h> |
1da177e4c Linux-2.6.12-rc2 |
61 |
|
1696a8bee locking: Move the... |
62 |
#include "locking/rtmutex_common.h" |
c87e2837b [PATCH] pi-futex:... |
63 |
|
99b60ce69 futexes: Document... |
64 |
/* |
d7e8af1af futex: update doc... |
65 66 67 68 |
* READ this before attempting to hack on futexes! * * Basic futex operation and ordering guarantees * ============================================= |
99b60ce69 futexes: Document... |
69 70 71 72 |
* * The waiter reads the futex value in user space and calls * futex_wait(). This function computes the hash bucket and acquires * the hash bucket lock. After that it reads the futex user space value |
b0c29f79e futexes: Avoid ta... |
73 74 75 |
* again and verifies that the data has not changed. If it has not changed * it enqueues itself into the hash bucket, releases the hash bucket lock * and schedules. |
99b60ce69 futexes: Document... |
76 77 |
* * The waker side modifies the user space value of the futex and calls |
b0c29f79e futexes: Avoid ta... |
78 79 80 |
* futex_wake(). This function computes the hash bucket and acquires the * hash bucket lock. Then it looks for waiters on that futex in the hash * bucket and wakes them. |
99b60ce69 futexes: Document... |
81 |
* |
b0c29f79e futexes: Avoid ta... |
82 83 84 85 86 |
* In futex wake up scenarios where no tasks are blocked on a futex, taking * the hb spinlock can be avoided and simply return. In order for this * optimization to work, ordering guarantees must exist so that the waiter * being added to the list is acknowledged when the list is concurrently being * checked by the waker, avoiding scenarios like the following: |
99b60ce69 futexes: Document... |
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 |
* * CPU 0 CPU 1 * val = *futex; * sys_futex(WAIT, futex, val); * futex_wait(futex, val); * uval = *futex; * *futex = newval; * sys_futex(WAKE, futex); * futex_wake(futex); * if (queue_empty()) * return; * if (uval == val) * lock(hash_bucket(futex)); * queue(); * unlock(hash_bucket(futex)); * schedule(); * * This would cause the waiter on CPU 0 to wait forever because it * missed the transition of the user space value from val to newval * and the waker did not find the waiter in the hash bucket queue. |
99b60ce69 futexes: Document... |
107 |
* |
b0c29f79e futexes: Avoid ta... |
108 109 110 111 112 |
* The correct serialization ensures that a waiter either observes * the changed user space value before blocking or is woken by a * concurrent waker: * * CPU 0 CPU 1 |
99b60ce69 futexes: Document... |
113 114 115 |
* val = *futex; * sys_futex(WAIT, futex, val); * futex_wait(futex, val); |
b0c29f79e futexes: Avoid ta... |
116 |
* |
d7e8af1af futex: update doc... |
117 |
* waiters++; (a) |
8ad7b378d futex: Rename bar... |
118 119 120 121 122 123 124 125 126 127 |
* smp_mb(); (A) <-- paired with -. * | * lock(hash_bucket(futex)); | * | * uval = *futex; | * | *futex = newval; * | sys_futex(WAKE, futex); * | futex_wake(futex); * | * `--------> smp_mb(); (B) |
99b60ce69 futexes: Document... |
128 |
* if (uval == val) |
b0c29f79e futexes: Avoid ta... |
129 |
* queue(); |
99b60ce69 futexes: Document... |
130 |
* unlock(hash_bucket(futex)); |
b0c29f79e futexes: Avoid ta... |
131 132 |
* schedule(); if (waiters) * lock(hash_bucket(futex)); |
d7e8af1af futex: update doc... |
133 134 |
* else wake_waiters(futex); * waiters--; (b) unlock(hash_bucket(futex)); |
b0c29f79e futexes: Avoid ta... |
135 |
* |
d7e8af1af futex: update doc... |
136 137 |
* Where (A) orders the waiters increment and the futex value read through * atomic operations (see hb_waiters_inc) and where (B) orders the write |
993b2ff22 futex: Mention ke... |
138 139 |
* to futex and the waiters read -- this is done by the barriers for both * shared and private futexes in get_futex_key_refs(). |
b0c29f79e futexes: Avoid ta... |
140 141 142 143 144 145 146 147 148 149 150 151 |
* * This yields the following case (where X:=waiters, Y:=futex): * * X = Y = 0 * * w[X]=1 w[Y]=1 * MB MB * r[Y]=y r[X]=x * * Which guarantees that x==0 && y==0 is impossible; which translates back into * the guarantee that we cannot both miss the futex variable change and the * enqueue. |
d7e8af1af futex: update doc... |
152 153 154 155 156 157 158 159 160 161 162 |
* * Note that a new waiter is accounted for in (a) even when it is possible that * the wait call can return error, in which case we backtrack from it in (b). * Refer to the comment in queue_lock(). * * Similarly, in order to account for waiters being requeued on another * address we always increment the waiters for the destination bucket before * acquiring the lock. It then decrements them again after releasing it - * the code that actually moves the futex(es) between hash buckets (requeue_futex) * will do the additional required waiter count housekeeping. This is done for * double_lock_hb() and double_unlock_hb(), respectively. |
99b60ce69 futexes: Document... |
163 |
*/ |
04e7712f4 y2038: futex: Mov... |
164 165 166 167 |
#ifdef CONFIG_HAVE_FUTEX_CMPXCHG #define futex_cmpxchg_enabled 1 #else static int __read_mostly futex_cmpxchg_enabled; |
03b8c7b62 futex: Allow arch... |
168 |
#endif |
a0c1e9073 futex: runtime en... |
169 |
|
1da177e4c Linux-2.6.12-rc2 |
170 |
/* |
b41277dc7 futex: Replace fs... |
171 172 173 |
* Futex flags used to encode options to functions and preserve them across * restarts. */ |
784bdf3bb futex: Assume all... |
174 175 176 177 178 179 180 181 182 |
#ifdef CONFIG_MMU # define FLAGS_SHARED 0x01 #else /* * NOMMU does not have per process address space. Let the compiler optimize * code away. */ # define FLAGS_SHARED 0x00 #endif |
b41277dc7 futex: Replace fs... |
183 184 185 186 |
#define FLAGS_CLOCKRT 0x02 #define FLAGS_HAS_TIMEOUT 0x04 /* |
c87e2837b [PATCH] pi-futex:... |
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 |
* Priority Inheritance state: */ struct futex_pi_state { /* * list of 'owned' pi_state instances - these have to be * cleaned up in do_exit() if the task exits prematurely: */ struct list_head list; /* * The PI object: */ struct rt_mutex pi_mutex; struct task_struct *owner; |
49262de22 futex: Convert fu... |
202 |
refcount_t refcount; |
c87e2837b [PATCH] pi-futex:... |
203 204 |
union futex_key key; |
3859a271a randstruct: Mark ... |
205 |
} __randomize_layout; |
c87e2837b [PATCH] pi-futex:... |
206 |
|
d8d88fbb1 futex: Correct fu... |
207 208 |
/** * struct futex_q - The hashed futex queue entry, one per waiting task |
fb62db2ba futex: Fix kernel... |
209 |
* @list: priority-sorted list of tasks waiting on this futex |
d8d88fbb1 futex: Correct fu... |
210 211 212 213 214 215 216 217 |
* @task: the task waiting on the futex * @lock_ptr: the hash bucket lock * @key: the key the futex is hashed on * @pi_state: optional priority inheritance state * @rt_waiter: rt_waiter storage for use with requeue_pi * @requeue_pi_key: the requeue_pi target futex key * @bitset: bitset for the optional bitmasked wakeup * |
ac6424b98 sched/wait: Renam... |
218 |
* We use this hashed waitqueue, instead of a normal wait_queue_entry_t, so |
1da177e4c Linux-2.6.12-rc2 |
219 220 221 |
* we can wake only the relevant ones (hashed queues may be shared). * * A futex_q has a woken state, just like tasks have TASK_RUNNING. |
ec92d0829 futex priority ba... |
222 |
* It is considered woken when plist_node_empty(&q->list) || q->lock_ptr == 0. |
fb62db2ba futex: Fix kernel... |
223 |
* The order of wakeup is always to make the first condition true, then |
d8d88fbb1 futex: Correct fu... |
224 225 226 227 |
* the second. * * PI futexes are typically woken before they are removed from the hash list via * the rt_mutex code. See unqueue_me_pi(). |
1da177e4c Linux-2.6.12-rc2 |
228 229 |
*/ struct futex_q { |
ec92d0829 futex priority ba... |
230 |
struct plist_node list; |
1da177e4c Linux-2.6.12-rc2 |
231 |
|
d8d88fbb1 futex: Correct fu... |
232 |
struct task_struct *task; |
1da177e4c Linux-2.6.12-rc2 |
233 |
spinlock_t *lock_ptr; |
1da177e4c Linux-2.6.12-rc2 |
234 |
union futex_key key; |
c87e2837b [PATCH] pi-futex:... |
235 |
struct futex_pi_state *pi_state; |
52400ba94 futex: add requeu... |
236 |
struct rt_mutex_waiter *rt_waiter; |
84bc4af59 futex: Detect mis... |
237 |
union futex_key *requeue_pi_key; |
cd689985c futex: Add bitset... |
238 |
u32 bitset; |
3859a271a randstruct: Mark ... |
239 |
} __randomize_layout; |
1da177e4c Linux-2.6.12-rc2 |
240 |
|
5bdb05f91 futex: Add futex_... |
241 242 243 244 245 |
static const struct futex_q futex_q_init = { /* list gets initialized in queue_me()*/ .key = FUTEX_KEY_INIT, .bitset = FUTEX_BITSET_MATCH_ANY }; |
1da177e4c Linux-2.6.12-rc2 |
246 |
/* |
b2d0994b1 futex: update fut... |
247 248 249 |
* Hash buckets are shared by all the futex_keys that hash to the same * location. Each key may have multiple futex_q structures, one for each task * waiting on a futex. |
1da177e4c Linux-2.6.12-rc2 |
250 251 |
*/ struct futex_hash_bucket { |
11d4616bd futex: revert bac... |
252 |
atomic_t waiters; |
ec92d0829 futex priority ba... |
253 254 |
spinlock_t lock; struct plist_head chain; |
a52b89ebb futexes: Increase... |
255 |
} ____cacheline_aligned_in_smp; |
1da177e4c Linux-2.6.12-rc2 |
256 |
|
ac742d371 futex: Force hot ... |
257 258 259 260 261 262 263 264 265 266 267 |
/* * The base of the bucket array and its size are always used together * (after initialization only in hash_futex()), so ensure that they * reside in the same cacheline. */ static struct { struct futex_hash_bucket *queues; unsigned long hashsize; } __futex_data __read_mostly __aligned(2*sizeof(long)); #define futex_queues (__futex_data.queues) #define futex_hashsize (__futex_data.hashsize) |
a52b89ebb futexes: Increase... |
268 |
|
1da177e4c Linux-2.6.12-rc2 |
269 |
|
ab51fbab3 futex: Fault/erro... |
270 271 272 273 274 275 276 |
/* * Fault injections for futexes. */ #ifdef CONFIG_FAIL_FUTEX static struct { struct fault_attr attr; |
621a5f7ad debugfs: Pass boo... |
277 |
bool ignore_private; |
ab51fbab3 futex: Fault/erro... |
278 279 |
} fail_futex = { .attr = FAULT_ATTR_INITIALIZER, |
621a5f7ad debugfs: Pass boo... |
280 |
.ignore_private = false, |
ab51fbab3 futex: Fault/erro... |
281 282 283 284 285 286 287 |
}; static int __init setup_fail_futex(char *str) { return setup_fault_attr(&fail_futex.attr, str); } __setup("fail_futex=", setup_fail_futex); |
5d285a7f3 futex: Make shoul... |
288 |
static bool should_fail_futex(bool fshared) |
ab51fbab3 futex: Fault/erro... |
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 |
{ if (fail_futex.ignore_private && !fshared) return false; return should_fail(&fail_futex.attr, 1); } #ifdef CONFIG_FAULT_INJECTION_DEBUG_FS static int __init fail_futex_debugfs(void) { umode_t mode = S_IFREG | S_IRUSR | S_IWUSR; struct dentry *dir; dir = fault_create_debugfs_attr("fail_futex", NULL, &fail_futex.attr); if (IS_ERR(dir)) return PTR_ERR(dir); |
0365aeba5 futex: No need to... |
307 308 |
debugfs_create_bool("ignore-private", mode, dir, &fail_futex.ignore_private); |
ab51fbab3 futex: Fault/erro... |
309 310 311 312 313 314 315 316 317 318 319 320 321 |
return 0; } late_initcall(fail_futex_debugfs); #endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */ #else static inline bool should_fail_futex(bool fshared) { return false; } #endif /* CONFIG_FAIL_FUTEX */ |
8012f98f9 futex: Move futex... |
322 323 324 325 326 |
#ifdef CONFIG_COMPAT static void compat_exit_robust_list(struct task_struct *curr); #else static inline void compat_exit_robust_list(struct task_struct *curr) { } #endif |
b0c29f79e futexes: Avoid ta... |
327 328 |
static inline void futex_get_mm(union futex_key *key) { |
f1f100764 mm: add new mmgra... |
329 |
mmgrab(key->private.mm); |
b0c29f79e futexes: Avoid ta... |
330 331 332 |
/* * Ensure futex_get_mm() implies a full barrier such that * get_futex_key() implies a full barrier. This is relied upon |
8ad7b378d futex: Rename bar... |
333 |
* as smp_mb(); (B), see the ordering comment above. |
b0c29f79e futexes: Avoid ta... |
334 |
*/ |
4e857c58e arch: Mass conver... |
335 |
smp_mb__after_atomic(); |
b0c29f79e futexes: Avoid ta... |
336 |
} |
11d4616bd futex: revert bac... |
337 338 339 340 |
/* * Reflects a new waiter being added to the waitqueue. */ static inline void hb_waiters_inc(struct futex_hash_bucket *hb) |
b0c29f79e futexes: Avoid ta... |
341 342 |
{ #ifdef CONFIG_SMP |
11d4616bd futex: revert bac... |
343 |
atomic_inc(&hb->waiters); |
b0c29f79e futexes: Avoid ta... |
344 |
/* |
11d4616bd futex: revert bac... |
345 |
* Full barrier (A), see the ordering comment above. |
b0c29f79e futexes: Avoid ta... |
346 |
*/ |
4e857c58e arch: Mass conver... |
347 |
smp_mb__after_atomic(); |
11d4616bd futex: revert bac... |
348 349 350 351 352 353 354 355 356 357 358 359 360 |
#endif } /* * Reflects a waiter being removed from the waitqueue by wakeup * paths. */ static inline void hb_waiters_dec(struct futex_hash_bucket *hb) { #ifdef CONFIG_SMP atomic_dec(&hb->waiters); #endif } |
b0c29f79e futexes: Avoid ta... |
361 |
|
11d4616bd futex: revert bac... |
362 363 364 365 |
static inline int hb_waiters_pending(struct futex_hash_bucket *hb) { #ifdef CONFIG_SMP return atomic_read(&hb->waiters); |
b0c29f79e futexes: Avoid ta... |
366 |
#else |
11d4616bd futex: revert bac... |
367 |
return 1; |
b0c29f79e futexes: Avoid ta... |
368 369 |
#endif } |
e8b61b3f2 futex: Add some m... |
370 371 372 373 374 375 |
/** * hash_futex - Return the hash bucket in the global hash * @key: Pointer to the futex key for which the hash is calculated * * We hash on the keys returned from get_futex_key (see below) and return the * corresponding hash bucket in the global hash. |
1da177e4c Linux-2.6.12-rc2 |
376 377 378 |
*/ static struct futex_hash_bucket *hash_futex(union futex_key *key) { |
163489b64 futex: Unbreak fu... |
379 |
u32 hash = jhash2((u32 *)key, offsetof(typeof(*key), both.offset) / 4, |
1da177e4c Linux-2.6.12-rc2 |
380 |
key->both.offset); |
163489b64 futex: Unbreak fu... |
381 |
|
a52b89ebb futexes: Increase... |
382 |
return &futex_queues[hash & (futex_hashsize - 1)]; |
1da177e4c Linux-2.6.12-rc2 |
383 |
} |
e8b61b3f2 futex: Add some m... |
384 385 386 387 388 389 |
/** * match_futex - Check whether two futex keys are equal * @key1: Pointer to key1 * @key2: Pointer to key2 * |
1da177e4c Linux-2.6.12-rc2 |
390 391 392 393 |
* Return 1 if two futex_keys are equal, 0 otherwise. */ static inline int match_futex(union futex_key *key1, union futex_key *key2) { |
2bc872036 futex: Check for ... |
394 395 |
return (key1 && key2 && key1->both.word == key2->both.word |
1da177e4c Linux-2.6.12-rc2 |
396 397 398 |
&& key1->both.ptr == key2->both.ptr && key1->both.offset == key2->both.offset); } |
38d47c1b7 futex: rely on ge... |
399 400 401 402 403 404 405 406 407 |
/* * Take a reference to the resource addressed by a key. * Can be called while holding spinlocks. * */ static void get_futex_key_refs(union futex_key *key) { if (!key->both.ptr) return; |
784bdf3bb futex: Assume all... |
408 409 410 411 412 413 414 415 416 |
/* * On MMU less systems futexes are always "private" as there is no per * process address space. We need the smp wmb nevertheless - yes, * arch/blackfin has MMU less SMP ... */ if (!IS_ENABLED(CONFIG_MMU)) { smp_mb(); /* explicit smp_mb(); (B) */ return; } |
38d47c1b7 futex: rely on ge... |
417 418 |
switch (key->both.offset & (FUT_OFF_INODE|FUT_OFF_MMSHARED)) { case FUT_OFF_INODE: |
553d46b07 futex: Fix inode ... |
419 |
smp_mb(); /* explicit smp_mb(); (B) */ |
38d47c1b7 futex: rely on ge... |
420 421 |
break; case FUT_OFF_MMSHARED: |
8ad7b378d futex: Rename bar... |
422 |
futex_get_mm(key); /* implies smp_mb(); (B) */ |
38d47c1b7 futex: rely on ge... |
423 |
break; |
76835b0eb futex: Ensure get... |
424 |
default: |
993b2ff22 futex: Mention ke... |
425 426 427 428 429 |
/* * Private futexes do not hold reference on an inode or * mm, therefore the only purpose of calling get_futex_key_refs * is because we need the barrier for the lockless waiter check. */ |
8ad7b378d futex: Rename bar... |
430 |
smp_mb(); /* explicit smp_mb(); (B) */ |
38d47c1b7 futex: rely on ge... |
431 432 433 434 435 |
} } /* * Drop a reference to the resource addressed by a key. |
993b2ff22 futex: Mention ke... |
436 437 438 |
* The hash bucket spinlock must not be held. This is * a no-op for private futexes, see comment in the get * counterpart. |
38d47c1b7 futex: rely on ge... |
439 440 441 |
*/ static void drop_futex_key_refs(union futex_key *key) { |
90621c40c futex: catch cert... |
442 443 444 |
if (!key->both.ptr) { /* If we're here then we tried to put a key we failed to get */ WARN_ON_ONCE(1); |
38d47c1b7 futex: rely on ge... |
445 |
return; |
90621c40c futex: catch cert... |
446 |
} |
38d47c1b7 futex: rely on ge... |
447 |
|
784bdf3bb futex: Assume all... |
448 449 |
if (!IS_ENABLED(CONFIG_MMU)) return; |
38d47c1b7 futex: rely on ge... |
450 451 |
switch (key->both.offset & (FUT_OFF_INODE|FUT_OFF_MMSHARED)) { case FUT_OFF_INODE: |
38d47c1b7 futex: rely on ge... |
452 453 454 455 456 457 |
break; case FUT_OFF_MMSHARED: mmdrop(key->private.mm); break; } } |
96d4f267e Remove 'type' arg... |
458 459 460 461 |
enum futex_access { FUTEX_READ, FUTEX_WRITE }; |
34f01cc1f FUTEX: new PRIVAT... |
462 |
/** |
5ca584d93 futex: Consolidat... |
463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 |
* futex_setup_timer - set up the sleeping hrtimer. * @time: ptr to the given timeout value * @timeout: the hrtimer_sleeper structure to be set up * @flags: futex flags * @range_ns: optional range in ns * * Return: Initialized hrtimer_sleeper structure or NULL if no timeout * value given */ static inline struct hrtimer_sleeper * futex_setup_timer(ktime_t *time, struct hrtimer_sleeper *timeout, int flags, u64 range_ns) { if (!time) return NULL; |
dbc1625fc hrtimer: Consolid... |
478 479 480 |
hrtimer_init_sleeper_on_stack(timeout, (flags & FLAGS_CLOCKRT) ? CLOCK_REALTIME : CLOCK_MONOTONIC, HRTIMER_MODE_ABS); |
5ca584d93 futex: Consolidat... |
481 482 483 484 485 486 487 488 |
/* * If range_ns is 0, calling hrtimer_set_expires_range_ns() is * effectively the same as calling hrtimer_set_expires(). */ hrtimer_set_expires_range_ns(&timeout->timer, *time, range_ns); return timeout; } |
553d46b07 futex: Fix inode ... |
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 |
/* * Generate a machine wide unique identifier for this inode. * * This relies on u64 not wrapping in the life-time of the machine; which with * 1ns resolution means almost 585 years. * * This further relies on the fact that a well formed program will not unmap * the file while it has a (shared) futex waiting on it. This mapping will have * a file reference which pins the mount and inode. * * If for some reason an inode gets evicted and read back in again, it will get * a new sequence number and will _NOT_ match, even though it is the exact same * file. * * It is important that match_futex() will never have a false-positive, esp. * for PI futexes that can mess up the state. The above argues that false-negatives * are only possible for malformed programs. */ static u64 get_inode_sequence_number(struct inode *inode) { static atomic64_t i_seq; u64 old; /* Does the inode already have a sequence number? */ old = atomic64_read(&inode->i_sequence); if (likely(old)) return old; for (;;) { u64 new = atomic64_add_return(1, &i_seq); if (WARN_ON_ONCE(!new)) continue; old = atomic64_cmpxchg_relaxed(&inode->i_sequence, 0, new); if (old) return old; return new; } } |
5ca584d93 futex: Consolidat... |
528 |
/** |
d96ee56ce futex: Make funct... |
529 530 531 532 |
* get_futex_key() - Get parameters which are the keys for a futex * @uaddr: virtual address of the futex * @fshared: 0 for a PROCESS_PRIVATE futex, 1 for PROCESS_SHARED * @key: address where result is stored. |
96d4f267e Remove 'type' arg... |
533 534 |
* @rw: mapping needs to be read/write (values: FUTEX_READ, * FUTEX_WRITE) |
34f01cc1f FUTEX: new PRIVAT... |
535 |
* |
6c23cbbd5 futex: fix kernel... |
536 537 |
* Return: a negative error code or 0 * |
7b4ff1adb mutex, futex: adj... |
538 |
* The key words are stored in @key on success. |
1da177e4c Linux-2.6.12-rc2 |
539 |
* |
553d46b07 futex: Fix inode ... |
540 541 542 543 544 545 546 547 548 |
* For shared mappings (when @fshared), the key is: * ( inode->i_sequence, page->index, offset_within_page ) * [ also see get_inode_sequence_number() ] * * For private mappings (or when !@fshared), the key is: * ( current->mm, address, 0 ) * * This allows (cross process, where applicable) identification of the futex * without keeping the page pinned for the duration of the FUTEX_WAIT. |
1da177e4c Linux-2.6.12-rc2 |
549 |
* |
b2d0994b1 futex: update fut... |
550 |
* lock_page() might sleep, the caller should not hold a spinlock. |
1da177e4c Linux-2.6.12-rc2 |
551 |
*/ |
64d1304a6 futex: setup writ... |
552 |
static int |
96d4f267e Remove 'type' arg... |
553 |
get_futex_key(u32 __user *uaddr, int fshared, union futex_key *key, enum futex_access rw) |
1da177e4c Linux-2.6.12-rc2 |
554 |
{ |
e2970f2fb [PATCH] pi-futex:... |
555 |
unsigned long address = (unsigned long)uaddr; |
1da177e4c Linux-2.6.12-rc2 |
556 |
struct mm_struct *mm = current->mm; |
077fa7aed futex: Calculate ... |
557 |
struct page *page, *tail; |
14d27abd1 futex, thp: remov... |
558 |
struct address_space *mapping; |
9ea71503a futex: Fix regres... |
559 |
int err, ro = 0; |
1da177e4c Linux-2.6.12-rc2 |
560 561 562 563 |
/* * The futex address must be "naturally" aligned. */ |
e2970f2fb [PATCH] pi-futex:... |
564 |
key->both.offset = address % PAGE_SIZE; |
34f01cc1f FUTEX: new PRIVAT... |
565 |
if (unlikely((address % sizeof(u32)) != 0)) |
1da177e4c Linux-2.6.12-rc2 |
566 |
return -EINVAL; |
e2970f2fb [PATCH] pi-futex:... |
567 |
address -= key->both.offset; |
1da177e4c Linux-2.6.12-rc2 |
568 |
|
96d4f267e Remove 'type' arg... |
569 |
if (unlikely(!access_ok(uaddr, sizeof(u32)))) |
5cdec2d83 futex: move user ... |
570 |
return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
571 572 |
if (unlikely(should_fail_futex(fshared))) return -EFAULT; |
1da177e4c Linux-2.6.12-rc2 |
573 |
/* |
34f01cc1f FUTEX: new PRIVAT... |
574 575 576 577 578 579 580 |
* PROCESS_PRIVATE futexes are fast. * As the mm cannot disappear under us and the 'key' only needs * virtual address, we dont even have to find the underlying vma. * Note : We do have to check 'uaddr' is a valid user address, * but access_ok() should be faster than find_vma() */ if (!fshared) { |
34f01cc1f FUTEX: new PRIVAT... |
581 582 |
key->private.mm = mm; key->private.address = address; |
8ad7b378d futex: Rename bar... |
583 |
get_futex_key_refs(key); /* implies smp_mb(); (B) */ |
34f01cc1f FUTEX: new PRIVAT... |
584 585 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
586 |
|
38d47c1b7 futex: rely on ge... |
587 |
again: |
ab51fbab3 futex: Fault/erro... |
588 589 590 |
/* Ignore any VERIFY_READ mapping (futex common case) */ if (unlikely(should_fail_futex(fshared))) return -EFAULT; |
73b0140bf mm/gup: change GU... |
591 |
err = get_user_pages_fast(address, 1, FOLL_WRITE, &page); |
9ea71503a futex: Fix regres... |
592 593 594 595 |
/* * If write access is not required (eg. FUTEX_WAIT), try * and get read-only access. */ |
96d4f267e Remove 'type' arg... |
596 |
if (err == -EFAULT && rw == FUTEX_READ) { |
9ea71503a futex: Fix regres... |
597 598 599 |
err = get_user_pages_fast(address, 1, 0, &page); ro = 1; } |
38d47c1b7 futex: rely on ge... |
600 601 |
if (err < 0) return err; |
9ea71503a futex: Fix regres... |
602 603 |
else err = 0; |
38d47c1b7 futex: rely on ge... |
604 |
|
65d8fc777 futex: Remove req... |
605 606 607 608 609 610 611 612 613 614 |
/* * The treatment of mapping from this point on is critical. The page * lock protects many things but in this context the page lock * stabilizes mapping, prevents inode freeing in the shared * file-backed region case and guards against movement to swap cache. * * Strictly speaking the page lock is not needed in all cases being * considered here and page lock forces unnecessarily serialization * From this point on, mapping will be re-verified if necessary and * page lock will be acquired only if it is unavoidable |
077fa7aed futex: Calculate ... |
615 616 617 618 619 620 621 |
* * Mapping checks require the head page for any compound page so the * head page and mapping is looked up now. For anonymous pages, it * does not matter if the page splits in the future as the key is * based on the address. For filesystem-backed pages, the tail is * required as the index of the page determines the key. For * base pages, there is no tail page and tail == page. |
65d8fc777 futex: Remove req... |
622 |
*/ |
077fa7aed futex: Calculate ... |
623 |
tail = page; |
65d8fc777 futex: Remove req... |
624 625 |
page = compound_head(page); mapping = READ_ONCE(page->mapping); |
e6780f724 futex: Fix uninte... |
626 |
/* |
14d27abd1 futex, thp: remov... |
627 |
* If page->mapping is NULL, then it cannot be a PageAnon |
e6780f724 futex: Fix uninte... |
628 629 630 631 632 633 634 635 636 637 638 |
* page; but it might be the ZERO_PAGE or in the gate area or * in a special mapping (all cases which we are happy to fail); * or it may have been a good file page when get_user_pages_fast * found it, but truncated or holepunched or subjected to * invalidate_complete_page2 before we got the page lock (also * cases which we are happy to fail). And we hold a reference, * so refcount care in invalidate_complete_page's remove_mapping * prevents drop_caches from setting mapping to NULL beneath us. * * The case we do have to guard against is when memory pressure made * shmem_writepage move it from filecache to swapcache beneath us: |
14d27abd1 futex, thp: remov... |
639 |
* an unlikely race, but we do need to retry for page->mapping. |
e6780f724 futex: Fix uninte... |
640 |
*/ |
65d8fc777 futex: Remove req... |
641 642 643 644 645 646 647 648 649 650 |
if (unlikely(!mapping)) { int shmem_swizzled; /* * Page lock is required to identify which special case above * applies. If this is really a shmem page then the page lock * will prevent unexpected transitions. */ lock_page(page); shmem_swizzled = PageSwapCache(page) || page->mapping; |
14d27abd1 futex, thp: remov... |
651 652 |
unlock_page(page); put_page(page); |
65d8fc777 futex: Remove req... |
653 |
|
e6780f724 futex: Fix uninte... |
654 655 |
if (shmem_swizzled) goto again; |
65d8fc777 futex: Remove req... |
656 |
|
e6780f724 futex: Fix uninte... |
657 |
return -EFAULT; |
38d47c1b7 futex: rely on ge... |
658 |
} |
1da177e4c Linux-2.6.12-rc2 |
659 660 661 662 |
/* * Private mappings are handled in a simple way. * |
65d8fc777 futex: Remove req... |
663 664 665 |
* If the futex key is stored on an anonymous page, then the associated * object is the mm which is implicitly pinned by the calling process. * |
1da177e4c Linux-2.6.12-rc2 |
666 667 |
* NOTE: When userspace waits on a MAP_SHARED mapping, even if * it's a read-only handle, it's expected that futexes attach to |
38d47c1b7 futex: rely on ge... |
668 |
* the object not the particular process. |
1da177e4c Linux-2.6.12-rc2 |
669 |
*/ |
14d27abd1 futex, thp: remov... |
670 |
if (PageAnon(page)) { |
9ea71503a futex: Fix regres... |
671 672 673 674 |
/* * A RO anonymous page will never change and thus doesn't make * sense for futex operations. */ |
ab51fbab3 futex: Fault/erro... |
675 |
if (unlikely(should_fail_futex(fshared)) || ro) { |
9ea71503a futex: Fix regres... |
676 677 678 |
err = -EFAULT; goto out; } |
38d47c1b7 futex: rely on ge... |
679 |
key->both.offset |= FUT_OFF_MMSHARED; /* ref taken on mm */ |
1da177e4c Linux-2.6.12-rc2 |
680 |
key->private.mm = mm; |
e2970f2fb [PATCH] pi-futex:... |
681 |
key->private.address = address; |
65d8fc777 futex: Remove req... |
682 |
|
38d47c1b7 futex: rely on ge... |
683 |
} else { |
65d8fc777 futex: Remove req... |
684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 |
struct inode *inode; /* * The associated futex object in this case is the inode and * the page->mapping must be traversed. Ordinarily this should * be stabilised under page lock but it's not strictly * necessary in this case as we just want to pin the inode, not * update the radix tree or anything like that. * * The RCU read lock is taken as the inode is finally freed * under RCU. If the mapping still matches expectations then the * mapping->host can be safely accessed as being a valid inode. */ rcu_read_lock(); if (READ_ONCE(page->mapping) != mapping) { rcu_read_unlock(); put_page(page); goto again; } inode = READ_ONCE(mapping->host); if (!inode) { rcu_read_unlock(); put_page(page); goto again; } |
38d47c1b7 futex: rely on ge... |
713 |
key->both.offset |= FUT_OFF_INODE; /* inode-based key */ |
553d46b07 futex: Fix inode ... |
714 |
key->shared.i_seq = get_inode_sequence_number(inode); |
077fa7aed futex: Calculate ... |
715 |
key->shared.pgoff = basepage_index(tail); |
65d8fc777 futex: Remove req... |
716 |
rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
717 |
} |
553d46b07 futex: Fix inode ... |
718 |
get_futex_key_refs(key); /* implies smp_mb(); (B) */ |
9ea71503a futex: Fix regres... |
719 |
out: |
14d27abd1 futex, thp: remov... |
720 |
put_page(page); |
9ea71503a futex: Fix regres... |
721 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
722 |
} |
ae791a2d2 futex: Cleanup st... |
723 |
static inline void put_futex_key(union futex_key *key) |
1da177e4c Linux-2.6.12-rc2 |
724 |
{ |
38d47c1b7 futex: rely on ge... |
725 |
drop_futex_key_refs(key); |
1da177e4c Linux-2.6.12-rc2 |
726 |
} |
d96ee56ce futex: Make funct... |
727 728 |
/** * fault_in_user_writeable() - Fault in user address and verify RW access |
d0725992c futex: Fix the wr... |
729 730 731 732 733 |
* @uaddr: pointer to faulting user space address * * Slow path to fixup the fault we just took in the atomic write * access to @uaddr. * |
fb62db2ba futex: Fix kernel... |
734 |
* We have no generic implementation of a non-destructive write to the |
d0725992c futex: Fix the wr... |
735 736 737 738 739 740 |
* user address. We know that we faulted in the atomic pagefault * disabled section so we can as well avoid the #PF overhead by * calling get_user_pages() right away. */ static int fault_in_user_writeable(u32 __user *uaddr) { |
722d01723 futex: Take mmap_... |
741 742 743 744 |
struct mm_struct *mm = current->mm; int ret; down_read(&mm->mmap_sem); |
2efaca927 mm/futex: fix fut... |
745 |
ret = fixup_user_fault(current, mm, (unsigned long)uaddr, |
4a9e1cda2 mm: bring in addi... |
746 |
FAULT_FLAG_WRITE, NULL); |
722d01723 futex: Take mmap_... |
747 |
up_read(&mm->mmap_sem); |
d0725992c futex: Fix the wr... |
748 749 |
return ret < 0 ? ret : 0; } |
4b1c486b3 futex: add helper... |
750 751 |
/** * futex_top_waiter() - Return the highest priority waiter on a futex |
d96ee56ce futex: Make funct... |
752 753 |
* @hb: the hash bucket the futex_q's reside in * @key: the futex key (to distinguish it from other futex futex_q's) |
4b1c486b3 futex: add helper... |
754 755 756 757 758 759 760 761 762 763 764 765 766 767 |
* * Must be called with the hb lock held. */ static struct futex_q *futex_top_waiter(struct futex_hash_bucket *hb, union futex_key *key) { struct futex_q *this; plist_for_each_entry(this, &hb->chain, list) { if (match_futex(&this->key, key)) return this; } return NULL; } |
37a9d912b futex: Sanitize c... |
768 769 |
static int cmpxchg_futex_value_locked(u32 *curval, u32 __user *uaddr, u32 uval, u32 newval) |
36cf3b5c3 FUTEX: Tidy up th... |
770 |
{ |
37a9d912b futex: Sanitize c... |
771 |
int ret; |
36cf3b5c3 FUTEX: Tidy up th... |
772 773 |
pagefault_disable(); |
37a9d912b futex: Sanitize c... |
774 |
ret = futex_atomic_cmpxchg_inatomic(curval, uaddr, uval, newval); |
36cf3b5c3 FUTEX: Tidy up th... |
775 |
pagefault_enable(); |
37a9d912b futex: Sanitize c... |
776 |
return ret; |
36cf3b5c3 FUTEX: Tidy up th... |
777 778 779 |
} static int get_futex_value_locked(u32 *dest, u32 __user *from) |
1da177e4c Linux-2.6.12-rc2 |
780 781 |
{ int ret; |
a866374ae [PATCH] mm: pagef... |
782 |
pagefault_disable(); |
bd28b1459 x86: remove more ... |
783 |
ret = __get_user(*dest, from); |
a866374ae [PATCH] mm: pagef... |
784 |
pagefault_enable(); |
1da177e4c Linux-2.6.12-rc2 |
785 786 787 |
return ret ? -EFAULT : 0; } |
c87e2837b [PATCH] pi-futex:... |
788 789 790 791 792 793 794 795 796 797 |
/* * PI code: */ static int refill_pi_state_cache(void) { struct futex_pi_state *pi_state; if (likely(current->pi_state_cache)) return 0; |
4668edc33 [PATCH] kernel co... |
798 |
pi_state = kzalloc(sizeof(*pi_state), GFP_KERNEL); |
c87e2837b [PATCH] pi-futex:... |
799 800 801 |
if (!pi_state) return -ENOMEM; |
c87e2837b [PATCH] pi-futex:... |
802 803 804 |
INIT_LIST_HEAD(&pi_state->list); /* pi_mutex gets initialized later */ pi_state->owner = NULL; |
49262de22 futex: Convert fu... |
805 |
refcount_set(&pi_state->refcount, 1); |
38d47c1b7 futex: rely on ge... |
806 |
pi_state->key = FUTEX_KEY_INIT; |
c87e2837b [PATCH] pi-futex:... |
807 808 809 810 811 |
current->pi_state_cache = pi_state; return 0; } |
bf92cf3a5 futex: Cleanup re... |
812 |
static struct futex_pi_state *alloc_pi_state(void) |
c87e2837b [PATCH] pi-futex:... |
813 814 815 816 817 818 819 820 |
{ struct futex_pi_state *pi_state = current->pi_state_cache; WARN_ON(!pi_state); current->pi_state_cache = NULL; return pi_state; } |
bf92cf3a5 futex: Cleanup re... |
821 822 |
static void get_pi_state(struct futex_pi_state *pi_state) { |
49262de22 futex: Convert fu... |
823 |
WARN_ON_ONCE(!refcount_inc_not_zero(&pi_state->refcount)); |
bf92cf3a5 futex: Cleanup re... |
824 |
} |
30a6b8031 futex: Fix a race... |
825 |
/* |
29e9ee5d4 futex: Rename fre... |
826 827 |
* Drops a reference to the pi_state object and frees or caches it * when the last reference is gone. |
30a6b8031 futex: Fix a race... |
828 |
*/ |
29e9ee5d4 futex: Rename fre... |
829 |
static void put_pi_state(struct futex_pi_state *pi_state) |
c87e2837b [PATCH] pi-futex:... |
830 |
{ |
30a6b8031 futex: Fix a race... |
831 832 |
if (!pi_state) return; |
49262de22 futex: Convert fu... |
833 |
if (!refcount_dec_and_test(&pi_state->refcount)) |
c87e2837b [PATCH] pi-futex:... |
834 835 836 837 838 839 840 |
return; /* * If pi_state->owner is NULL, the owner is most probably dying * and has cleaned up the pi_state already */ if (pi_state->owner) { |
c74aef2d0 futex: Fix pi_sta... |
841 |
struct task_struct *owner; |
c87e2837b [PATCH] pi-futex:... |
842 |
|
c74aef2d0 futex: Fix pi_sta... |
843 844 845 846 847 848 849 850 851 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); owner = pi_state->owner; if (owner) { raw_spin_lock(&owner->pi_lock); list_del_init(&pi_state->list); raw_spin_unlock(&owner->pi_lock); } rt_mutex_proxy_unlock(&pi_state->pi_mutex, owner); raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
c87e2837b [PATCH] pi-futex:... |
852 |
} |
c74aef2d0 futex: Fix pi_sta... |
853 |
if (current->pi_state_cache) { |
c87e2837b [PATCH] pi-futex:... |
854 |
kfree(pi_state); |
c74aef2d0 futex: Fix pi_sta... |
855 |
} else { |
c87e2837b [PATCH] pi-futex:... |
856 857 858 859 860 861 |
/* * pi_state->list is already empty. * clear pi_state->owner. * refcount is at 0 - put it back to 1. */ pi_state->owner = NULL; |
49262de22 futex: Convert fu... |
862 |
refcount_set(&pi_state->refcount, 1); |
c87e2837b [PATCH] pi-futex:... |
863 864 865 |
current->pi_state_cache = pi_state; } } |
bc2eecd7e futex: Allow for ... |
866 |
#ifdef CONFIG_FUTEX_PI |
c87e2837b [PATCH] pi-futex:... |
867 868 869 870 871 |
/* * This task is holding PI mutexes at exit time => bad. * Kernel cleans up PI-state, but userspace is likely hosed. * (Robust-futex cleanup is separate and might save the day for userspace.) */ |
8012f98f9 futex: Move futex... |
872 |
static void exit_pi_state_list(struct task_struct *curr) |
c87e2837b [PATCH] pi-futex:... |
873 |
{ |
c87e2837b [PATCH] pi-futex:... |
874 875 |
struct list_head *next, *head = &curr->pi_state_list; struct futex_pi_state *pi_state; |
627371d73 [PATCH] pi-futex:... |
876 |
struct futex_hash_bucket *hb; |
38d47c1b7 futex: rely on ge... |
877 |
union futex_key key = FUTEX_KEY_INIT; |
c87e2837b [PATCH] pi-futex:... |
878 |
|
a0c1e9073 futex: runtime en... |
879 880 |
if (!futex_cmpxchg_enabled) return; |
c87e2837b [PATCH] pi-futex:... |
881 882 883 |
/* * We are a ZOMBIE and nobody can enqueue itself on * pi_state_list anymore, but we have to be careful |
627371d73 [PATCH] pi-futex:... |
884 |
* versus waiters unqueueing themselves: |
c87e2837b [PATCH] pi-futex:... |
885 |
*/ |
1d6154825 sched: Convert pi... |
886 |
raw_spin_lock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
887 |
while (!list_empty(head)) { |
c87e2837b [PATCH] pi-futex:... |
888 889 890 |
next = head->next; pi_state = list_entry(next, struct futex_pi_state, list); key = pi_state->key; |
627371d73 [PATCH] pi-futex:... |
891 |
hb = hash_futex(&key); |
153fbd122 futex: Fix more p... |
892 893 894 895 896 897 898 899 900 901 902 |
/* * We can race against put_pi_state() removing itself from the * list (a waiter going away). put_pi_state() will first * decrement the reference count and then modify the list, so * its possible to see the list entry but fail this reference * acquire. * * In that case; drop the locks to let put_pi_state() make * progress and retry the loop. */ |
49262de22 futex: Convert fu... |
903 |
if (!refcount_inc_not_zero(&pi_state->refcount)) { |
153fbd122 futex: Fix more p... |
904 905 906 907 908 |
raw_spin_unlock_irq(&curr->pi_lock); cpu_relax(); raw_spin_lock_irq(&curr->pi_lock); continue; } |
1d6154825 sched: Convert pi... |
909 |
raw_spin_unlock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
910 |
|
c87e2837b [PATCH] pi-futex:... |
911 |
spin_lock(&hb->lock); |
c74aef2d0 futex: Fix pi_sta... |
912 913 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); raw_spin_lock(&curr->pi_lock); |
627371d73 [PATCH] pi-futex:... |
914 915 916 917 |
/* * We dropped the pi-lock, so re-check whether this * task still owns the PI-state: */ |
c87e2837b [PATCH] pi-futex:... |
918 |
if (head->next != next) { |
153fbd122 futex: Fix more p... |
919 |
/* retain curr->pi_lock for the loop invariant */ |
c74aef2d0 futex: Fix pi_sta... |
920 |
raw_spin_unlock(&pi_state->pi_mutex.wait_lock); |
c87e2837b [PATCH] pi-futex:... |
921 |
spin_unlock(&hb->lock); |
153fbd122 futex: Fix more p... |
922 |
put_pi_state(pi_state); |
c87e2837b [PATCH] pi-futex:... |
923 924 |
continue; } |
c87e2837b [PATCH] pi-futex:... |
925 |
WARN_ON(pi_state->owner != curr); |
627371d73 [PATCH] pi-futex:... |
926 927 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
c87e2837b [PATCH] pi-futex:... |
928 |
pi_state->owner = NULL; |
c87e2837b [PATCH] pi-futex:... |
929 |
|
153fbd122 futex: Fix more p... |
930 |
raw_spin_unlock(&curr->pi_lock); |
c74aef2d0 futex: Fix pi_sta... |
931 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
c87e2837b [PATCH] pi-futex:... |
932 |
spin_unlock(&hb->lock); |
16ffa12d7 futex: Pull rt_mu... |
933 934 |
rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); |
1d6154825 sched: Convert pi... |
935 |
raw_spin_lock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
936 |
} |
1d6154825 sched: Convert pi... |
937 |
raw_spin_unlock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
938 |
} |
8012f98f9 futex: Move futex... |
939 940 |
#else static inline void exit_pi_state_list(struct task_struct *curr) { } |
bc2eecd7e futex: Allow for ... |
941 |
#endif |
54a217887 futex: Make looku... |
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 |
/* * We need to check the following states: * * Waiter | pi_state | pi->owner | uTID | uODIED | ? * * [1] NULL | --- | --- | 0 | 0/1 | Valid * [2] NULL | --- | --- | >0 | 0/1 | Valid * * [3] Found | NULL | -- | Any | 0/1 | Invalid * * [4] Found | Found | NULL | 0 | 1 | Valid * [5] Found | Found | NULL | >0 | 1 | Invalid * * [6] Found | Found | task | 0 | 1 | Valid * * [7] Found | Found | NULL | Any | 0 | Invalid * * [8] Found | Found | task | ==taskTID | 0/1 | Valid * [9] Found | Found | task | 0 | 0 | Invalid * [10] Found | Found | task | !=taskTID | 0/1 | Invalid * * [1] Indicates that the kernel can acquire the futex atomically. We * came came here due to a stale FUTEX_WAITERS/FUTEX_OWNER_DIED bit. * * [2] Valid, if TID does not belong to a kernel thread. If no matching * thread is found then it indicates that the owner TID has died. * * [3] Invalid. The waiter is queued on a non PI futex * * [4] Valid state after exit_robust_list(), which sets the user space * value to FUTEX_WAITERS | FUTEX_OWNER_DIED. * * [5] The user space value got manipulated between exit_robust_list() * and exit_pi_state_list() * * [6] Valid state after exit_pi_state_list() which sets the new owner in * the pi_state but cannot access the user space value. * * [7] pi_state->owner can only be NULL when the OWNER_DIED bit is set. * * [8] Owner and user space value match * * [9] There is no transient state which sets the user space TID to 0 * except exit_robust_list(), but this is indicated by the * FUTEX_OWNER_DIED bit. See [4] * * [10] There is no transient state which leaves owner and user space * TID out of sync. |
734009e96 futex: Change loc... |
990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 |
* * * Serialization and lifetime rules: * * hb->lock: * * hb -> futex_q, relation * futex_q -> pi_state, relation * * (cannot be raw because hb can contain arbitrary amount * of futex_q's) * * pi_mutex->wait_lock: * * {uval, pi_state} * * (and pi_mutex 'obviously') * * p->pi_lock: * * p->pi_state_list -> pi_state->list, relation * * pi_state->refcount: * * pi_state lifetime * * * Lock order: * * hb->lock * pi_mutex->wait_lock * p->pi_lock * |
54a217887 futex: Make looku... |
1023 |
*/ |
e60cbc5ce futex: Split out ... |
1024 1025 1026 1027 1028 1029 |
/* * Validate that the existing waiter has a pi_state and sanity check * the pi_state against the user space value. If correct, attach to * it. */ |
734009e96 futex: Change loc... |
1030 1031 |
static int attach_to_pi_state(u32 __user *uaddr, u32 uval, struct futex_pi_state *pi_state, |
e60cbc5ce futex: Split out ... |
1032 |
struct futex_pi_state **ps) |
c87e2837b [PATCH] pi-futex:... |
1033 |
{ |
778e9a9c3 pi-futex: fix exi... |
1034 |
pid_t pid = uval & FUTEX_TID_MASK; |
94ffac5d8 futex: Fix small ... |
1035 1036 |
u32 uval2; int ret; |
c87e2837b [PATCH] pi-futex:... |
1037 |
|
e60cbc5ce futex: Split out ... |
1038 1039 1040 1041 1042 |
/* * Userspace might have messed up non-PI and PI futexes [3] */ if (unlikely(!pi_state)) return -EINVAL; |
06a9ec291 [PATCH] pi-futex:... |
1043 |
|
734009e96 futex: Change loc... |
1044 1045 1046 1047 1048 1049 |
/* * We get here with hb->lock held, and having found a * futex_top_waiter(). This means that futex_lock_pi() of said futex_q * has dropped the hb->lock in between queue_me() and unqueue_me_pi(), * which in turn means that futex_lock_pi() still has a reference on * our pi_state. |
16ffa12d7 futex: Pull rt_mu... |
1050 1051 1052 1053 1054 |
* * The waiter holding a reference on @pi_state also protects against * the unlocked put_pi_state() in futex_unlock_pi(), futex_lock_pi() * and futex_wait_requeue_pi() as it cannot go to 0 and consequently * free pi_state before we can take a reference ourselves. |
734009e96 futex: Change loc... |
1055 |
*/ |
49262de22 futex: Convert fu... |
1056 |
WARN_ON(!refcount_read(&pi_state->refcount)); |
59647b6ac futex: Handle fut... |
1057 |
|
e60cbc5ce futex: Split out ... |
1058 |
/* |
734009e96 futex: Change loc... |
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 |
* Now that we have a pi_state, we can acquire wait_lock * and do the state validation. */ raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); /* * Since {uval, pi_state} is serialized by wait_lock, and our current * uval was read without holding it, it can have changed. Verify it * still is what we expect it to be, otherwise retry the entire * operation. */ if (get_futex_value_locked(&uval2, uaddr)) goto out_efault; if (uval != uval2) goto out_eagain; /* |
e60cbc5ce futex: Split out ... |
1077 1078 1079 |
* Handle the owner died case: */ if (uval & FUTEX_OWNER_DIED) { |
bd1dbcc67 futex: Use futex_... |
1080 |
/* |
e60cbc5ce futex: Split out ... |
1081 1082 1083 |
* exit_pi_state_list sets owner to NULL and wakes the * topmost waiter. The task which acquires the * pi_state->rt_mutex will fixup owner. |
bd1dbcc67 futex: Use futex_... |
1084 |
*/ |
e60cbc5ce futex: Split out ... |
1085 |
if (!pi_state->owner) { |
59647b6ac futex: Handle fut... |
1086 |
/* |
e60cbc5ce futex: Split out ... |
1087 1088 |
* No pi state owner, but the user space TID * is not 0. Inconsistent state. [5] |
59647b6ac futex: Handle fut... |
1089 |
*/ |
e60cbc5ce futex: Split out ... |
1090 |
if (pid) |
734009e96 futex: Change loc... |
1091 |
goto out_einval; |
bd1dbcc67 futex: Use futex_... |
1092 |
/* |
e60cbc5ce futex: Split out ... |
1093 |
* Take a ref on the state and return success. [4] |
866293ee5 futex: Add anothe... |
1094 |
*/ |
734009e96 futex: Change loc... |
1095 |
goto out_attach; |
c87e2837b [PATCH] pi-futex:... |
1096 |
} |
bd1dbcc67 futex: Use futex_... |
1097 1098 |
/* |
e60cbc5ce futex: Split out ... |
1099 1100 1101 1102 1103 1104 1105 1106 |
* If TID is 0, then either the dying owner has not * yet executed exit_pi_state_list() or some waiter * acquired the rtmutex in the pi state, but did not * yet fixup the TID in user space. * * Take a ref on the state and return success. [6] */ if (!pid) |
734009e96 futex: Change loc... |
1107 |
goto out_attach; |
e60cbc5ce futex: Split out ... |
1108 1109 1110 1111 |
} else { /* * If the owner died bit is not set, then the pi_state * must have an owner. [7] |
bd1dbcc67 futex: Use futex_... |
1112 |
*/ |
e60cbc5ce futex: Split out ... |
1113 |
if (!pi_state->owner) |
734009e96 futex: Change loc... |
1114 |
goto out_einval; |
c87e2837b [PATCH] pi-futex:... |
1115 1116 1117 |
} /* |
e60cbc5ce futex: Split out ... |
1118 1119 1120 1121 1122 |
* Bail out if user space manipulated the futex value. If pi * state exists then the owner TID must be the same as the * user space TID. [9/10] */ if (pid != task_pid_vnr(pi_state->owner)) |
734009e96 futex: Change loc... |
1123 1124 1125 |
goto out_einval; out_attach: |
bf92cf3a5 futex: Cleanup re... |
1126 |
get_pi_state(pi_state); |
734009e96 futex: Change loc... |
1127 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
e60cbc5ce futex: Split out ... |
1128 1129 |
*ps = pi_state; return 0; |
734009e96 futex: Change loc... |
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 |
out_einval: ret = -EINVAL; goto out_error; out_eagain: ret = -EAGAIN; goto out_error; out_efault: ret = -EFAULT; goto out_error; out_error: raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); return ret; |
e60cbc5ce futex: Split out ... |
1146 |
} |
fc3b55ef2 futex: Prevent ex... |
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 |
/** * wait_for_owner_exiting - Block until the owner has exited * @exiting: Pointer to the exiting task * * Caller must hold a refcount on @exiting. */ static void wait_for_owner_exiting(int ret, struct task_struct *exiting) { if (ret != -EBUSY) { WARN_ON_ONCE(exiting); return; } if (WARN_ON_ONCE(ret == -EBUSY && !exiting)) return; mutex_lock(&exiting->futex_exit_mutex); /* * No point in doing state checking here. If the waiter got here * while the task was in exec()->exec_futex_release() then it can * have any FUTEX_STATE_* value when the waiter has acquired the * mutex. OK, if running, EXITING or DEAD if it reached exit() * already. Highly unlikely and not a problem. Just one more round * through the futex maze. */ mutex_unlock(&exiting->futex_exit_mutex); put_task_struct(exiting); } |
da791a667 futex: Cure exit ... |
1176 1177 1178 1179 1180 1181 |
static int handle_exit_race(u32 __user *uaddr, u32 uval, struct task_struct *tsk) { u32 uval2; /* |
56690612a futex: Provide di... |
1182 1183 |
* If the futex exit state is not yet FUTEX_STATE_DEAD, tell the * caller that the alleged owner is busy. |
da791a667 futex: Cure exit ... |
1184 |
*/ |
52507cfaf futex: Replace PF... |
1185 |
if (tsk && tsk->futex_state != FUTEX_STATE_DEAD) |
56690612a futex: Provide di... |
1186 |
return -EBUSY; |
da791a667 futex: Cure exit ... |
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 |
/* * Reread the user space value to handle the following situation: * * CPU0 CPU1 * * sys_exit() sys_futex() * do_exit() futex_lock_pi() * futex_lock_pi_atomic() * exit_signals(tsk) No waiters: * tsk->flags |= PF_EXITING; *uaddr == 0x00000PID * mm_release(tsk) Set waiter bit * exit_robust_list(tsk) { *uaddr = 0x80000PID; * Set owner died attach_to_pi_owner() { * *uaddr = 0xC0000000; tsk = get_task(PID); * } if (!tsk->flags & PF_EXITING) { * ... attach(); |
52507cfaf futex: Replace PF... |
1204 1205 1206 |
* tsk->futex_state = } else { * FUTEX_STATE_DEAD; if (tsk->futex_state != * FUTEX_STATE_DEAD) |
da791a667 futex: Cure exit ... |
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 |
* return -EAGAIN; * return -ESRCH; <--- FAIL * } * * Returning ESRCH unconditionally is wrong here because the * user space value has been changed by the exiting task. * * The same logic applies to the case where the exiting task is * already gone. */ if (get_futex_value_locked(&uval2, uaddr)) return -EFAULT; /* If the user space value has changed, try again. */ if (uval2 != uval) return -EAGAIN; /* * The exiting task did not have a robust list, the robust list was * corrupted or the user space value in *uaddr is simply bogus. * Give up and tell user space. */ return -ESRCH; } |
04e1b2e52 futex: Split out ... |
1231 1232 1233 1234 |
/* * Lookup the task for the TID provided from user space and attach to * it after doing proper sanity checks. */ |
da791a667 futex: Cure exit ... |
1235 |
static int attach_to_pi_owner(u32 __user *uaddr, u32 uval, union futex_key *key, |
fc3b55ef2 futex: Prevent ex... |
1236 1237 |
struct futex_pi_state **ps, struct task_struct **exiting) |
e60cbc5ce futex: Split out ... |
1238 |
{ |
e60cbc5ce futex: Split out ... |
1239 |
pid_t pid = uval & FUTEX_TID_MASK; |
04e1b2e52 futex: Split out ... |
1240 1241 |
struct futex_pi_state *pi_state; struct task_struct *p; |
e60cbc5ce futex: Split out ... |
1242 1243 |
/* |
e3f2ddeac [PATCH] pi-futex:... |
1244 |
* We are the first waiter - try to look up the real owner and attach |
54a217887 futex: Make looku... |
1245 |
* the new pi_state to it, but bail out when TID = 0 [1] |
da791a667 futex: Cure exit ... |
1246 1247 1248 |
* * The !pid check is paranoid. None of the call sites should end up * with pid == 0, but better safe than sorry. Let the caller retry |
c87e2837b [PATCH] pi-futex:... |
1249 |
*/ |
778e9a9c3 pi-futex: fix exi... |
1250 |
if (!pid) |
da791a667 futex: Cure exit ... |
1251 |
return -EAGAIN; |
2ee082608 pids: introduce f... |
1252 |
p = find_get_task_by_vpid(pid); |
7a0ea09ad futex: futex_find... |
1253 |
if (!p) |
da791a667 futex: Cure exit ... |
1254 |
return handle_exit_race(uaddr, uval, NULL); |
778e9a9c3 pi-futex: fix exi... |
1255 |
|
a21294644 locking/futex: Ch... |
1256 |
if (unlikely(p->flags & PF_KTHREAD)) { |
f0d71b3dc futex: Prevent at... |
1257 1258 1259 |
put_task_struct(p); return -EPERM; } |
778e9a9c3 pi-futex: fix exi... |
1260 |
/* |
52507cfaf futex: Replace PF... |
1261 1262 1263 |
* We need to look at the task state to figure out, whether the * task is exiting. To protect against the change of the task state * in futex_exit_release(), we do this protected by p->pi_lock: |
778e9a9c3 pi-futex: fix exi... |
1264 |
*/ |
1d6154825 sched: Convert pi... |
1265 |
raw_spin_lock_irq(&p->pi_lock); |
52507cfaf futex: Replace PF... |
1266 |
if (unlikely(p->futex_state != FUTEX_STATE_OK)) { |
778e9a9c3 pi-futex: fix exi... |
1267 |
/* |
52507cfaf futex: Replace PF... |
1268 1269 1270 |
* The task is on the way out. When the futex state is * FUTEX_STATE_DEAD, we know that the task has finished * the cleanup: |
778e9a9c3 pi-futex: fix exi... |
1271 |
*/ |
da791a667 futex: Cure exit ... |
1272 |
int ret = handle_exit_race(uaddr, uval, p); |
778e9a9c3 pi-futex: fix exi... |
1273 |
|
1d6154825 sched: Convert pi... |
1274 |
raw_spin_unlock_irq(&p->pi_lock); |
fc3b55ef2 futex: Prevent ex... |
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 |
/* * If the owner task is between FUTEX_STATE_EXITING and * FUTEX_STATE_DEAD then store the task pointer and keep * the reference on the task struct. The calling code will * drop all locks, wait for the task to reach * FUTEX_STATE_DEAD and then drop the refcount. This is * required to prevent a live lock when the current task * preempted the exiting task between the two states. */ if (ret == -EBUSY) *exiting = p; else put_task_struct(p); |
778e9a9c3 pi-futex: fix exi... |
1288 1289 |
return ret; } |
c87e2837b [PATCH] pi-futex:... |
1290 |
|
54a217887 futex: Make looku... |
1291 1292 |
/* * No existing pi state. First waiter. [2] |
734009e96 futex: Change loc... |
1293 1294 1295 |
* * This creates pi_state, we have hb->lock held, this means nothing can * observe this state, wait_lock is irrelevant. |
54a217887 futex: Make looku... |
1296 |
*/ |
c87e2837b [PATCH] pi-futex:... |
1297 1298 1299 |
pi_state = alloc_pi_state(); /* |
04e1b2e52 futex: Split out ... |
1300 |
* Initialize the pi_mutex in locked state and make @p |
c87e2837b [PATCH] pi-futex:... |
1301 1302 1303 1304 1305 |
* the owner of it: */ rt_mutex_init_proxy_locked(&pi_state->pi_mutex, p); /* Store the key for possible exit cleanups: */ |
d0aa7a70b futex_requeue_pi ... |
1306 |
pi_state->key = *key; |
c87e2837b [PATCH] pi-futex:... |
1307 |
|
627371d73 [PATCH] pi-futex:... |
1308 |
WARN_ON(!list_empty(&pi_state->list)); |
c87e2837b [PATCH] pi-futex:... |
1309 |
list_add(&pi_state->list, &p->pi_state_list); |
c74aef2d0 futex: Fix pi_sta... |
1310 1311 1312 1313 |
/* * Assignment without holding pi_state->pi_mutex.wait_lock is safe * because there is no concurrency as the object is not published yet. */ |
c87e2837b [PATCH] pi-futex:... |
1314 |
pi_state->owner = p; |
1d6154825 sched: Convert pi... |
1315 |
raw_spin_unlock_irq(&p->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
1316 1317 |
put_task_struct(p); |
d0aa7a70b futex_requeue_pi ... |
1318 |
*ps = pi_state; |
c87e2837b [PATCH] pi-futex:... |
1319 1320 1321 |
return 0; } |
734009e96 futex: Change loc... |
1322 1323 |
static int lookup_pi_state(u32 __user *uaddr, u32 uval, struct futex_hash_bucket *hb, |
fc3b55ef2 futex: Prevent ex... |
1324 1325 |
union futex_key *key, struct futex_pi_state **ps, struct task_struct **exiting) |
04e1b2e52 futex: Split out ... |
1326 |
{ |
499f5aca2 futex: Cleanup va... |
1327 |
struct futex_q *top_waiter = futex_top_waiter(hb, key); |
04e1b2e52 futex: Split out ... |
1328 1329 1330 1331 1332 |
/* * If there is a waiter on that futex, validate it and * attach to the pi_state when the validation succeeds. */ |
499f5aca2 futex: Cleanup va... |
1333 |
if (top_waiter) |
734009e96 futex: Change loc... |
1334 |
return attach_to_pi_state(uaddr, uval, top_waiter->pi_state, ps); |
04e1b2e52 futex: Split out ... |
1335 1336 1337 1338 1339 |
/* * We are the first waiter - try to look up the owner based on * @uval and attach to it. */ |
fc3b55ef2 futex: Prevent ex... |
1340 |
return attach_to_pi_owner(uaddr, uval, key, ps, exiting); |
04e1b2e52 futex: Split out ... |
1341 |
} |
af54d6a1c futex: Simplify f... |
1342 1343 |
static int lock_pi_update_atomic(u32 __user *uaddr, u32 uval, u32 newval) { |
6b4f4bc9c locking/futex: Al... |
1344 |
int err; |
af54d6a1c futex: Simplify f... |
1345 |
u32 uninitialized_var(curval); |
ab51fbab3 futex: Fault/erro... |
1346 1347 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
6b4f4bc9c locking/futex: Al... |
1348 1349 1350 |
err = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (unlikely(err)) return err; |
af54d6a1c futex: Simplify f... |
1351 |
|
734009e96 futex: Change loc... |
1352 |
/* If user space value changed, let the caller retry */ |
af54d6a1c futex: Simplify f... |
1353 1354 |
return curval != uval ? -EAGAIN : 0; } |
1a52084d0 futex: split out ... |
1355 |
/** |
d96ee56ce futex: Make funct... |
1356 |
* futex_lock_pi_atomic() - Atomic work required to acquire a pi aware futex |
bab5bc9e8 futex: fixup unlo... |
1357 1358 1359 1360 1361 1362 1363 |
* @uaddr: the pi futex user address * @hb: the pi futex hash bucket * @key: the futex key associated with uaddr and hb * @ps: the pi_state pointer where we store the result of the * lookup * @task: the task to perform the atomic lock work for. This will * be "current" except in the case of requeue pi. |
fc3b55ef2 futex: Prevent ex... |
1364 1365 |
* @exiting: Pointer to store the task pointer of the owner task * which is in the middle of exiting |
bab5bc9e8 futex: fixup unlo... |
1366 |
* @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) |
1a52084d0 futex: split out ... |
1367 |
* |
6c23cbbd5 futex: fix kernel... |
1368 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1369 1370 1371 |
* - 0 - ready to wait; * - 1 - acquired the lock; * - <0 - error |
1a52084d0 futex: split out ... |
1372 1373 |
* * The hb->lock and futex_key refs shall be held by the caller. |
fc3b55ef2 futex: Prevent ex... |
1374 1375 1376 1377 |
* * @exiting is only set when the return value is -EBUSY. If so, this holds * a refcount on the exiting task on return and the caller needs to drop it * after waiting for the exit to complete. |
1a52084d0 futex: split out ... |
1378 1379 1380 1381 |
*/ static int futex_lock_pi_atomic(u32 __user *uaddr, struct futex_hash_bucket *hb, union futex_key *key, struct futex_pi_state **ps, |
fc3b55ef2 futex: Prevent ex... |
1382 1383 1384 |
struct task_struct *task, struct task_struct **exiting, int set_waiters) |
1a52084d0 futex: split out ... |
1385 |
{ |
af54d6a1c futex: Simplify f... |
1386 |
u32 uval, newval, vpid = task_pid_vnr(task); |
499f5aca2 futex: Cleanup va... |
1387 |
struct futex_q *top_waiter; |
af54d6a1c futex: Simplify f... |
1388 |
int ret; |
1a52084d0 futex: split out ... |
1389 1390 |
/* |
af54d6a1c futex: Simplify f... |
1391 1392 |
* Read the user space value first so we can validate a few * things before proceeding further. |
1a52084d0 futex: split out ... |
1393 |
*/ |
af54d6a1c futex: Simplify f... |
1394 |
if (get_futex_value_locked(&uval, uaddr)) |
1a52084d0 futex: split out ... |
1395 |
return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
1396 1397 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
1a52084d0 futex: split out ... |
1398 1399 1400 |
/* * Detect deadlocks. */ |
af54d6a1c futex: Simplify f... |
1401 |
if ((unlikely((uval & FUTEX_TID_MASK) == vpid))) |
1a52084d0 futex: split out ... |
1402 |
return -EDEADLK; |
ab51fbab3 futex: Fault/erro... |
1403 1404 |
if ((unlikely(should_fail_futex(true)))) return -EDEADLK; |
1a52084d0 futex: split out ... |
1405 |
/* |
af54d6a1c futex: Simplify f... |
1406 1407 |
* Lookup existing state first. If it exists, try to attach to * its pi_state. |
1a52084d0 futex: split out ... |
1408 |
*/ |
499f5aca2 futex: Cleanup va... |
1409 1410 |
top_waiter = futex_top_waiter(hb, key); if (top_waiter) |
734009e96 futex: Change loc... |
1411 |
return attach_to_pi_state(uaddr, uval, top_waiter->pi_state, ps); |
1a52084d0 futex: split out ... |
1412 1413 |
/* |
af54d6a1c futex: Simplify f... |
1414 1415 1416 1417 |
* No waiter and user TID is 0. We are here because the * waiters or the owner died bit is set or called from * requeue_cmp_pi or for whatever reason something took the * syscall. |
1a52084d0 futex: split out ... |
1418 |
*/ |
af54d6a1c futex: Simplify f... |
1419 |
if (!(uval & FUTEX_TID_MASK)) { |
59fa62451 futex: Handle fut... |
1420 |
/* |
af54d6a1c futex: Simplify f... |
1421 1422 |
* We take over the futex. No other waiters and the user space * TID is 0. We preserve the owner died bit. |
59fa62451 futex: Handle fut... |
1423 |
*/ |
af54d6a1c futex: Simplify f... |
1424 1425 |
newval = uval & FUTEX_OWNER_DIED; newval |= vpid; |
1a52084d0 futex: split out ... |
1426 |
|
af54d6a1c futex: Simplify f... |
1427 1428 1429 1430 1431 1432 1433 1434 |
/* The futex requeue_pi code can enforce the waiters bit */ if (set_waiters) newval |= FUTEX_WAITERS; ret = lock_pi_update_atomic(uaddr, uval, newval); /* If the take over worked, return 1 */ return ret < 0 ? ret : 1; } |
1a52084d0 futex: split out ... |
1435 1436 |
/* |
af54d6a1c futex: Simplify f... |
1437 1438 1439 |
* First waiter. Set the waiters bit before attaching ourself to * the owner. If owner tries to unlock, it will be forced into * the kernel and blocked on hb->lock. |
1a52084d0 futex: split out ... |
1440 |
*/ |
af54d6a1c futex: Simplify f... |
1441 1442 1443 1444 |
newval = uval | FUTEX_WAITERS; ret = lock_pi_update_atomic(uaddr, uval, newval); if (ret) return ret; |
1a52084d0 futex: split out ... |
1445 |
/* |
af54d6a1c futex: Simplify f... |
1446 1447 1448 |
* If the update of the user space value succeeded, we try to * attach to the owner. If that fails, no harm done, we only * set the FUTEX_WAITERS bit in the user space variable. |
1a52084d0 futex: split out ... |
1449 |
*/ |
fc3b55ef2 futex: Prevent ex... |
1450 |
return attach_to_pi_owner(uaddr, newval, key, ps, exiting); |
1a52084d0 futex: split out ... |
1451 |
} |
2e12978a9 futex,plist: Pass... |
1452 1453 1454 1455 1456 1457 1458 1459 1460 |
/** * __unqueue_futex() - Remove the futex_q from its futex_hash_bucket * @q: The futex_q to unqueue * * The q->lock_ptr must not be NULL and must be held by the caller. */ static void __unqueue_futex(struct futex_q *q) { struct futex_hash_bucket *hb; |
4de1a293a futex: Replace sp... |
1461 |
if (WARN_ON_SMP(!q->lock_ptr) || WARN_ON(plist_node_empty(&q->list))) |
2e12978a9 futex,plist: Pass... |
1462 |
return; |
4de1a293a futex: Replace sp... |
1463 |
lockdep_assert_held(q->lock_ptr); |
2e12978a9 futex,plist: Pass... |
1464 1465 1466 |
hb = container_of(q->lock_ptr, struct futex_hash_bucket, lock); plist_del(&q->list, &hb->chain); |
11d4616bd futex: revert bac... |
1467 |
hb_waiters_dec(hb); |
2e12978a9 futex,plist: Pass... |
1468 |
} |
c87e2837b [PATCH] pi-futex:... |
1469 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1470 |
* The hash bucket lock must be held when this is called. |
1d0dcb3ad futex: Implement ... |
1471 1472 1473 |
* Afterwards, the futex_q must not be accessed. Callers * must ensure to later call wake_up_q() for the actual * wakeups to occur. |
1da177e4c Linux-2.6.12-rc2 |
1474 |
*/ |
1d0dcb3ad futex: Implement ... |
1475 |
static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q) |
1da177e4c Linux-2.6.12-rc2 |
1476 |
{ |
f1a11e057 futex: remove the... |
1477 |
struct task_struct *p = q->task; |
aa10990e0 futex: avoid wake... |
1478 1479 1480 |
if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex ")) return; |
b061c38be futex: Fix (possi... |
1481 |
get_task_struct(p); |
2e12978a9 futex,plist: Pass... |
1482 |
__unqueue_futex(q); |
1da177e4c Linux-2.6.12-rc2 |
1483 |
/* |
38fcd06e9 futex: Clarify ma... |
1484 1485 1486 1487 1488 |
* The waiting task can free the futex_q as soon as q->lock_ptr = NULL * is written, without taking any locks. This is possible in the event * of a spurious wakeup, for example. A memory barrier is required here * to prevent the following store to lock_ptr from getting ahead of the * plist_del in __unqueue_futex(). |
1da177e4c Linux-2.6.12-rc2 |
1489 |
*/ |
1b367ece0 futex: Use smp_st... |
1490 |
smp_store_release(&q->lock_ptr, NULL); |
b061c38be futex: Fix (possi... |
1491 1492 1493 1494 1495 |
/* * Queue the task for later wakeup for after we've released * the hb->lock. wake_q_add() grabs reference to p. */ |
07879c6a3 sched/wake_q: Red... |
1496 |
wake_q_add_safe(wake_q, p); |
1da177e4c Linux-2.6.12-rc2 |
1497 |
} |
16ffa12d7 futex: Pull rt_mu... |
1498 1499 1500 1501 |
/* * Caller must hold a reference on @pi_state. */ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_pi_state *pi_state) |
c87e2837b [PATCH] pi-futex:... |
1502 |
{ |
7cfdaf38d futex: uninitiali... |
1503 |
u32 uninitialized_var(curval), newval; |
16ffa12d7 futex: Pull rt_mu... |
1504 |
struct task_struct *new_owner; |
aa2bfe553 rtmutex: Clean up |
1505 |
bool postunlock = false; |
194a6b5b9 sched/wake_q: Ren... |
1506 |
DEFINE_WAKE_Q(wake_q); |
13fbca4c6 futex: Always cle... |
1507 |
int ret = 0; |
c87e2837b [PATCH] pi-futex:... |
1508 |
|
c87e2837b [PATCH] pi-futex:... |
1509 |
new_owner = rt_mutex_next_owner(&pi_state->pi_mutex); |
bebe5b514 futex: Futex_unlo... |
1510 |
if (WARN_ON_ONCE(!new_owner)) { |
16ffa12d7 futex: Pull rt_mu... |
1511 |
/* |
bebe5b514 futex: Futex_unlo... |
1512 |
* As per the comment in futex_unlock_pi() this should not happen. |
16ffa12d7 futex: Pull rt_mu... |
1513 1514 1515 1516 1517 1518 1519 1520 |
* * When this happens, give up our locks and try again, giving * the futex_lock_pi() instance time to complete, either by * waiting on the rtmutex or removing itself from the futex * queue. */ ret = -EAGAIN; goto out_unlock; |
73d786bd0 futex: Rework inc... |
1521 |
} |
c87e2837b [PATCH] pi-futex:... |
1522 1523 |
/* |
16ffa12d7 futex: Pull rt_mu... |
1524 1525 1526 |
* We pass it to the next owner. The WAITERS bit is always kept * enabled while there is PI state around. We cleanup the owner * died bit, because we are the owner. |
c87e2837b [PATCH] pi-futex:... |
1527 |
*/ |
13fbca4c6 futex: Always cle... |
1528 |
newval = FUTEX_WAITERS | task_pid_vnr(new_owner); |
e3f2ddeac [PATCH] pi-futex:... |
1529 |
|
ab51fbab3 futex: Fault/erro... |
1530 1531 |
if (unlikely(should_fail_futex(true))) ret = -EFAULT; |
6b4f4bc9c locking/futex: Al... |
1532 1533 |
ret = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (!ret && (curval != uval)) { |
89e9e66ba futex: Handle unl... |
1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 |
/* * If a unconditional UNLOCK_PI operation (user space did not * try the TID->0 transition) raced with a waiter setting the * FUTEX_WAITERS flag between get_user() and locking the hash * bucket lock, retry the operation. */ if ((FUTEX_TID_MASK & curval) == uval) ret = -EAGAIN; else ret = -EINVAL; } |
734009e96 futex: Change loc... |
1545 |
|
16ffa12d7 futex: Pull rt_mu... |
1546 1547 |
if (ret) goto out_unlock; |
c87e2837b [PATCH] pi-futex:... |
1548 |
|
94ffac5d8 futex: Fix small ... |
1549 1550 1551 1552 |
/* * This is a point of no return; once we modify the uval there is no * going back and subsequent operations must not fail. */ |
b4abf9104 rtmutex: Make wai... |
1553 |
raw_spin_lock(&pi_state->owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1554 1555 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
b4abf9104 rtmutex: Make wai... |
1556 |
raw_spin_unlock(&pi_state->owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1557 |
|
b4abf9104 rtmutex: Make wai... |
1558 |
raw_spin_lock(&new_owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1559 |
WARN_ON(!list_empty(&pi_state->list)); |
c87e2837b [PATCH] pi-futex:... |
1560 1561 |
list_add(&pi_state->list, &new_owner->pi_state_list); pi_state->owner = new_owner; |
b4abf9104 rtmutex: Make wai... |
1562 |
raw_spin_unlock(&new_owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1563 |
|
aa2bfe553 rtmutex: Clean up |
1564 |
postunlock = __rt_mutex_futex_unlock(&pi_state->pi_mutex, &wake_q); |
5293c2efd futex,rt_mutex: P... |
1565 |
|
16ffa12d7 futex: Pull rt_mu... |
1566 |
out_unlock: |
5293c2efd futex,rt_mutex: P... |
1567 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
5293c2efd futex,rt_mutex: P... |
1568 |
|
aa2bfe553 rtmutex: Clean up |
1569 1570 |
if (postunlock) rt_mutex_postunlock(&wake_q); |
c87e2837b [PATCH] pi-futex:... |
1571 |
|
16ffa12d7 futex: Pull rt_mu... |
1572 |
return ret; |
c87e2837b [PATCH] pi-futex:... |
1573 |
} |
1da177e4c Linux-2.6.12-rc2 |
1574 |
/* |
8b8f319fc [PATCH] lockdep: ... |
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 |
* Express the locking dependencies for lockdep: */ static inline void double_lock_hb(struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2) { if (hb1 <= hb2) { spin_lock(&hb1->lock); if (hb1 < hb2) spin_lock_nested(&hb2->lock, SINGLE_DEPTH_NESTING); } else { /* hb1 > hb2 */ spin_lock(&hb2->lock); spin_lock_nested(&hb1->lock, SINGLE_DEPTH_NESTING); } } |
5eb3dc62f futex: add double... |
1589 1590 1591 |
static inline void double_unlock_hb(struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2) { |
f061d3515 futex: remove the... |
1592 |
spin_unlock(&hb1->lock); |
88f502fed futex: remove the... |
1593 1594 |
if (hb1 != hb2) spin_unlock(&hb2->lock); |
5eb3dc62f futex: add double... |
1595 |
} |
8b8f319fc [PATCH] lockdep: ... |
1596 |
/* |
b2d0994b1 futex: update fut... |
1597 |
* Wake up waiters matching bitset queued on this futex (uaddr). |
1da177e4c Linux-2.6.12-rc2 |
1598 |
*/ |
b41277dc7 futex: Replace fs... |
1599 1600 |
static int futex_wake(u32 __user *uaddr, unsigned int flags, int nr_wake, u32 bitset) |
1da177e4c Linux-2.6.12-rc2 |
1601 |
{ |
e2970f2fb [PATCH] pi-futex:... |
1602 |
struct futex_hash_bucket *hb; |
1da177e4c Linux-2.6.12-rc2 |
1603 |
struct futex_q *this, *next; |
38d47c1b7 futex: rely on ge... |
1604 |
union futex_key key = FUTEX_KEY_INIT; |
1da177e4c Linux-2.6.12-rc2 |
1605 |
int ret; |
194a6b5b9 sched/wake_q: Ren... |
1606 |
DEFINE_WAKE_Q(wake_q); |
1da177e4c Linux-2.6.12-rc2 |
1607 |
|
cd689985c futex: Add bitset... |
1608 1609 |
if (!bitset) return -EINVAL; |
96d4f267e Remove 'type' arg... |
1610 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &key, FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1611 1612 |
if (unlikely(ret != 0)) goto out; |
e2970f2fb [PATCH] pi-futex:... |
1613 |
hb = hash_futex(&key); |
b0c29f79e futexes: Avoid ta... |
1614 1615 1616 1617 |
/* Make sure we really have tasks to wakeup */ if (!hb_waiters_pending(hb)) goto out_put_key; |
e2970f2fb [PATCH] pi-futex:... |
1618 |
spin_lock(&hb->lock); |
1da177e4c Linux-2.6.12-rc2 |
1619 |
|
0d00c7b20 futexes: Clean up... |
1620 |
plist_for_each_entry_safe(this, next, &hb->chain, list) { |
1da177e4c Linux-2.6.12-rc2 |
1621 |
if (match_futex (&this->key, &key)) { |
52400ba94 futex: add requeu... |
1622 |
if (this->pi_state || this->rt_waiter) { |
ed6f7b10e [PATCH] pi-futex:... |
1623 1624 1625 |
ret = -EINVAL; break; } |
cd689985c futex: Add bitset... |
1626 1627 1628 1629 |
/* Check if one of the bits is set in both bitsets */ if (!(this->bitset & bitset)) continue; |
1d0dcb3ad futex: Implement ... |
1630 |
mark_wake_futex(&wake_q, this); |
1da177e4c Linux-2.6.12-rc2 |
1631 1632 1633 1634 |
if (++ret >= nr_wake) break; } } |
e2970f2fb [PATCH] pi-futex:... |
1635 |
spin_unlock(&hb->lock); |
1d0dcb3ad futex: Implement ... |
1636 |
wake_up_q(&wake_q); |
b0c29f79e futexes: Avoid ta... |
1637 |
out_put_key: |
ae791a2d2 futex: Cleanup st... |
1638 |
put_futex_key(&key); |
42d35d48c futex: make futex... |
1639 |
out: |
1da177e4c Linux-2.6.12-rc2 |
1640 1641 |
return ret; } |
30d6e0a41 futex: Remove dup... |
1642 1643 1644 1645 |
static int futex_atomic_op_inuser(unsigned int encoded_op, u32 __user *uaddr) { unsigned int op = (encoded_op & 0x70000000) >> 28; unsigned int cmp = (encoded_op & 0x0f000000) >> 24; |
d70ef2289 futex: futex_wake... |
1646 1647 |
int oparg = sign_extend32((encoded_op & 0x00fff000) >> 12, 11); int cmparg = sign_extend32(encoded_op & 0x00000fff, 11); |
30d6e0a41 futex: Remove dup... |
1648 1649 1650 |
int oldval, ret; if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28)) { |
e78c38f6b futex: futex_wake... |
1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 |
if (oparg < 0 || oparg > 31) { char comm[sizeof(current->comm)]; /* * kill this print and return -EINVAL when userspace * is sane again */ pr_info_ratelimited("futex_wake_op: %s tries to shift op by %d; fix this program ", get_task_comm(comm, current), oparg); oparg &= 31; } |
30d6e0a41 futex: Remove dup... |
1662 1663 |
oparg = 1 << oparg; } |
96d4f267e Remove 'type' arg... |
1664 |
if (!access_ok(uaddr, sizeof(u32))) |
30d6e0a41 futex: Remove dup... |
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 |
return -EFAULT; ret = arch_futex_atomic_op_inuser(op, oparg, &oldval, uaddr); if (ret) return ret; switch (cmp) { case FUTEX_OP_CMP_EQ: return oldval == cmparg; case FUTEX_OP_CMP_NE: return oldval != cmparg; case FUTEX_OP_CMP_LT: return oldval < cmparg; case FUTEX_OP_CMP_GE: return oldval >= cmparg; case FUTEX_OP_CMP_LE: return oldval <= cmparg; case FUTEX_OP_CMP_GT: return oldval > cmparg; default: return -ENOSYS; } } |
1da177e4c Linux-2.6.12-rc2 |
1688 |
/* |
4732efbeb [PATCH] FUTEX_WAK... |
1689 1690 1691 |
* Wake up all waiters hashed on the physical page that is mapped * to this virtual address: */ |
e2970f2fb [PATCH] pi-futex:... |
1692 |
static int |
b41277dc7 futex: Replace fs... |
1693 |
futex_wake_op(u32 __user *uaddr1, unsigned int flags, u32 __user *uaddr2, |
e2970f2fb [PATCH] pi-futex:... |
1694 |
int nr_wake, int nr_wake2, int op) |
4732efbeb [PATCH] FUTEX_WAK... |
1695 |
{ |
38d47c1b7 futex: rely on ge... |
1696 |
union futex_key key1 = FUTEX_KEY_INIT, key2 = FUTEX_KEY_INIT; |
e2970f2fb [PATCH] pi-futex:... |
1697 |
struct futex_hash_bucket *hb1, *hb2; |
4732efbeb [PATCH] FUTEX_WAK... |
1698 |
struct futex_q *this, *next; |
e4dc5b7a3 futex: clean up f... |
1699 |
int ret, op_ret; |
194a6b5b9 sched/wake_q: Ren... |
1700 |
DEFINE_WAKE_Q(wake_q); |
4732efbeb [PATCH] FUTEX_WAK... |
1701 |
|
e4dc5b7a3 futex: clean up f... |
1702 |
retry: |
96d4f267e Remove 'type' arg... |
1703 |
ret = get_futex_key(uaddr1, flags & FLAGS_SHARED, &key1, FUTEX_READ); |
4732efbeb [PATCH] FUTEX_WAK... |
1704 1705 |
if (unlikely(ret != 0)) goto out; |
96d4f267e Remove 'type' arg... |
1706 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, FUTEX_WRITE); |
4732efbeb [PATCH] FUTEX_WAK... |
1707 |
if (unlikely(ret != 0)) |
42d35d48c futex: make futex... |
1708 |
goto out_put_key1; |
4732efbeb [PATCH] FUTEX_WAK... |
1709 |
|
e2970f2fb [PATCH] pi-futex:... |
1710 1711 |
hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); |
4732efbeb [PATCH] FUTEX_WAK... |
1712 |
|
e4dc5b7a3 futex: clean up f... |
1713 |
retry_private: |
eaaea8036 futex: Fix lockin... |
1714 |
double_lock_hb(hb1, hb2); |
e2970f2fb [PATCH] pi-futex:... |
1715 |
op_ret = futex_atomic_op_inuser(op, uaddr2); |
4732efbeb [PATCH] FUTEX_WAK... |
1716 |
if (unlikely(op_ret < 0)) { |
5eb3dc62f futex: add double... |
1717 |
double_unlock_hb(hb1, hb2); |
4732efbeb [PATCH] FUTEX_WAK... |
1718 |
|
6b4f4bc9c locking/futex: Al... |
1719 1720 1721 1722 1723 1724 |
if (!IS_ENABLED(CONFIG_MMU) || unlikely(op_ret != -EFAULT && op_ret != -EAGAIN)) { /* * we don't get EFAULT from MMU faults if we don't have * an MMU, but we might get them from range checking */ |
796f8d9b9 [PATCH] FUTEX_WAK... |
1725 |
ret = op_ret; |
42d35d48c futex: make futex... |
1726 |
goto out_put_keys; |
796f8d9b9 [PATCH] FUTEX_WAK... |
1727 |
} |
6b4f4bc9c locking/futex: Al... |
1728 1729 1730 1731 1732 |
if (op_ret == -EFAULT) { ret = fault_in_user_writeable(uaddr2); if (ret) goto out_put_keys; } |
4732efbeb [PATCH] FUTEX_WAK... |
1733 |
|
6b4f4bc9c locking/futex: Al... |
1734 1735 |
if (!(flags & FLAGS_SHARED)) { cond_resched(); |
e4dc5b7a3 futex: clean up f... |
1736 |
goto retry_private; |
6b4f4bc9c locking/futex: Al... |
1737 |
} |
e4dc5b7a3 futex: clean up f... |
1738 |
|
ae791a2d2 futex: Cleanup st... |
1739 1740 |
put_futex_key(&key2); put_futex_key(&key1); |
6b4f4bc9c locking/futex: Al... |
1741 |
cond_resched(); |
e4dc5b7a3 futex: clean up f... |
1742 |
goto retry; |
4732efbeb [PATCH] FUTEX_WAK... |
1743 |
} |
0d00c7b20 futexes: Clean up... |
1744 |
plist_for_each_entry_safe(this, next, &hb1->chain, list) { |
4732efbeb [PATCH] FUTEX_WAK... |
1745 |
if (match_futex (&this->key, &key1)) { |
aa10990e0 futex: avoid wake... |
1746 1747 1748 1749 |
if (this->pi_state || this->rt_waiter) { ret = -EINVAL; goto out_unlock; } |
1d0dcb3ad futex: Implement ... |
1750 |
mark_wake_futex(&wake_q, this); |
4732efbeb [PATCH] FUTEX_WAK... |
1751 1752 1753 1754 1755 1756 |
if (++ret >= nr_wake) break; } } if (op_ret > 0) { |
4732efbeb [PATCH] FUTEX_WAK... |
1757 |
op_ret = 0; |
0d00c7b20 futexes: Clean up... |
1758 |
plist_for_each_entry_safe(this, next, &hb2->chain, list) { |
4732efbeb [PATCH] FUTEX_WAK... |
1759 |
if (match_futex (&this->key, &key2)) { |
aa10990e0 futex: avoid wake... |
1760 1761 1762 1763 |
if (this->pi_state || this->rt_waiter) { ret = -EINVAL; goto out_unlock; } |
1d0dcb3ad futex: Implement ... |
1764 |
mark_wake_futex(&wake_q, this); |
4732efbeb [PATCH] FUTEX_WAK... |
1765 1766 1767 1768 1769 1770 |
if (++op_ret >= nr_wake2) break; } } ret += op_ret; } |
aa10990e0 futex: avoid wake... |
1771 |
out_unlock: |
5eb3dc62f futex: add double... |
1772 |
double_unlock_hb(hb1, hb2); |
1d0dcb3ad futex: Implement ... |
1773 |
wake_up_q(&wake_q); |
42d35d48c futex: make futex... |
1774 |
out_put_keys: |
ae791a2d2 futex: Cleanup st... |
1775 |
put_futex_key(&key2); |
42d35d48c futex: make futex... |
1776 |
out_put_key1: |
ae791a2d2 futex: Cleanup st... |
1777 |
put_futex_key(&key1); |
42d35d48c futex: make futex... |
1778 |
out: |
4732efbeb [PATCH] FUTEX_WAK... |
1779 1780 |
return ret; } |
9121e4783 futex: distangle ... |
1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 |
/** * requeue_futex() - Requeue a futex_q from one hb to another * @q: the futex_q to requeue * @hb1: the source hash_bucket * @hb2: the target hash_bucket * @key2: the new key for the requeued futex_q */ static inline void requeue_futex(struct futex_q *q, struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2, union futex_key *key2) { /* * If key1 and key2 hash to the same bucket, no need to * requeue. */ if (likely(&hb1->chain != &hb2->chain)) { plist_del(&q->list, &hb1->chain); |
11d4616bd futex: revert bac... |
1799 |
hb_waiters_dec(hb1); |
11d4616bd futex: revert bac... |
1800 |
hb_waiters_inc(hb2); |
fe1bce9e2 futex: Acknowledg... |
1801 |
plist_add(&q->list, &hb2->chain); |
9121e4783 futex: distangle ... |
1802 |
q->lock_ptr = &hb2->lock; |
9121e4783 futex: distangle ... |
1803 1804 1805 1806 |
} get_futex_key_refs(key2); q->key = *key2; } |
52400ba94 futex: add requeu... |
1807 1808 |
/** * requeue_pi_wake_futex() - Wake a task that acquired the lock during requeue |
d96ee56ce futex: Make funct... |
1809 1810 1811 |
* @q: the futex_q * @key: the key of the requeue target futex * @hb: the hash_bucket of the requeue target futex |
52400ba94 futex: add requeu... |
1812 1813 1814 1815 1816 |
* * During futex_requeue, with requeue_pi=1, it is possible to acquire the * target futex if it is uncontended or via a lock steal. Set the futex_q key * to the requeue target futex so the waiter can detect the wakeup on the right * futex, but remove it from the hb and NULL the rt_waiter so it can detect |
beda2c7ea futex: Update fut... |
1817 1818 1819 |
* atomic lock acquisition. Set the q->lock_ptr to the requeue target hb->lock * to protect access to the pi_state to fixup the owner later. Must be called * with both q->lock_ptr and hb->lock held. |
52400ba94 futex: add requeu... |
1820 1821 |
*/ static inline |
beda2c7ea futex: Update fut... |
1822 1823 |
void requeue_pi_wake_futex(struct futex_q *q, union futex_key *key, struct futex_hash_bucket *hb) |
52400ba94 futex: add requeu... |
1824 |
{ |
52400ba94 futex: add requeu... |
1825 1826 |
get_futex_key_refs(key); q->key = *key; |
2e12978a9 futex,plist: Pass... |
1827 |
__unqueue_futex(q); |
52400ba94 futex: add requeu... |
1828 1829 1830 |
WARN_ON(!q->rt_waiter); q->rt_waiter = NULL; |
beda2c7ea futex: Update fut... |
1831 |
q->lock_ptr = &hb->lock; |
beda2c7ea futex: Update fut... |
1832 |
|
f1a11e057 futex: remove the... |
1833 |
wake_up_state(q->task, TASK_NORMAL); |
52400ba94 futex: add requeu... |
1834 1835 1836 1837 |
} /** * futex_proxy_trylock_atomic() - Attempt an atomic lock for the top waiter |
bab5bc9e8 futex: fixup unlo... |
1838 1839 1840 1841 1842 1843 |
* @pifutex: the user address of the to futex * @hb1: the from futex hash bucket, must be locked by the caller * @hb2: the to futex hash bucket, must be locked by the caller * @key1: the from futex key * @key2: the to futex key * @ps: address to store the pi_state pointer |
fc3b55ef2 futex: Prevent ex... |
1844 1845 |
* @exiting: Pointer to store the task pointer of the owner task * which is in the middle of exiting |
bab5bc9e8 futex: fixup unlo... |
1846 |
* @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) |
52400ba94 futex: add requeu... |
1847 1848 |
* * Try and get the lock on behalf of the top waiter if we can do it atomically. |
bab5bc9e8 futex: fixup unlo... |
1849 1850 1851 |
* Wake the top waiter if we succeed. If the caller specified set_waiters, * then direct futex_lock_pi_atomic() to force setting the FUTEX_WAITERS bit. * hb1 and hb2 must be held by the caller. |
52400ba94 futex: add requeu... |
1852 |
* |
fc3b55ef2 futex: Prevent ex... |
1853 1854 1855 1856 |
* @exiting is only set when the return value is -EBUSY. If so, this holds * a refcount on the exiting task on return and the caller needs to drop it * after waiting for the exit to complete. * |
6c23cbbd5 futex: fix kernel... |
1857 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1858 1859 1860 |
* - 0 - failed to acquire the lock atomically; * - >0 - acquired the lock, return value is vpid of the top_waiter * - <0 - error |
52400ba94 futex: add requeu... |
1861 |
*/ |
fc3b55ef2 futex: Prevent ex... |
1862 1863 1864 1865 1866 |
static int futex_proxy_trylock_atomic(u32 __user *pifutex, struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2, union futex_key *key1, union futex_key *key2, struct futex_pi_state **ps, struct task_struct **exiting, int set_waiters) |
52400ba94 futex: add requeu... |
1867 |
{ |
bab5bc9e8 futex: fixup unlo... |
1868 |
struct futex_q *top_waiter = NULL; |
52400ba94 futex: add requeu... |
1869 |
u32 curval; |
866293ee5 futex: Add anothe... |
1870 |
int ret, vpid; |
52400ba94 futex: add requeu... |
1871 1872 1873 |
if (get_futex_value_locked(&curval, pifutex)) return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
1874 1875 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
bab5bc9e8 futex: fixup unlo... |
1876 1877 1878 1879 1880 1881 1882 1883 |
/* * Find the top_waiter and determine if there are additional waiters. * If the caller intends to requeue more than 1 waiter to pifutex, * force futex_lock_pi_atomic() to set the FUTEX_WAITERS bit now, * as we have means to handle the possible fault. If not, don't set * the bit unecessarily as it will force the subsequent unlock to enter * the kernel. */ |
52400ba94 futex: add requeu... |
1884 1885 1886 1887 1888 |
top_waiter = futex_top_waiter(hb1, key1); /* There are no waiters, nothing for us to do. */ if (!top_waiter) return 0; |
84bc4af59 futex: Detect mis... |
1889 1890 1891 |
/* Ensure we requeue to the expected futex. */ if (!match_futex(top_waiter->requeue_pi_key, key2)) return -EINVAL; |
52400ba94 futex: add requeu... |
1892 |
/* |
bab5bc9e8 futex: fixup unlo... |
1893 1894 1895 |
* Try to take the lock for top_waiter. Set the FUTEX_WAITERS bit in * the contended case or if set_waiters is 1. The pi_state is returned * in ps in contended cases. |
52400ba94 futex: add requeu... |
1896 |
*/ |
866293ee5 futex: Add anothe... |
1897 |
vpid = task_pid_vnr(top_waiter->task); |
bab5bc9e8 futex: fixup unlo... |
1898 |
ret = futex_lock_pi_atomic(pifutex, hb2, key2, ps, top_waiter->task, |
fc3b55ef2 futex: Prevent ex... |
1899 |
exiting, set_waiters); |
866293ee5 futex: Add anothe... |
1900 |
if (ret == 1) { |
beda2c7ea futex: Update fut... |
1901 |
requeue_pi_wake_futex(top_waiter, key2, hb2); |
866293ee5 futex: Add anothe... |
1902 1903 |
return vpid; } |
52400ba94 futex: add requeu... |
1904 1905 1906 1907 1908 |
return ret; } /** * futex_requeue() - Requeue waiters from uaddr1 to uaddr2 |
fb62db2ba futex: Fix kernel... |
1909 |
* @uaddr1: source futex user address |
b41277dc7 futex: Replace fs... |
1910 |
* @flags: futex flags (FLAGS_SHARED, etc.) |
fb62db2ba futex: Fix kernel... |
1911 1912 1913 1914 1915 |
* @uaddr2: target futex user address * @nr_wake: number of waiters to wake (must be 1 for requeue_pi) * @nr_requeue: number of waiters to requeue (0-INT_MAX) * @cmpval: @uaddr1 expected value (or %NULL) * @requeue_pi: if we are attempting to requeue from a non-pi futex to a |
b41277dc7 futex: Replace fs... |
1916 |
* pi futex (pi to pi requeue is not supported) |
52400ba94 futex: add requeu... |
1917 1918 1919 1920 |
* * Requeue waiters on uaddr1 to uaddr2. In the requeue_pi case, try to acquire * uaddr2 atomically on behalf of the top waiter. * |
6c23cbbd5 futex: fix kernel... |
1921 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1922 1923 |
* - >=0 - on success, the number of tasks requeued or woken; * - <0 - on error |
1da177e4c Linux-2.6.12-rc2 |
1924 |
*/ |
b41277dc7 futex: Replace fs... |
1925 1926 1927 |
static int futex_requeue(u32 __user *uaddr1, unsigned int flags, u32 __user *uaddr2, int nr_wake, int nr_requeue, u32 *cmpval, int requeue_pi) |
1da177e4c Linux-2.6.12-rc2 |
1928 |
{ |
38d47c1b7 futex: rely on ge... |
1929 |
union futex_key key1 = FUTEX_KEY_INIT, key2 = FUTEX_KEY_INIT; |
52400ba94 futex: add requeu... |
1930 1931 |
int drop_count = 0, task_count = 0, ret; struct futex_pi_state *pi_state = NULL; |
e2970f2fb [PATCH] pi-futex:... |
1932 |
struct futex_hash_bucket *hb1, *hb2; |
1da177e4c Linux-2.6.12-rc2 |
1933 |
struct futex_q *this, *next; |
194a6b5b9 sched/wake_q: Ren... |
1934 |
DEFINE_WAKE_Q(wake_q); |
52400ba94 futex: add requeu... |
1935 |
|
fbe0e839d futex: Prevent ov... |
1936 1937 |
if (nr_wake < 0 || nr_requeue < 0) return -EINVAL; |
bc2eecd7e futex: Allow for ... |
1938 1939 1940 1941 1942 1943 1944 1945 |
/* * When PI not supported: return -ENOSYS if requeue_pi is true, * consequently the compiler knows requeue_pi is always false past * this point which will optimize away all the conditional code * further down. */ if (!IS_ENABLED(CONFIG_FUTEX_PI) && requeue_pi) return -ENOSYS; |
52400ba94 futex: add requeu... |
1946 1947 |
if (requeue_pi) { /* |
e9c243a5a futex-prevent-req... |
1948 1949 1950 1951 1952 1953 1954 |
* Requeue PI only works on two distinct uaddrs. This * check is only valid for private futexes. See below. */ if (uaddr1 == uaddr2) return -EINVAL; /* |
52400ba94 futex: add requeu... |
1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 |
* requeue_pi requires a pi_state, try to allocate it now * without any locks in case it fails. */ if (refill_pi_state_cache()) return -ENOMEM; /* * requeue_pi must wake as many tasks as it can, up to nr_wake * + nr_requeue, since it acquires the rt_mutex prior to * returning to userspace, so as to not leave the rt_mutex with * waiters and no owner. However, second and third wake-ups * cannot be predicted as they involve race conditions with the * first wake and a fault while looking up the pi_state. Both * pthread_cond_signal() and pthread_cond_broadcast() should * use nr_wake=1. */ if (nr_wake != 1) return -EINVAL; } |
1da177e4c Linux-2.6.12-rc2 |
1973 |
|
42d35d48c futex: make futex... |
1974 |
retry: |
96d4f267e Remove 'type' arg... |
1975 |
ret = get_futex_key(uaddr1, flags & FLAGS_SHARED, &key1, FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1976 1977 |
if (unlikely(ret != 0)) goto out; |
9ea71503a futex: Fix regres... |
1978 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, |
96d4f267e Remove 'type' arg... |
1979 |
requeue_pi ? FUTEX_WRITE : FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1980 |
if (unlikely(ret != 0)) |
42d35d48c futex: make futex... |
1981 |
goto out_put_key1; |
1da177e4c Linux-2.6.12-rc2 |
1982 |
|
e9c243a5a futex-prevent-req... |
1983 1984 1985 1986 1987 1988 1989 1990 |
/* * The check above which compares uaddrs is not sufficient for * shared futexes. We need to compare the keys: */ if (requeue_pi && match_futex(&key1, &key2)) { ret = -EINVAL; goto out_put_keys; } |
e2970f2fb [PATCH] pi-futex:... |
1991 1992 |
hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); |
1da177e4c Linux-2.6.12-rc2 |
1993 |
|
e4dc5b7a3 futex: clean up f... |
1994 |
retry_private: |
69cd9eba3 futex: avoid race... |
1995 |
hb_waiters_inc(hb2); |
8b8f319fc [PATCH] lockdep: ... |
1996 |
double_lock_hb(hb1, hb2); |
1da177e4c Linux-2.6.12-rc2 |
1997 |
|
e2970f2fb [PATCH] pi-futex:... |
1998 1999 |
if (likely(cmpval != NULL)) { u32 curval; |
1da177e4c Linux-2.6.12-rc2 |
2000 |
|
e2970f2fb [PATCH] pi-futex:... |
2001 |
ret = get_futex_value_locked(&curval, uaddr1); |
1da177e4c Linux-2.6.12-rc2 |
2002 2003 |
if (unlikely(ret)) { |
5eb3dc62f futex: add double... |
2004 |
double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
2005 |
hb_waiters_dec(hb2); |
1da177e4c Linux-2.6.12-rc2 |
2006 |
|
e2970f2fb [PATCH] pi-futex:... |
2007 |
ret = get_user(curval, uaddr1); |
e4dc5b7a3 futex: clean up f... |
2008 2009 |
if (ret) goto out_put_keys; |
1da177e4c Linux-2.6.12-rc2 |
2010 |
|
b41277dc7 futex: Replace fs... |
2011 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
2012 |
goto retry_private; |
1da177e4c Linux-2.6.12-rc2 |
2013 |
|
ae791a2d2 futex: Cleanup st... |
2014 2015 |
put_futex_key(&key2); put_futex_key(&key1); |
e4dc5b7a3 futex: clean up f... |
2016 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
2017 |
} |
e2970f2fb [PATCH] pi-futex:... |
2018 |
if (curval != *cmpval) { |
1da177e4c Linux-2.6.12-rc2 |
2019 2020 2021 2022 |
ret = -EAGAIN; goto out_unlock; } } |
52400ba94 futex: add requeu... |
2023 |
if (requeue_pi && (task_count - nr_wake < nr_requeue)) { |
fc3b55ef2 futex: Prevent ex... |
2024 |
struct task_struct *exiting = NULL; |
bab5bc9e8 futex: fixup unlo... |
2025 2026 2027 2028 2029 2030 |
/* * Attempt to acquire uaddr2 and wake the top waiter. If we * intend to requeue waiters, force setting the FUTEX_WAITERS * bit. We force this here where we are able to easily handle * faults rather in the requeue loop below. */ |
52400ba94 futex: add requeu... |
2031 |
ret = futex_proxy_trylock_atomic(uaddr2, hb1, hb2, &key1, |
fc3b55ef2 futex: Prevent ex... |
2032 2033 |
&key2, &pi_state, &exiting, nr_requeue); |
52400ba94 futex: add requeu... |
2034 2035 2036 2037 2038 |
/* * At this point the top_waiter has either taken uaddr2 or is * waiting on it. If the former, then the pi_state will not * exist yet, look it up one more time to ensure we have a |
866293ee5 futex: Add anothe... |
2039 2040 |
* reference to it. If the lock was taken, ret contains the * vpid of the top waiter task. |
ecb38b78f futex: Document p... |
2041 2042 |
* If the lock was not taken, we have pi_state and an initial * refcount on it. In case of an error we have nothing. |
52400ba94 futex: add requeu... |
2043 |
*/ |
866293ee5 futex: Add anothe... |
2044 |
if (ret > 0) { |
52400ba94 futex: add requeu... |
2045 |
WARN_ON(pi_state); |
89061d3d5 futex: Move drop_... |
2046 |
drop_count++; |
52400ba94 futex: add requeu... |
2047 |
task_count++; |
866293ee5 futex: Add anothe... |
2048 |
/* |
ecb38b78f futex: Document p... |
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 |
* If we acquired the lock, then the user space value * of uaddr2 should be vpid. It cannot be changed by * the top waiter as it is blocked on hb2 lock if it * tries to do so. If something fiddled with it behind * our back the pi state lookup might unearth it. So * we rather use the known value than rereading and * handing potential crap to lookup_pi_state. * * If that call succeeds then we have pi_state and an * initial refcount on it. |
866293ee5 futex: Add anothe... |
2059 |
*/ |
fc3b55ef2 futex: Prevent ex... |
2060 2061 |
ret = lookup_pi_state(uaddr2, ret, hb2, &key2, &pi_state, &exiting); |
52400ba94 futex: add requeu... |
2062 2063 2064 2065 |
} switch (ret) { case 0: |
ecb38b78f futex: Document p... |
2066 |
/* We hold a reference on the pi state. */ |
52400ba94 futex: add requeu... |
2067 |
break; |
4959f2de1 futex: Remove poi... |
2068 2069 |
/* If the above failed, then pi_state is NULL */ |
52400ba94 futex: add requeu... |
2070 2071 |
case -EFAULT: double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
2072 |
hb_waiters_dec(hb2); |
ae791a2d2 futex: Cleanup st... |
2073 2074 |
put_futex_key(&key2); put_futex_key(&key1); |
d0725992c futex: Fix the wr... |
2075 |
ret = fault_in_user_writeable(uaddr2); |
52400ba94 futex: add requeu... |
2076 2077 2078 |
if (!ret) goto retry; goto out; |
56690612a futex: Provide di... |
2079 |
case -EBUSY: |
52400ba94 futex: add requeu... |
2080 |
case -EAGAIN: |
af54d6a1c futex: Simplify f... |
2081 2082 |
/* * Two reasons for this: |
56690612a futex: Provide di... |
2083 |
* - EBUSY: Owner is exiting and we just wait for the |
af54d6a1c futex: Simplify f... |
2084 |
* exit to complete. |
56690612a futex: Provide di... |
2085 |
* - EAGAIN: The user space value changed. |
af54d6a1c futex: Simplify f... |
2086 |
*/ |
52400ba94 futex: add requeu... |
2087 |
double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
2088 |
hb_waiters_dec(hb2); |
ae791a2d2 futex: Cleanup st... |
2089 2090 |
put_futex_key(&key2); put_futex_key(&key1); |
fc3b55ef2 futex: Prevent ex... |
2091 2092 2093 2094 2095 2096 |
/* * Handle the case where the owner is in the middle of * exiting. Wait for the exit to complete otherwise * this task might loop forever, aka. live lock. */ wait_for_owner_exiting(ret, exiting); |
52400ba94 futex: add requeu... |
2097 2098 2099 2100 2101 2102 |
cond_resched(); goto retry; default: goto out_unlock; } } |
0d00c7b20 futexes: Clean up... |
2103 |
plist_for_each_entry_safe(this, next, &hb1->chain, list) { |
52400ba94 futex: add requeu... |
2104 2105 2106 2107 |
if (task_count - nr_wake >= nr_requeue) break; if (!match_futex(&this->key, &key1)) |
1da177e4c Linux-2.6.12-rc2 |
2108 |
continue; |
52400ba94 futex: add requeu... |
2109 |
|
392741e0a futex: Fix handli... |
2110 2111 2112 |
/* * FUTEX_WAIT_REQEUE_PI and FUTEX_CMP_REQUEUE_PI should always * be paired with each other and no other futex ops. |
aa10990e0 futex: avoid wake... |
2113 2114 2115 |
* * We should never be requeueing a futex_q with a pi_state, * which is awaiting a futex_unlock_pi(). |
392741e0a futex: Fix handli... |
2116 2117 |
*/ if ((requeue_pi && !this->rt_waiter) || |
aa10990e0 futex: avoid wake... |
2118 2119 |
(!requeue_pi && this->rt_waiter) || this->pi_state) { |
392741e0a futex: Fix handli... |
2120 2121 2122 |
ret = -EINVAL; break; } |
52400ba94 futex: add requeu... |
2123 2124 2125 2126 2127 2128 2129 |
/* * Wake nr_wake waiters. For requeue_pi, if we acquired the * lock, we already woke the top_waiter. If not, it will be * woken by futex_unlock_pi(). */ if (++task_count <= nr_wake && !requeue_pi) { |
1d0dcb3ad futex: Implement ... |
2130 |
mark_wake_futex(&wake_q, this); |
52400ba94 futex: add requeu... |
2131 2132 |
continue; } |
1da177e4c Linux-2.6.12-rc2 |
2133 |
|
84bc4af59 futex: Detect mis... |
2134 2135 2136 2137 2138 |
/* Ensure we requeue to the expected futex for requeue_pi. */ if (requeue_pi && !match_futex(this->requeue_pi_key, &key2)) { ret = -EINVAL; break; } |
52400ba94 futex: add requeu... |
2139 2140 2141 2142 2143 |
/* * Requeue nr_requeue waiters and possibly one more in the case * of requeue_pi if we couldn't acquire the lock atomically. */ if (requeue_pi) { |
ecb38b78f futex: Document p... |
2144 2145 2146 2147 2148 |
/* * Prepare the waiter to take the rt_mutex. Take a * refcount on the pi_state and store the pointer in * the futex_q object of the waiter. */ |
bf92cf3a5 futex: Cleanup re... |
2149 |
get_pi_state(pi_state); |
52400ba94 futex: add requeu... |
2150 2151 2152 |
this->pi_state = pi_state; ret = rt_mutex_start_proxy_lock(&pi_state->pi_mutex, this->rt_waiter, |
c051b21f7 rtmutex: Confine ... |
2153 |
this->task); |
52400ba94 futex: add requeu... |
2154 |
if (ret == 1) { |
ecb38b78f futex: Document p... |
2155 2156 2157 2158 2159 2160 2161 2162 |
/* * We got the lock. We do neither drop the * refcount on pi_state nor clear * this->pi_state because the waiter needs the * pi_state for cleaning up the user space * value. It will drop the refcount after * doing so. */ |
beda2c7ea futex: Update fut... |
2163 |
requeue_pi_wake_futex(this, &key2, hb2); |
89061d3d5 futex: Move drop_... |
2164 |
drop_count++; |
52400ba94 futex: add requeu... |
2165 2166 |
continue; } else if (ret) { |
ecb38b78f futex: Document p... |
2167 2168 2169 2170 2171 2172 2173 2174 |
/* * rt_mutex_start_proxy_lock() detected a * potential deadlock when we tried to queue * that waiter. Drop the pi_state reference * which we took above and remove the pointer * to the state from the waiters futex_q * object. */ |
52400ba94 futex: add requeu... |
2175 |
this->pi_state = NULL; |
29e9ee5d4 futex: Rename fre... |
2176 |
put_pi_state(pi_state); |
885c2cb77 futex: Cleanup th... |
2177 2178 2179 2180 2181 |
/* * We stop queueing more waiters and let user * space deal with the mess. */ break; |
52400ba94 futex: add requeu... |
2182 |
} |
1da177e4c Linux-2.6.12-rc2 |
2183 |
} |
52400ba94 futex: add requeu... |
2184 2185 |
requeue_futex(this, hb1, hb2, &key2); drop_count++; |
1da177e4c Linux-2.6.12-rc2 |
2186 |
} |
ecb38b78f futex: Document p... |
2187 2188 2189 2190 2191 |
/* * We took an extra initial reference to the pi_state either * in futex_proxy_trylock_atomic() or in lookup_pi_state(). We * need to drop it here again. */ |
29e9ee5d4 futex: Rename fre... |
2192 |
put_pi_state(pi_state); |
885c2cb77 futex: Cleanup th... |
2193 2194 |
out_unlock: |
5eb3dc62f futex: add double... |
2195 |
double_unlock_hb(hb1, hb2); |
1d0dcb3ad futex: Implement ... |
2196 |
wake_up_q(&wake_q); |
69cd9eba3 futex: avoid race... |
2197 |
hb_waiters_dec(hb2); |
1da177e4c Linux-2.6.12-rc2 |
2198 |
|
cd84a42f3 futex: comment re... |
2199 2200 2201 2202 2203 2204 |
/* * drop_futex_key_refs() must be called outside the spinlocks. During * the requeue we moved futex_q's from the hash bucket at key1 to the * one at key2 and updated their key pointer. We no longer need to * hold the references to key1. */ |
1da177e4c Linux-2.6.12-rc2 |
2205 |
while (--drop_count >= 0) |
9adef58b1 futex: get_futex_... |
2206 |
drop_futex_key_refs(&key1); |
1da177e4c Linux-2.6.12-rc2 |
2207 |
|
42d35d48c futex: make futex... |
2208 |
out_put_keys: |
ae791a2d2 futex: Cleanup st... |
2209 |
put_futex_key(&key2); |
42d35d48c futex: make futex... |
2210 |
out_put_key1: |
ae791a2d2 futex: Cleanup st... |
2211 |
put_futex_key(&key1); |
42d35d48c futex: make futex... |
2212 |
out: |
52400ba94 futex: add requeu... |
2213 |
return ret ? ret : task_count; |
1da177e4c Linux-2.6.12-rc2 |
2214 2215 2216 |
} /* The key must be already stored in q->key. */ |
82af7aca5 Removal of FUTEX_FD |
2217 |
static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) |
15e408cd6 futex: Add lock c... |
2218 |
__acquires(&hb->lock) |
1da177e4c Linux-2.6.12-rc2 |
2219 |
{ |
e2970f2fb [PATCH] pi-futex:... |
2220 |
struct futex_hash_bucket *hb; |
1da177e4c Linux-2.6.12-rc2 |
2221 |
|
e2970f2fb [PATCH] pi-futex:... |
2222 |
hb = hash_futex(&q->key); |
11d4616bd futex: revert bac... |
2223 2224 2225 2226 2227 2228 2229 2230 2231 |
/* * Increment the counter before taking the lock so that * a potential waker won't miss a to-be-slept task that is * waiting for the spinlock. This is safe as all queue_lock() * users end up calling queue_me(). Similarly, for housekeeping, * decrement the counter at queue_unlock() when some error has * occurred and we don't end up adding the task to the list. */ |
6f568ebe2 futex: Fix barrie... |
2232 |
hb_waiters_inc(hb); /* implies smp_mb(); (A) */ |
11d4616bd futex: revert bac... |
2233 |
|
e2970f2fb [PATCH] pi-futex:... |
2234 |
q->lock_ptr = &hb->lock; |
1da177e4c Linux-2.6.12-rc2 |
2235 |
|
6f568ebe2 futex: Fix barrie... |
2236 |
spin_lock(&hb->lock); |
e2970f2fb [PATCH] pi-futex:... |
2237 |
return hb; |
1da177e4c Linux-2.6.12-rc2 |
2238 |
} |
d40d65c8d futex: Correct qu... |
2239 |
static inline void |
0d00c7b20 futexes: Clean up... |
2240 |
queue_unlock(struct futex_hash_bucket *hb) |
15e408cd6 futex: Add lock c... |
2241 |
__releases(&hb->lock) |
d40d65c8d futex: Correct qu... |
2242 2243 |
{ spin_unlock(&hb->lock); |
11d4616bd futex: revert bac... |
2244 |
hb_waiters_dec(hb); |
d40d65c8d futex: Correct qu... |
2245 |
} |
cfafcd117 futex: Rework fut... |
2246 |
static inline void __queue_me(struct futex_q *q, struct futex_hash_bucket *hb) |
1da177e4c Linux-2.6.12-rc2 |
2247 |
{ |
ec92d0829 futex priority ba... |
2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 |
int prio; /* * The priority used to register this element is * - either the real thread-priority for the real-time threads * (i.e. threads with a priority lower than MAX_RT_PRIO) * - or MAX_RT_PRIO for non-RT threads. * Thus, all RT-threads are woken first in priority order, and * the others are woken last, in FIFO order. */ prio = min(current->normal_prio, MAX_RT_PRIO); plist_node_init(&q->list, prio); |
ec92d0829 futex priority ba... |
2261 |
plist_add(&q->list, &hb->chain); |
c87e2837b [PATCH] pi-futex:... |
2262 |
q->task = current; |
cfafcd117 futex: Rework fut... |
2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 |
} /** * queue_me() - Enqueue the futex_q on the futex_hash_bucket * @q: The futex_q to enqueue * @hb: The destination hash bucket * * The hb->lock must be held by the caller, and is released here. A call to * queue_me() is typically paired with exactly one call to unqueue_me(). The * exceptions involve the PI related operations, which may use unqueue_me_pi() * or nothing if the unqueue is done as part of the wake process and the unqueue * state is implicit in the state of woken task (see futex_wait_requeue_pi() for * an example). */ static inline void queue_me(struct futex_q *q, struct futex_hash_bucket *hb) __releases(&hb->lock) { __queue_me(q, hb); |
e2970f2fb [PATCH] pi-futex:... |
2281 |
spin_unlock(&hb->lock); |
1da177e4c Linux-2.6.12-rc2 |
2282 |
} |
d40d65c8d futex: Correct qu... |
2283 2284 2285 2286 2287 2288 2289 |
/** * unqueue_me() - Remove the futex_q from its futex_hash_bucket * @q: The futex_q to unqueue * * The q->lock_ptr must not be held by the caller. A call to unqueue_me() must * be paired with exactly one earlier call to queue_me(). * |
6c23cbbd5 futex: fix kernel... |
2290 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2291 2292 |
* - 1 - if the futex_q was still queued (and we removed unqueued it); * - 0 - if the futex_q was already removed by the waking thread |
1da177e4c Linux-2.6.12-rc2 |
2293 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
2294 2295 |
static int unqueue_me(struct futex_q *q) { |
1da177e4c Linux-2.6.12-rc2 |
2296 |
spinlock_t *lock_ptr; |
e2970f2fb [PATCH] pi-futex:... |
2297 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
2298 2299 |
/* In the common case we don't take the spinlock, which is nice. */ |
42d35d48c futex: make futex... |
2300 |
retry: |
29b75eb2d futex: Replace ba... |
2301 2302 2303 2304 2305 2306 |
/* * q->lock_ptr can change between this read and the following spin_lock. * Use READ_ONCE to forbid the compiler from reloading q->lock_ptr and * optimizing lock_ptr out of the logic below. */ lock_ptr = READ_ONCE(q->lock_ptr); |
c80544dc0 sparse pointer us... |
2307 |
if (lock_ptr != NULL) { |
1da177e4c Linux-2.6.12-rc2 |
2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 |
spin_lock(lock_ptr); /* * q->lock_ptr can change between reading it and * spin_lock(), causing us to take the wrong lock. This * corrects the race condition. * * Reasoning goes like this: if we have the wrong lock, * q->lock_ptr must have changed (maybe several times) * between reading it and the spin_lock(). It can * change again after the spin_lock() but only if it was * already changed before the spin_lock(). It cannot, * however, change back to the original value. Therefore * we can detect whether we acquired the correct lock. */ if (unlikely(lock_ptr != q->lock_ptr)) { spin_unlock(lock_ptr); goto retry; } |
2e12978a9 futex,plist: Pass... |
2326 |
__unqueue_futex(q); |
c87e2837b [PATCH] pi-futex:... |
2327 2328 |
BUG_ON(q->pi_state); |
1da177e4c Linux-2.6.12-rc2 |
2329 2330 2331 |
spin_unlock(lock_ptr); ret = 1; } |
9adef58b1 futex: get_futex_... |
2332 |
drop_futex_key_refs(&q->key); |
1da177e4c Linux-2.6.12-rc2 |
2333 2334 |
return ret; } |
c87e2837b [PATCH] pi-futex:... |
2335 2336 |
/* * PI futexes can not be requeued and must remove themself from the |
d0aa7a70b futex_requeue_pi ... |
2337 2338 |
* hash bucket. The hash bucket lock (i.e. lock_ptr) is held on entry * and dropped here. |
c87e2837b [PATCH] pi-futex:... |
2339 |
*/ |
d0aa7a70b futex_requeue_pi ... |
2340 |
static void unqueue_me_pi(struct futex_q *q) |
15e408cd6 futex: Add lock c... |
2341 |
__releases(q->lock_ptr) |
c87e2837b [PATCH] pi-futex:... |
2342 |
{ |
2e12978a9 futex,plist: Pass... |
2343 |
__unqueue_futex(q); |
c87e2837b [PATCH] pi-futex:... |
2344 2345 |
BUG_ON(!q->pi_state); |
29e9ee5d4 futex: Rename fre... |
2346 |
put_pi_state(q->pi_state); |
c87e2837b [PATCH] pi-futex:... |
2347 |
q->pi_state = NULL; |
d0aa7a70b futex_requeue_pi ... |
2348 |
spin_unlock(q->lock_ptr); |
c87e2837b [PATCH] pi-futex:... |
2349 |
} |
778e9a9c3 pi-futex: fix exi... |
2350 |
static int fixup_pi_state_owner(u32 __user *uaddr, struct futex_q *q, |
c1e2f0eaf futex: Avoid viol... |
2351 |
struct task_struct *argowner) |
d0aa7a70b futex_requeue_pi ... |
2352 |
{ |
d0aa7a70b futex_requeue_pi ... |
2353 |
struct futex_pi_state *pi_state = q->pi_state; |
7cfdaf38d futex: uninitiali... |
2354 |
u32 uval, uninitialized_var(curval), newval; |
c1e2f0eaf futex: Avoid viol... |
2355 2356 |
struct task_struct *oldowner, *newowner; u32 newtid; |
6b4f4bc9c locking/futex: Al... |
2357 |
int ret, err = 0; |
d0aa7a70b futex_requeue_pi ... |
2358 |
|
c1e2f0eaf futex: Avoid viol... |
2359 |
lockdep_assert_held(q->lock_ptr); |
734009e96 futex: Change loc... |
2360 2361 2362 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); oldowner = pi_state->owner; |
1b7558e45 futexes: fix faul... |
2363 2364 |
/* |
c1e2f0eaf futex: Avoid viol... |
2365 |
* We are here because either: |
16ffa12d7 futex: Pull rt_mu... |
2366 |
* |
c1e2f0eaf futex: Avoid viol... |
2367 2368 2369 2370 2371 2372 2373 2374 2375 |
* - we stole the lock and pi_state->owner needs updating to reflect * that (@argowner == current), * * or: * * - someone stole our lock and we need to fix things to point to the * new owner (@argowner == NULL). * * Either way, we have to replace the TID in the user space variable. |
8161239a8 rtmutex: Simplify... |
2376 |
* This must be atomic as we have to preserve the owner died bit here. |
1b7558e45 futexes: fix faul... |
2377 |
* |
b2d0994b1 futex: update fut... |
2378 2379 2380 |
* Note: We write the user space value _before_ changing the pi_state * because we can fault here. Imagine swapped out pages or a fork * that marked all the anonymous memory readonly for cow. |
1b7558e45 futexes: fix faul... |
2381 |
* |
734009e96 futex: Change loc... |
2382 2383 2384 2385 |
* Modifying pi_state _before_ the user space value would leave the * pi_state in an inconsistent state when we fault here, because we * need to drop the locks to handle the fault. This might be observed * in the PID check in lookup_pi_state. |
1b7558e45 futexes: fix faul... |
2386 2387 |
*/ retry: |
c1e2f0eaf futex: Avoid viol... |
2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 |
if (!argowner) { if (oldowner != current) { /* * We raced against a concurrent self; things are * already fixed up. Nothing to do. */ ret = 0; goto out_unlock; } if (__rt_mutex_futex_trylock(&pi_state->pi_mutex)) { /* We got the lock after all, nothing to fix. */ ret = 0; goto out_unlock; } /* * Since we just failed the trylock; there must be an owner. */ newowner = rt_mutex_owner(&pi_state->pi_mutex); BUG_ON(!newowner); } else { WARN_ON_ONCE(argowner != current); if (oldowner == current) { /* * We raced against a concurrent self; things are * already fixed up. Nothing to do. */ ret = 0; goto out_unlock; } newowner = argowner; } newtid = task_pid_vnr(newowner) | FUTEX_WAITERS; |
a97cb0e7b futex: Fix OWNER_... |
2423 2424 2425 |
/* Owner died? */ if (!pi_state->owner) newtid |= FUTEX_OWNER_DIED; |
c1e2f0eaf futex: Avoid viol... |
2426 |
|
6b4f4bc9c locking/futex: Al... |
2427 2428 2429 |
err = get_futex_value_locked(&uval, uaddr); if (err) goto handle_err; |
1b7558e45 futexes: fix faul... |
2430 |
|
16ffa12d7 futex: Pull rt_mu... |
2431 |
for (;;) { |
1b7558e45 futexes: fix faul... |
2432 |
newval = (uval & FUTEX_OWNER_DIED) | newtid; |
6b4f4bc9c locking/futex: Al... |
2433 2434 2435 |
err = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (err) goto handle_err; |
1b7558e45 futexes: fix faul... |
2436 2437 2438 2439 2440 2441 2442 2443 2444 |
if (curval == uval) break; uval = curval; } /* * We fixed up user space. Now we need to fix the pi_state * itself. */ |
d0aa7a70b futex_requeue_pi ... |
2445 |
if (pi_state->owner != NULL) { |
734009e96 futex: Change loc... |
2446 |
raw_spin_lock(&pi_state->owner->pi_lock); |
d0aa7a70b futex_requeue_pi ... |
2447 2448 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
734009e96 futex: Change loc... |
2449 |
raw_spin_unlock(&pi_state->owner->pi_lock); |
1b7558e45 futexes: fix faul... |
2450 |
} |
d0aa7a70b futex_requeue_pi ... |
2451 |
|
cdf71a10c futex: Prevent st... |
2452 |
pi_state->owner = newowner; |
d0aa7a70b futex_requeue_pi ... |
2453 |
|
734009e96 futex: Change loc... |
2454 |
raw_spin_lock(&newowner->pi_lock); |
d0aa7a70b futex_requeue_pi ... |
2455 |
WARN_ON(!list_empty(&pi_state->list)); |
cdf71a10c futex: Prevent st... |
2456 |
list_add(&pi_state->list, &newowner->pi_state_list); |
734009e96 futex: Change loc... |
2457 2458 |
raw_spin_unlock(&newowner->pi_lock); raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
1b7558e45 futexes: fix faul... |
2459 |
return 0; |
d0aa7a70b futex_requeue_pi ... |
2460 |
|
d0aa7a70b futex_requeue_pi ... |
2461 |
/* |
6b4f4bc9c locking/futex: Al... |
2462 2463 2464 2465 2466 2467 2468 |
* In order to reschedule or handle a page fault, we need to drop the * locks here. In the case of a fault, this gives the other task * (either the highest priority waiter itself or the task which stole * the rtmutex) the chance to try the fixup of the pi_state. So once we * are back from handling the fault we need to check the pi_state after * reacquiring the locks and before trying to do another fixup. When * the fixup has been done already we simply return. |
734009e96 futex: Change loc... |
2469 2470 2471 2472 |
* * Note: we hold both hb->lock and pi_mutex->wait_lock. We can safely * drop hb->lock since the caller owns the hb -> futex_q relation. * Dropping the pi_mutex->wait_lock requires the state revalidate. |
d0aa7a70b futex_requeue_pi ... |
2473 |
*/ |
6b4f4bc9c locking/futex: Al... |
2474 |
handle_err: |
734009e96 futex: Change loc... |
2475 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
1b7558e45 futexes: fix faul... |
2476 |
spin_unlock(q->lock_ptr); |
778e9a9c3 pi-futex: fix exi... |
2477 |
|
6b4f4bc9c locking/futex: Al... |
2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 |
switch (err) { case -EFAULT: ret = fault_in_user_writeable(uaddr); break; case -EAGAIN: cond_resched(); ret = 0; break; default: WARN_ON_ONCE(1); ret = err; break; } |
778e9a9c3 pi-futex: fix exi... |
2493 |
|
1b7558e45 futexes: fix faul... |
2494 |
spin_lock(q->lock_ptr); |
734009e96 futex: Change loc... |
2495 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); |
778e9a9c3 pi-futex: fix exi... |
2496 |
|
1b7558e45 futexes: fix faul... |
2497 2498 2499 |
/* * Check if someone else fixed it for us: */ |
734009e96 futex: Change loc... |
2500 2501 2502 2503 |
if (pi_state->owner != oldowner) { ret = 0; goto out_unlock; } |
1b7558e45 futexes: fix faul... |
2504 2505 |
if (ret) |
734009e96 futex: Change loc... |
2506 |
goto out_unlock; |
1b7558e45 futexes: fix faul... |
2507 2508 |
goto retry; |
734009e96 futex: Change loc... |
2509 2510 2511 2512 |
out_unlock: raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); return ret; |
d0aa7a70b futex_requeue_pi ... |
2513 |
} |
72c1bbf30 futex: restartabl... |
2514 |
static long futex_wait_restart(struct restart_block *restart); |
36cf3b5c3 FUTEX: Tidy up th... |
2515 |
|
ca5f9524d futex: separate f... |
2516 |
/** |
dd9739980 futex: split out ... |
2517 2518 |
* fixup_owner() - Post lock pi_state and corner case management * @uaddr: user address of the futex |
dd9739980 futex: split out ... |
2519 2520 2521 2522 2523 2524 2525 |
* @q: futex_q (contains pi_state and access to the rt_mutex) * @locked: if the attempt to take the rt_mutex succeeded (1) or not (0) * * After attempting to lock an rt_mutex, this function is called to cleanup * the pi_state owner as well as handle race conditions that may allow us to * acquire the lock. Must be called with the hb lock held. * |
6c23cbbd5 futex: fix kernel... |
2526 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2527 2528 2529 |
* - 1 - success, lock taken; * - 0 - success, lock not taken; * - <0 - on error (-EFAULT) |
dd9739980 futex: split out ... |
2530 |
*/ |
ae791a2d2 futex: Cleanup st... |
2531 |
static int fixup_owner(u32 __user *uaddr, struct futex_q *q, int locked) |
dd9739980 futex: split out ... |
2532 |
{ |
dd9739980 futex: split out ... |
2533 2534 2535 2536 2537 2538 |
int ret = 0; if (locked) { /* * Got the lock. We might not be the anticipated owner if we * did a lock-steal - fix up the PI-state in that case: |
16ffa12d7 futex: Pull rt_mu... |
2539 |
* |
c1e2f0eaf futex: Avoid viol... |
2540 2541 2542 |
* Speculative pi_state->owner read (we don't hold wait_lock); * since we own the lock pi_state->owner == current is the * stable state, anything else needs more attention. |
dd9739980 futex: split out ... |
2543 2544 |
*/ if (q->pi_state->owner != current) |
ae791a2d2 futex: Cleanup st... |
2545 |
ret = fixup_pi_state_owner(uaddr, q, current); |
dd9739980 futex: split out ... |
2546 2547 2548 2549 |
goto out; } /* |
c1e2f0eaf futex: Avoid viol... |
2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 |
* If we didn't get the lock; check if anybody stole it from us. In * that case, we need to fix up the uval to point to them instead of * us, otherwise bad things happen. [10] * * Another speculative read; pi_state->owner == current is unstable * but needs our attention. */ if (q->pi_state->owner == current) { ret = fixup_pi_state_owner(uaddr, q, NULL); goto out; } /* |
dd9739980 futex: split out ... |
2563 |
* Paranoia check. If we did not take the lock, then we should not be |
8161239a8 rtmutex: Simplify... |
2564 |
* the owner of the rt_mutex. |
dd9739980 futex: split out ... |
2565 |
*/ |
73d786bd0 futex: Rework inc... |
2566 |
if (rt_mutex_owner(&q->pi_state->pi_mutex) == current) { |
dd9739980 futex: split out ... |
2567 2568 2569 2570 2571 |
printk(KERN_ERR "fixup_owner: ret = %d pi-mutex: %p " "pi-state %p ", ret, q->pi_state->pi_mutex.owner, q->pi_state->owner); |
73d786bd0 futex: Rework inc... |
2572 |
} |
dd9739980 futex: split out ... |
2573 2574 2575 2576 2577 2578 |
out: return ret ? ret : locked; } /** |
ca5f9524d futex: separate f... |
2579 2580 2581 2582 |
* futex_wait_queue_me() - queue_me() and wait for wakeup, timeout, or signal * @hb: the futex hash bucket, must be locked by the caller * @q: the futex_q to queue up on * @timeout: the prepared hrtimer_sleeper, or null for no timeout |
ca5f9524d futex: separate f... |
2583 2584 |
*/ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, |
f1a11e057 futex: remove the... |
2585 |
struct hrtimer_sleeper *timeout) |
ca5f9524d futex: separate f... |
2586 |
{ |
9beba3c54 futex: Add memory... |
2587 2588 |
/* * The task state is guaranteed to be set before another task can |
b92b8b35a locking/arch: Ren... |
2589 |
* wake it. set_current_state() is implemented using smp_store_mb() and |
9beba3c54 futex: Add memory... |
2590 2591 2592 |
* queue_me() calls spin_unlock() upon completion, both serializing * access to the hash list and forcing another memory barrier. */ |
f1a11e057 futex: remove the... |
2593 |
set_current_state(TASK_INTERRUPTIBLE); |
0729e1961 futex: Fix wakeup... |
2594 |
queue_me(q, hb); |
ca5f9524d futex: separate f... |
2595 2596 |
/* Arm the timer */ |
2e4b0d3fe futex: Remove bog... |
2597 |
if (timeout) |
9dd8813ed hrtimer/treewide:... |
2598 |
hrtimer_sleeper_start_expires(timeout, HRTIMER_MODE_ABS); |
ca5f9524d futex: separate f... |
2599 2600 |
/* |
0729e1961 futex: Fix wakeup... |
2601 2602 |
* If we have been removed from the hash list, then another task * has tried to wake us, and we can skip the call to schedule(). |
ca5f9524d futex: separate f... |
2603 2604 2605 2606 2607 2608 2609 2610 |
*/ if (likely(!plist_node_empty(&q->list))) { /* * If the timer has already expired, current will already be * flagged for rescheduling. Only call schedule if there * is no timeout, or if it has yet to expire. */ if (!timeout || timeout->task) |
88c8004fd futex: Use freeza... |
2611 |
freezable_schedule(); |
ca5f9524d futex: separate f... |
2612 2613 2614 |
} __set_current_state(TASK_RUNNING); } |
f801073f8 futex: split out ... |
2615 2616 2617 2618 |
/** * futex_wait_setup() - Prepare to wait on a futex * @uaddr: the futex userspace address * @val: the expected value |
b41277dc7 futex: Replace fs... |
2619 |
* @flags: futex flags (FLAGS_SHARED, etc.) |
f801073f8 futex: split out ... |
2620 2621 2622 2623 2624 2625 2626 2627 |
* @q: the associated futex_q * @hb: storage for hash_bucket pointer to be returned to caller * * Setup the futex_q and locate the hash_bucket. Get the futex value and * compare it with the expected value. Handle atomic faults internally. * Return with the hb lock held and a q.key reference on success, and unlocked * with no q.key reference on failure. * |
6c23cbbd5 futex: fix kernel... |
2628 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2629 2630 |
* - 0 - uaddr contains val and hb has been locked; * - <1 - -EFAULT or -EWOULDBLOCK (uaddr does not contain val) and hb is unlocked |
f801073f8 futex: split out ... |
2631 |
*/ |
b41277dc7 futex: Replace fs... |
2632 |
static int futex_wait_setup(u32 __user *uaddr, u32 val, unsigned int flags, |
f801073f8 futex: split out ... |
2633 |
struct futex_q *q, struct futex_hash_bucket **hb) |
1da177e4c Linux-2.6.12-rc2 |
2634 |
{ |
e2970f2fb [PATCH] pi-futex:... |
2635 2636 |
u32 uval; int ret; |
1da177e4c Linux-2.6.12-rc2 |
2637 |
|
1da177e4c Linux-2.6.12-rc2 |
2638 |
/* |
b2d0994b1 futex: update fut... |
2639 |
* Access the page AFTER the hash-bucket is locked. |
1da177e4c Linux-2.6.12-rc2 |
2640 2641 2642 2643 2644 2645 2646 |
* Order is important: * * Userspace waiter: val = var; if (cond(val)) futex_wait(&var, val); * Userspace waker: if (cond(var)) { var = new; futex_wake(&var); } * * The basic logical guarantee of a futex is that it blocks ONLY * if cond(var) is known to be true at the time of blocking, for |
8fe8f545c futex: Update fut... |
2647 2648 |
* any cond. If we locked the hash-bucket after testing *uaddr, that * would open a race condition where we could block indefinitely with |
1da177e4c Linux-2.6.12-rc2 |
2649 2650 |
* cond(var) false, which would violate the guarantee. * |
8fe8f545c futex: Update fut... |
2651 2652 2653 2654 |
* On the other hand, we insert q and release the hash-bucket only * after testing *uaddr. This guarantees that futex_wait() will NOT * absorb a wakeup if *uaddr does not match the desired values * while the syscall executes. |
1da177e4c Linux-2.6.12-rc2 |
2655 |
*/ |
f801073f8 futex: split out ... |
2656 |
retry: |
96d4f267e Remove 'type' arg... |
2657 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &q->key, FUTEX_READ); |
f801073f8 futex: split out ... |
2658 |
if (unlikely(ret != 0)) |
a5a2a0c7f futex: fix futex_... |
2659 |
return ret; |
f801073f8 futex: split out ... |
2660 2661 2662 |
retry_private: *hb = queue_lock(q); |
e2970f2fb [PATCH] pi-futex:... |
2663 |
ret = get_futex_value_locked(&uval, uaddr); |
1da177e4c Linux-2.6.12-rc2 |
2664 |
|
f801073f8 futex: split out ... |
2665 |
if (ret) { |
0d00c7b20 futexes: Clean up... |
2666 |
queue_unlock(*hb); |
1da177e4c Linux-2.6.12-rc2 |
2667 |
|
e2970f2fb [PATCH] pi-futex:... |
2668 |
ret = get_user(uval, uaddr); |
e4dc5b7a3 futex: clean up f... |
2669 |
if (ret) |
f801073f8 futex: split out ... |
2670 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
2671 |
|
b41277dc7 futex: Replace fs... |
2672 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
2673 |
goto retry_private; |
ae791a2d2 futex: Cleanup st... |
2674 |
put_futex_key(&q->key); |
e4dc5b7a3 futex: clean up f... |
2675 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
2676 |
} |
ca5f9524d futex: separate f... |
2677 |
|
f801073f8 futex: split out ... |
2678 |
if (uval != val) { |
0d00c7b20 futexes: Clean up... |
2679 |
queue_unlock(*hb); |
f801073f8 futex: split out ... |
2680 |
ret = -EWOULDBLOCK; |
2fff78c78 futex: fix refere... |
2681 |
} |
1da177e4c Linux-2.6.12-rc2 |
2682 |
|
f801073f8 futex: split out ... |
2683 2684 |
out: if (ret) |
ae791a2d2 futex: Cleanup st... |
2685 |
put_futex_key(&q->key); |
f801073f8 futex: split out ... |
2686 2687 |
return ret; } |
b41277dc7 futex: Replace fs... |
2688 2689 |
static int futex_wait(u32 __user *uaddr, unsigned int flags, u32 val, ktime_t *abs_time, u32 bitset) |
f801073f8 futex: split out ... |
2690 |
{ |
5ca584d93 futex: Consolidat... |
2691 |
struct hrtimer_sleeper timeout, *to; |
f801073f8 futex: split out ... |
2692 2693 |
struct restart_block *restart; struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
2694 |
struct futex_q q = futex_q_init; |
f801073f8 futex: split out ... |
2695 2696 2697 2698 |
int ret; if (!bitset) return -EINVAL; |
f801073f8 futex: split out ... |
2699 |
q.bitset = bitset; |
5ca584d93 futex: Consolidat... |
2700 2701 |
to = futex_setup_timer(abs_time, &timeout, flags, current->timer_slack_ns); |
d58e6576b futex: Handle spu... |
2702 |
retry: |
7ada876a8 futex: Fix errors... |
2703 2704 2705 2706 |
/* * Prepare to wait on uaddr. On success, holds hb lock and increments * q.key refs. */ |
b41277dc7 futex: Replace fs... |
2707 |
ret = futex_wait_setup(uaddr, val, flags, &q, &hb); |
f801073f8 futex: split out ... |
2708 2709 |
if (ret) goto out; |
ca5f9524d futex: separate f... |
2710 |
/* queue_me and wait for wakeup, timeout, or a signal. */ |
f1a11e057 futex: remove the... |
2711 |
futex_wait_queue_me(hb, &q, to); |
1da177e4c Linux-2.6.12-rc2 |
2712 2713 |
/* If we were woken (and unqueued), we succeeded, whatever. */ |
2fff78c78 futex: fix refere... |
2714 |
ret = 0; |
7ada876a8 futex: Fix errors... |
2715 |
/* unqueue_me() drops q.key ref */ |
1da177e4c Linux-2.6.12-rc2 |
2716 |
if (!unqueue_me(&q)) |
7ada876a8 futex: Fix errors... |
2717 |
goto out; |
2fff78c78 futex: fix refere... |
2718 |
ret = -ETIMEDOUT; |
ca5f9524d futex: separate f... |
2719 |
if (to && !to->task) |
7ada876a8 futex: Fix errors... |
2720 |
goto out; |
72c1bbf30 futex: restartabl... |
2721 |
|
e2970f2fb [PATCH] pi-futex:... |
2722 |
/* |
d58e6576b futex: Handle spu... |
2723 2724 |
* We expect signal_pending(current), but we might be the * victim of a spurious wakeup as well. |
e2970f2fb [PATCH] pi-futex:... |
2725 |
*/ |
7ada876a8 futex: Fix errors... |
2726 |
if (!signal_pending(current)) |
d58e6576b futex: Handle spu... |
2727 |
goto retry; |
d58e6576b futex: Handle spu... |
2728 |
|
2fff78c78 futex: fix refere... |
2729 |
ret = -ERESTARTSYS; |
c19384b5b Make futex_wait()... |
2730 |
if (!abs_time) |
7ada876a8 futex: Fix errors... |
2731 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
2732 |
|
f56141e3e all arches, signa... |
2733 |
restart = ¤t->restart_block; |
2fff78c78 futex: fix refere... |
2734 |
restart->fn = futex_wait_restart; |
a3c74c525 futex: Mark resta... |
2735 |
restart->futex.uaddr = uaddr; |
2fff78c78 futex: fix refere... |
2736 |
restart->futex.val = val; |
2456e8553 ktime: Get rid of... |
2737 |
restart->futex.time = *abs_time; |
2fff78c78 futex: fix refere... |
2738 |
restart->futex.bitset = bitset; |
0cd9c6494 futex: Set FLAGS_... |
2739 |
restart->futex.flags = flags | FLAGS_HAS_TIMEOUT; |
42d35d48c futex: make futex... |
2740 |
|
2fff78c78 futex: fix refere... |
2741 |
ret = -ERESTART_RESTARTBLOCK; |
42d35d48c futex: make futex... |
2742 |
out: |
ca5f9524d futex: separate f... |
2743 2744 2745 2746 |
if (to) { hrtimer_cancel(&to->timer); destroy_hrtimer_on_stack(&to->timer); } |
c87e2837b [PATCH] pi-futex:... |
2747 2748 |
return ret; } |
72c1bbf30 futex: restartabl... |
2749 2750 2751 |
static long futex_wait_restart(struct restart_block *restart) { |
a3c74c525 futex: Mark resta... |
2752 |
u32 __user *uaddr = restart->futex.uaddr; |
a72188d8a futex: add FUTEX_... |
2753 |
ktime_t t, *tp = NULL; |
72c1bbf30 futex: restartabl... |
2754 |
|
a72188d8a futex: add FUTEX_... |
2755 |
if (restart->futex.flags & FLAGS_HAS_TIMEOUT) { |
2456e8553 ktime: Get rid of... |
2756 |
t = restart->futex.time; |
a72188d8a futex: add FUTEX_... |
2757 2758 |
tp = &t; } |
72c1bbf30 futex: restartabl... |
2759 |
restart->fn = do_no_restart_syscall; |
b41277dc7 futex: Replace fs... |
2760 2761 2762 |
return (long)futex_wait(uaddr, restart->futex.flags, restart->futex.val, tp, restart->futex.bitset); |
72c1bbf30 futex: restartabl... |
2763 |
} |
c87e2837b [PATCH] pi-futex:... |
2764 2765 2766 |
/* * Userspace tried a 0 -> TID atomic transition of the futex value * and failed. The kernel side here does the whole locking operation: |
767f509ca futex: Enhance co... |
2767 2768 2769 2770 2771 |
* if there are waiters then it will block as a consequence of relying * on rt-mutexes, it does PI, etc. (Due to races the kernel might see * a 0 value of the futex too.). * * Also serves as futex trylock_pi()'ing, and due semantics. |
c87e2837b [PATCH] pi-futex:... |
2772 |
*/ |
996636dda futex: Fix argume... |
2773 |
static int futex_lock_pi(u32 __user *uaddr, unsigned int flags, |
b41277dc7 futex: Replace fs... |
2774 |
ktime_t *time, int trylock) |
c87e2837b [PATCH] pi-futex:... |
2775 |
{ |
5ca584d93 futex: Consolidat... |
2776 |
struct hrtimer_sleeper timeout, *to; |
16ffa12d7 futex: Pull rt_mu... |
2777 |
struct futex_pi_state *pi_state = NULL; |
fc3b55ef2 futex: Prevent ex... |
2778 |
struct task_struct *exiting = NULL; |
cfafcd117 futex: Rework fut... |
2779 |
struct rt_mutex_waiter rt_waiter; |
c87e2837b [PATCH] pi-futex:... |
2780 |
struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
2781 |
struct futex_q q = futex_q_init; |
dd9739980 futex: split out ... |
2782 |
int res, ret; |
c87e2837b [PATCH] pi-futex:... |
2783 |
|
bc2eecd7e futex: Allow for ... |
2784 2785 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
c87e2837b [PATCH] pi-futex:... |
2786 2787 |
if (refill_pi_state_cache()) return -ENOMEM; |
5ca584d93 futex: Consolidat... |
2788 |
to = futex_setup_timer(time, &timeout, FLAGS_CLOCKRT, 0); |
c5780e976 [PATCH] Use the c... |
2789 |
|
42d35d48c futex: make futex... |
2790 |
retry: |
96d4f267e Remove 'type' arg... |
2791 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &q.key, FUTEX_WRITE); |
c87e2837b [PATCH] pi-futex:... |
2792 |
if (unlikely(ret != 0)) |
42d35d48c futex: make futex... |
2793 |
goto out; |
c87e2837b [PATCH] pi-futex:... |
2794 |
|
e4dc5b7a3 futex: clean up f... |
2795 |
retry_private: |
82af7aca5 Removal of FUTEX_FD |
2796 |
hb = queue_lock(&q); |
c87e2837b [PATCH] pi-futex:... |
2797 |
|
fc3b55ef2 futex: Prevent ex... |
2798 2799 |
ret = futex_lock_pi_atomic(uaddr, hb, &q.key, &q.pi_state, current, &exiting, 0); |
c87e2837b [PATCH] pi-futex:... |
2800 |
if (unlikely(ret)) { |
767f509ca futex: Enhance co... |
2801 2802 2803 2804 |
/* * Atomic work succeeded and we got the lock, * or failed. Either way, we do _not_ block. */ |
778e9a9c3 pi-futex: fix exi... |
2805 |
switch (ret) { |
1a52084d0 futex: split out ... |
2806 2807 2808 2809 2810 2811 |
case 1: /* We got the lock. */ ret = 0; goto out_unlock_put_key; case -EFAULT: goto uaddr_faulted; |
56690612a futex: Provide di... |
2812 |
case -EBUSY: |
778e9a9c3 pi-futex: fix exi... |
2813 2814 |
case -EAGAIN: /* |
af54d6a1c futex: Simplify f... |
2815 |
* Two reasons for this: |
56690612a futex: Provide di... |
2816 |
* - EBUSY: Task is exiting and we just wait for the |
af54d6a1c futex: Simplify f... |
2817 |
* exit to complete. |
56690612a futex: Provide di... |
2818 |
* - EAGAIN: The user space value changed. |
778e9a9c3 pi-futex: fix exi... |
2819 |
*/ |
0d00c7b20 futexes: Clean up... |
2820 |
queue_unlock(hb); |
ae791a2d2 futex: Cleanup st... |
2821 |
put_futex_key(&q.key); |
fc3b55ef2 futex: Prevent ex... |
2822 2823 2824 2825 2826 2827 |
/* * Handle the case where the owner is in the middle of * exiting. Wait for the exit to complete otherwise * this task might loop forever, aka. live lock. */ wait_for_owner_exiting(ret, exiting); |
778e9a9c3 pi-futex: fix exi... |
2828 2829 |
cond_resched(); goto retry; |
778e9a9c3 pi-futex: fix exi... |
2830 |
default: |
42d35d48c futex: make futex... |
2831 |
goto out_unlock_put_key; |
c87e2837b [PATCH] pi-futex:... |
2832 |
} |
c87e2837b [PATCH] pi-futex:... |
2833 |
} |
cfafcd117 futex: Rework fut... |
2834 |
WARN_ON(!q.pi_state); |
c87e2837b [PATCH] pi-futex:... |
2835 2836 2837 |
/* * Only actually queue now that the atomic ops are done: */ |
cfafcd117 futex: Rework fut... |
2838 |
__queue_me(&q, hb); |
c87e2837b [PATCH] pi-futex:... |
2839 |
|
cfafcd117 futex: Rework fut... |
2840 |
if (trylock) { |
5293c2efd futex,rt_mutex: P... |
2841 |
ret = rt_mutex_futex_trylock(&q.pi_state->pi_mutex); |
c87e2837b [PATCH] pi-futex:... |
2842 2843 |
/* Fixup the trylock return value: */ ret = ret ? 0 : -EWOULDBLOCK; |
cfafcd117 futex: Rework fut... |
2844 |
goto no_block; |
c87e2837b [PATCH] pi-futex:... |
2845 |
} |
56222b212 futex: Drop hb->l... |
2846 |
rt_mutex_init_waiter(&rt_waiter); |
cfafcd117 futex: Rework fut... |
2847 |
/* |
56222b212 futex: Drop hb->l... |
2848 2849 2850 2851 2852 2853 2854 |
* On PREEMPT_RT_FULL, when hb->lock becomes an rt_mutex, we must not * hold it while doing rt_mutex_start_proxy(), because then it will * include hb->lock in the blocking chain, even through we'll not in * fact hold it while blocking. This will lead it to report -EDEADLK * and BUG when futex_unlock_pi() interleaves with this. * * Therefore acquire wait_lock while holding hb->lock, but drop the |
1a1fb985f futex: Handle ear... |
2855 2856 2857 2858 |
* latter before calling __rt_mutex_start_proxy_lock(). This * interleaves with futex_unlock_pi() -- which does a similar lock * handoff -- such that the latter can observe the futex_q::pi_state * before __rt_mutex_start_proxy_lock() is done. |
cfafcd117 futex: Rework fut... |
2859 |
*/ |
56222b212 futex: Drop hb->l... |
2860 2861 |
raw_spin_lock_irq(&q.pi_state->pi_mutex.wait_lock); spin_unlock(q.lock_ptr); |
1a1fb985f futex: Handle ear... |
2862 2863 2864 2865 2866 |
/* * __rt_mutex_start_proxy_lock() unconditionally enqueues the @rt_waiter * such that futex_unlock_pi() is guaranteed to observe the waiter when * it sees the futex_q::pi_state. */ |
56222b212 futex: Drop hb->l... |
2867 2868 |
ret = __rt_mutex_start_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter, current); raw_spin_unlock_irq(&q.pi_state->pi_mutex.wait_lock); |
cfafcd117 futex: Rework fut... |
2869 2870 2871 |
if (ret) { if (ret == 1) ret = 0; |
1a1fb985f futex: Handle ear... |
2872 |
goto cleanup; |
cfafcd117 futex: Rework fut... |
2873 |
} |
cfafcd117 futex: Rework fut... |
2874 |
if (unlikely(to)) |
9dd8813ed hrtimer/treewide:... |
2875 |
hrtimer_sleeper_start_expires(to, HRTIMER_MODE_ABS); |
cfafcd117 futex: Rework fut... |
2876 2877 |
ret = rt_mutex_wait_proxy_lock(&q.pi_state->pi_mutex, to, &rt_waiter); |
1a1fb985f futex: Handle ear... |
2878 |
cleanup: |
a99e4e413 [PATCH] pi-futex:... |
2879 |
spin_lock(q.lock_ptr); |
dd9739980 futex: split out ... |
2880 |
/* |
1a1fb985f futex: Handle ear... |
2881 |
* If we failed to acquire the lock (deadlock/signal/timeout), we must |
cfafcd117 futex: Rework fut... |
2882 |
* first acquire the hb->lock before removing the lock from the |
1a1fb985f futex: Handle ear... |
2883 2884 |
* rt_mutex waitqueue, such that we can keep the hb and rt_mutex wait * lists consistent. |
56222b212 futex: Drop hb->l... |
2885 2886 2887 |
* * In particular; it is important that futex_unlock_pi() can not * observe this inconsistency. |
cfafcd117 futex: Rework fut... |
2888 2889 2890 2891 2892 2893 |
*/ if (ret && !rt_mutex_cleanup_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter)) ret = 0; no_block: /* |
dd9739980 futex: split out ... |
2894 2895 2896 |
* Fixup the pi_state owner and possibly acquire the lock if we * haven't already. */ |
ae791a2d2 futex: Cleanup st... |
2897 |
res = fixup_owner(uaddr, &q, !ret); |
dd9739980 futex: split out ... |
2898 2899 2900 2901 2902 2903 |
/* * If fixup_owner() returned an error, proprogate that. If it acquired * the lock, clear our -ETIMEDOUT or -EINTR. */ if (res) ret = (res < 0) ? res : 0; |
c87e2837b [PATCH] pi-futex:... |
2904 |
|
e8f6386c0 futex: unlock bef... |
2905 |
/* |
dd9739980 futex: split out ... |
2906 2907 |
* If fixup_owner() faulted and was unable to handle the fault, unlock * it and return the fault to userspace. |
e8f6386c0 futex: unlock bef... |
2908 |
*/ |
16ffa12d7 futex: Pull rt_mu... |
2909 2910 2911 2912 |
if (ret && (rt_mutex_owner(&q.pi_state->pi_mutex) == current)) { pi_state = q.pi_state; get_pi_state(pi_state); } |
e8f6386c0 futex: unlock bef... |
2913 |
|
778e9a9c3 pi-futex: fix exi... |
2914 2915 |
/* Unqueue and drop the lock */ unqueue_me_pi(&q); |
c87e2837b [PATCH] pi-futex:... |
2916 |
|
16ffa12d7 futex: Pull rt_mu... |
2917 2918 2919 2920 |
if (pi_state) { rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); } |
5ecb01cfd futex_lock_pi() k... |
2921 |
goto out_put_key; |
c87e2837b [PATCH] pi-futex:... |
2922 |
|
42d35d48c futex: make futex... |
2923 |
out_unlock_put_key: |
0d00c7b20 futexes: Clean up... |
2924 |
queue_unlock(hb); |
c87e2837b [PATCH] pi-futex:... |
2925 |
|
42d35d48c futex: make futex... |
2926 |
out_put_key: |
ae791a2d2 futex: Cleanup st... |
2927 |
put_futex_key(&q.key); |
42d35d48c futex: make futex... |
2928 |
out: |
97181f9bd futex: Avoid free... |
2929 2930 |
if (to) { hrtimer_cancel(&to->timer); |
237fc6e7a add hrtimer speci... |
2931 |
destroy_hrtimer_on_stack(&to->timer); |
97181f9bd futex: Avoid free... |
2932 |
} |
dd9739980 futex: split out ... |
2933 |
return ret != -EINTR ? ret : -ERESTARTNOINTR; |
c87e2837b [PATCH] pi-futex:... |
2934 |
|
42d35d48c futex: make futex... |
2935 |
uaddr_faulted: |
0d00c7b20 futexes: Clean up... |
2936 |
queue_unlock(hb); |
778e9a9c3 pi-futex: fix exi... |
2937 |
|
d0725992c futex: Fix the wr... |
2938 |
ret = fault_in_user_writeable(uaddr); |
e4dc5b7a3 futex: clean up f... |
2939 2940 |
if (ret) goto out_put_key; |
c87e2837b [PATCH] pi-futex:... |
2941 |
|
b41277dc7 futex: Replace fs... |
2942 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
2943 |
goto retry_private; |
ae791a2d2 futex: Cleanup st... |
2944 |
put_futex_key(&q.key); |
e4dc5b7a3 futex: clean up f... |
2945 |
goto retry; |
c87e2837b [PATCH] pi-futex:... |
2946 2947 2948 |
} /* |
c87e2837b [PATCH] pi-futex:... |
2949 2950 2951 2952 |
* Userspace attempted a TID -> 0 atomic transition, and failed. * This is the in-kernel slowpath: we look up the PI state (if any), * and do the rt-mutex unlock. */ |
b41277dc7 futex: Replace fs... |
2953 |
static int futex_unlock_pi(u32 __user *uaddr, unsigned int flags) |
c87e2837b [PATCH] pi-futex:... |
2954 |
{ |
ccf9e6a80 futex: Make unloc... |
2955 |
u32 uninitialized_var(curval), uval, vpid = task_pid_vnr(current); |
38d47c1b7 futex: rely on ge... |
2956 |
union futex_key key = FUTEX_KEY_INIT; |
ccf9e6a80 futex: Make unloc... |
2957 |
struct futex_hash_bucket *hb; |
499f5aca2 futex: Cleanup va... |
2958 |
struct futex_q *top_waiter; |
e4dc5b7a3 futex: clean up f... |
2959 |
int ret; |
c87e2837b [PATCH] pi-futex:... |
2960 |
|
bc2eecd7e futex: Allow for ... |
2961 2962 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
c87e2837b [PATCH] pi-futex:... |
2963 2964 2965 2966 2967 2968 |
retry: if (get_user(uval, uaddr)) return -EFAULT; /* * We release only a lock we actually own: */ |
c0c9ed150 futex: Avoid redu... |
2969 |
if ((uval & FUTEX_TID_MASK) != vpid) |
c87e2837b [PATCH] pi-futex:... |
2970 |
return -EPERM; |
c87e2837b [PATCH] pi-futex:... |
2971 |
|
96d4f267e Remove 'type' arg... |
2972 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &key, FUTEX_WRITE); |
ccf9e6a80 futex: Make unloc... |
2973 2974 |
if (ret) return ret; |
c87e2837b [PATCH] pi-futex:... |
2975 2976 2977 |
hb = hash_futex(&key); spin_lock(&hb->lock); |
c87e2837b [PATCH] pi-futex:... |
2978 |
/* |
ccf9e6a80 futex: Make unloc... |
2979 2980 2981 |
* Check waiters first. We do not trust user space values at * all and we at least want to know if user space fiddled * with the futex value instead of blindly unlocking. |
c87e2837b [PATCH] pi-futex:... |
2982 |
*/ |
499f5aca2 futex: Cleanup va... |
2983 2984 |
top_waiter = futex_top_waiter(hb, &key); if (top_waiter) { |
16ffa12d7 futex: Pull rt_mu... |
2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 |
struct futex_pi_state *pi_state = top_waiter->pi_state; ret = -EINVAL; if (!pi_state) goto out_unlock; /* * If current does not own the pi_state then the futex is * inconsistent and user space fiddled with the futex value. */ if (pi_state->owner != current) goto out_unlock; |
bebe5b514 futex: Futex_unlo... |
2997 |
get_pi_state(pi_state); |
802ab58da futex: Lower the ... |
2998 |
/* |
bebe5b514 futex: Futex_unlo... |
2999 3000 3001 3002 |
* By taking wait_lock while still holding hb->lock, we ensure * there is no point where we hold neither; and therefore * wake_futex_pi() must observe a state consistent with what we * observed. |
1a1fb985f futex: Handle ear... |
3003 3004 3005 3006 |
* * In particular; this forces __rt_mutex_start_proxy() to * complete such that we're guaranteed to observe the * rt_waiter. Also see the WARN in wake_futex_pi(). |
16ffa12d7 futex: Pull rt_mu... |
3007 |
*/ |
bebe5b514 futex: Futex_unlo... |
3008 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); |
16ffa12d7 futex: Pull rt_mu... |
3009 |
spin_unlock(&hb->lock); |
c74aef2d0 futex: Fix pi_sta... |
3010 |
/* drops pi_state->pi_mutex.wait_lock */ |
16ffa12d7 futex: Pull rt_mu... |
3011 3012 3013 3014 3015 3016 |
ret = wake_futex_pi(uaddr, uval, pi_state); put_pi_state(pi_state); /* * Success, we're done! No tricky corner cases. |
802ab58da futex: Lower the ... |
3017 3018 3019 |
*/ if (!ret) goto out_putkey; |
c87e2837b [PATCH] pi-futex:... |
3020 |
/* |
ccf9e6a80 futex: Make unloc... |
3021 3022 |
* The atomic access to the futex value generated a * pagefault, so retry the user-access and the wakeup: |
c87e2837b [PATCH] pi-futex:... |
3023 3024 3025 |
*/ if (ret == -EFAULT) goto pi_faulted; |
802ab58da futex: Lower the ... |
3026 |
/* |
89e9e66ba futex: Handle unl... |
3027 3028 3029 |
* A unconditional UNLOCK_PI op raced against a waiter * setting the FUTEX_WAITERS bit. Try again. */ |
6b4f4bc9c locking/futex: Al... |
3030 3031 |
if (ret == -EAGAIN) goto pi_retry; |
89e9e66ba futex: Handle unl... |
3032 |
/* |
802ab58da futex: Lower the ... |
3033 3034 3035 |
* wake_futex_pi has detected invalid state. Tell user * space. */ |
16ffa12d7 futex: Pull rt_mu... |
3036 |
goto out_putkey; |
c87e2837b [PATCH] pi-futex:... |
3037 |
} |
ccf9e6a80 futex: Make unloc... |
3038 |
|
c87e2837b [PATCH] pi-futex:... |
3039 |
/* |
ccf9e6a80 futex: Make unloc... |
3040 3041 3042 3043 3044 |
* We have no kernel internal state, i.e. no waiters in the * kernel. Waiters which are about to queue themselves are stuck * on hb->lock. So we can safely ignore them. We do neither * preserve the WAITERS bit not the OWNER_DIED one. We are the * owner. |
c87e2837b [PATCH] pi-futex:... |
3045 |
*/ |
6b4f4bc9c locking/futex: Al... |
3046 |
if ((ret = cmpxchg_futex_value_locked(&curval, uaddr, uval, 0))) { |
16ffa12d7 futex: Pull rt_mu... |
3047 |
spin_unlock(&hb->lock); |
6b4f4bc9c locking/futex: Al... |
3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 |
switch (ret) { case -EFAULT: goto pi_faulted; case -EAGAIN: goto pi_retry; default: WARN_ON_ONCE(1); goto out_putkey; } |
16ffa12d7 futex: Pull rt_mu... |
3059 |
} |
c87e2837b [PATCH] pi-futex:... |
3060 |
|
ccf9e6a80 futex: Make unloc... |
3061 3062 3063 3064 |
/* * If uval has changed, let user space handle it. */ ret = (curval == uval) ? 0 : -EAGAIN; |
c87e2837b [PATCH] pi-futex:... |
3065 3066 |
out_unlock: spin_unlock(&hb->lock); |
802ab58da futex: Lower the ... |
3067 |
out_putkey: |
ae791a2d2 futex: Cleanup st... |
3068 |
put_futex_key(&key); |
c87e2837b [PATCH] pi-futex:... |
3069 |
return ret; |
6b4f4bc9c locking/futex: Al... |
3070 3071 3072 3073 |
pi_retry: put_futex_key(&key); cond_resched(); goto retry; |
c87e2837b [PATCH] pi-futex:... |
3074 |
pi_faulted: |
ae791a2d2 futex: Cleanup st... |
3075 |
put_futex_key(&key); |
c87e2837b [PATCH] pi-futex:... |
3076 |
|
d0725992c futex: Fix the wr... |
3077 |
ret = fault_in_user_writeable(uaddr); |
b56863630 futex: clean up f... |
3078 |
if (!ret) |
c87e2837b [PATCH] pi-futex:... |
3079 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
3080 3081 |
return ret; } |
52400ba94 futex: add requeu... |
3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 |
/** * handle_early_requeue_pi_wakeup() - Detect early wakeup on the initial futex * @hb: the hash_bucket futex_q was original enqueued on * @q: the futex_q woken while waiting to be requeued * @key2: the futex_key of the requeue target futex * @timeout: the timeout associated with the wait (NULL if none) * * Detect if the task was woken on the initial futex as opposed to the requeue * target futex. If so, determine if it was a timeout or a signal that caused * the wakeup and return the appropriate error code to the caller. Must be * called with the hb lock held. * |
6c23cbbd5 futex: fix kernel... |
3094 |
* Return: |
7b4ff1adb mutex, futex: adj... |
3095 3096 |
* - 0 = no early wakeup detected; * - <0 = -ETIMEDOUT or -ERESTARTNOINTR |
52400ba94 futex: add requeu... |
3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 |
*/ static inline int handle_early_requeue_pi_wakeup(struct futex_hash_bucket *hb, struct futex_q *q, union futex_key *key2, struct hrtimer_sleeper *timeout) { int ret = 0; /* * With the hb lock held, we avoid races while we process the wakeup. * We only need to hold hb (and not hb2) to ensure atomicity as the * wakeup code can't change q.key from uaddr to uaddr2 if we hold hb. * It can't be requeued from uaddr2 to something else since we don't * support a PI aware source futex for requeue. */ if (!match_futex(&q->key, key2)) { WARN_ON(q->lock_ptr && (&hb->lock != q->lock_ptr)); /* * We were woken prior to requeue by a timeout or a signal. * Unqueue the futex_q and determine which it was. */ |
2e12978a9 futex,plist: Pass... |
3118 |
plist_del(&q->list, &hb->chain); |
11d4616bd futex: revert bac... |
3119 |
hb_waiters_dec(hb); |
52400ba94 futex: add requeu... |
3120 |
|
d58e6576b futex: Handle spu... |
3121 |
/* Handle spurious wakeups gracefully */ |
11df6dddc futex: Fix spurio... |
3122 |
ret = -EWOULDBLOCK; |
52400ba94 futex: add requeu... |
3123 3124 |
if (timeout && !timeout->task) ret = -ETIMEDOUT; |
d58e6576b futex: Handle spu... |
3125 |
else if (signal_pending(current)) |
1c840c149 futex: fix restar... |
3126 |
ret = -ERESTARTNOINTR; |
52400ba94 futex: add requeu... |
3127 3128 3129 3130 3131 3132 |
} return ret; } /** * futex_wait_requeue_pi() - Wait on uaddr and take uaddr2 |
56ec1607b futex: Correct fu... |
3133 |
* @uaddr: the futex we initially wait on (non-pi) |
b41277dc7 futex: Replace fs... |
3134 |
* @flags: futex flags (FLAGS_SHARED, FLAGS_CLOCKRT, etc.), they must be |
ab51fbab3 futex: Fault/erro... |
3135 |
* the same type, no requeueing from private to shared, etc. |
52400ba94 futex: add requeu... |
3136 3137 |
* @val: the expected value of uaddr * @abs_time: absolute timeout |
56ec1607b futex: Correct fu... |
3138 |
* @bitset: 32 bit wakeup bitset set by userspace, defaults to all |
52400ba94 futex: add requeu... |
3139 3140 3141 |
* @uaddr2: the pi futex we will take prior to returning to user-space * * The caller will wait on uaddr and will be requeued by futex_requeue() to |
6f7b0a2a5 futex: Forbid uad... |
3142 3143 3144 3145 3146 |
* uaddr2 which must be PI aware and unique from uaddr. Normal wakeup will wake * on uaddr2 and complete the acquisition of the rt_mutex prior to returning to * userspace. This ensures the rt_mutex maintains an owner when it has waiters; * without one, the pi logic would not know which task to boost/deboost, if * there was a need to. |
52400ba94 futex: add requeu... |
3147 3148 |
* * We call schedule in futex_wait_queue_me() when we enqueue and return there |
6c23cbbd5 futex: fix kernel... |
3149 |
* via the following-- |
52400ba94 futex: add requeu... |
3150 |
* 1) wakeup on uaddr2 after an atomic lock acquisition by futex_requeue() |
cc6db4e60 futex: Correct fu... |
3151 3152 3153 |
* 2) wakeup on uaddr2 after a requeue * 3) signal * 4) timeout |
52400ba94 futex: add requeu... |
3154 |
* |
cc6db4e60 futex: Correct fu... |
3155 |
* If 3, cleanup and return -ERESTARTNOINTR. |
52400ba94 futex: add requeu... |
3156 3157 3158 3159 3160 3161 3162 |
* * If 2, we may then block on trying to take the rt_mutex and return via: * 5) successful lock * 6) signal * 7) timeout * 8) other lock acquisition failure * |
cc6db4e60 futex: Correct fu... |
3163 |
* If 6, return -EWOULDBLOCK (restarting the syscall would do the same). |
52400ba94 futex: add requeu... |
3164 3165 3166 |
* * If 4 or 7, we cleanup and return with -ETIMEDOUT. * |
6c23cbbd5 futex: fix kernel... |
3167 |
* Return: |
7b4ff1adb mutex, futex: adj... |
3168 3169 |
* - 0 - On success; * - <0 - On error |
52400ba94 futex: add requeu... |
3170 |
*/ |
b41277dc7 futex: Replace fs... |
3171 |
static int futex_wait_requeue_pi(u32 __user *uaddr, unsigned int flags, |
52400ba94 futex: add requeu... |
3172 |
u32 val, ktime_t *abs_time, u32 bitset, |
b41277dc7 futex: Replace fs... |
3173 |
u32 __user *uaddr2) |
52400ba94 futex: add requeu... |
3174 |
{ |
5ca584d93 futex: Consolidat... |
3175 |
struct hrtimer_sleeper timeout, *to; |
16ffa12d7 futex: Pull rt_mu... |
3176 |
struct futex_pi_state *pi_state = NULL; |
52400ba94 futex: add requeu... |
3177 |
struct rt_mutex_waiter rt_waiter; |
52400ba94 futex: add requeu... |
3178 |
struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
3179 3180 |
union futex_key key2 = FUTEX_KEY_INIT; struct futex_q q = futex_q_init; |
52400ba94 futex: add requeu... |
3181 |
int res, ret; |
52400ba94 futex: add requeu... |
3182 |
|
bc2eecd7e futex: Allow for ... |
3183 3184 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
6f7b0a2a5 futex: Forbid uad... |
3185 3186 |
if (uaddr == uaddr2) return -EINVAL; |
52400ba94 futex: add requeu... |
3187 3188 |
if (!bitset) return -EINVAL; |
5ca584d93 futex: Consolidat... |
3189 3190 |
to = futex_setup_timer(abs_time, &timeout, flags, current->timer_slack_ns); |
52400ba94 futex: add requeu... |
3191 3192 3193 3194 3195 |
/* * The waiter is allocated on our stack, manipulated by the requeue * code while we sleep on uaddr. */ |
50809358d futex,rt_mutex: I... |
3196 |
rt_mutex_init_waiter(&rt_waiter); |
52400ba94 futex: add requeu... |
3197 |
|
96d4f267e Remove 'type' arg... |
3198 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, FUTEX_WRITE); |
52400ba94 futex: add requeu... |
3199 3200 |
if (unlikely(ret != 0)) goto out; |
84bc4af59 futex: Detect mis... |
3201 3202 3203 |
q.bitset = bitset; q.rt_waiter = &rt_waiter; q.requeue_pi_key = &key2; |
7ada876a8 futex: Fix errors... |
3204 3205 3206 3207 |
/* * Prepare to wait on uaddr. On success, increments q.key (key1) ref * count. */ |
b41277dc7 futex: Replace fs... |
3208 |
ret = futex_wait_setup(uaddr, val, flags, &q, &hb); |
c8b15a706 futex: cleanup er... |
3209 3210 |
if (ret) goto out_key2; |
52400ba94 futex: add requeu... |
3211 |
|
e9c243a5a futex-prevent-req... |
3212 3213 3214 3215 3216 |
/* * The check above which compares uaddrs is not sufficient for * shared futexes. We need to compare the keys: */ if (match_futex(&q.key, &key2)) { |
13c42c2f4 futex: Unlock hb-... |
3217 |
queue_unlock(hb); |
e9c243a5a futex-prevent-req... |
3218 3219 3220 |
ret = -EINVAL; goto out_put_keys; } |
52400ba94 futex: add requeu... |
3221 |
/* Queue the futex_q, drop the hb lock, wait for wakeup. */ |
f1a11e057 futex: remove the... |
3222 |
futex_wait_queue_me(hb, &q, to); |
52400ba94 futex: add requeu... |
3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 |
spin_lock(&hb->lock); ret = handle_early_requeue_pi_wakeup(hb, &q, &key2, to); spin_unlock(&hb->lock); if (ret) goto out_put_keys; /* * In order for us to be here, we know our q.key == key2, and since * we took the hb->lock above, we also know that futex_requeue() has * completed and we no longer have to concern ourselves with a wakeup |
7ada876a8 futex: Fix errors... |
3234 3235 3236 |
* race with the atomic proxy lock acquisition by the requeue code. The * futex_requeue dropped our key1 reference and incremented our key2 * reference count. |
52400ba94 futex: add requeu... |
3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 |
*/ /* Check if the requeue code acquired the second futex for us. */ if (!q.rt_waiter) { /* * Got the lock. We might not be the anticipated owner if we * did a lock-steal - fix up the PI-state in that case. */ if (q.pi_state && (q.pi_state->owner != current)) { spin_lock(q.lock_ptr); |
ae791a2d2 futex: Cleanup st... |
3247 |
ret = fixup_pi_state_owner(uaddr2, &q, current); |
16ffa12d7 futex: Pull rt_mu... |
3248 3249 3250 3251 |
if (ret && rt_mutex_owner(&q.pi_state->pi_mutex) == current) { pi_state = q.pi_state; get_pi_state(pi_state); } |
fb75a4282 futex: Drop refco... |
3252 3253 3254 3255 |
/* * Drop the reference to the pi state which * the requeue_pi() code acquired for us. */ |
29e9ee5d4 futex: Rename fre... |
3256 |
put_pi_state(q.pi_state); |
52400ba94 futex: add requeu... |
3257 3258 3259 |
spin_unlock(q.lock_ptr); } } else { |
c236c8e95 futex: Fix potent... |
3260 |
struct rt_mutex *pi_mutex; |
52400ba94 futex: add requeu... |
3261 3262 3263 3264 3265 |
/* * We have been woken up by futex_unlock_pi(), a timeout, or a * signal. futex_unlock_pi() will not destroy the lock_ptr nor * the pi_state. */ |
f27071cb7 futex: Fix bug in... |
3266 |
WARN_ON(!q.pi_state); |
52400ba94 futex: add requeu... |
3267 |
pi_mutex = &q.pi_state->pi_mutex; |
38d589f2f futex,rt_mutex: R... |
3268 |
ret = rt_mutex_wait_proxy_lock(pi_mutex, to, &rt_waiter); |
52400ba94 futex: add requeu... |
3269 3270 |
spin_lock(q.lock_ptr); |
38d589f2f futex,rt_mutex: R... |
3271 3272 3273 3274 |
if (ret && !rt_mutex_cleanup_proxy_lock(pi_mutex, &rt_waiter)) ret = 0; debug_rt_mutex_free_waiter(&rt_waiter); |
52400ba94 futex: add requeu... |
3275 3276 3277 3278 |
/* * Fixup the pi_state owner and possibly acquire the lock if we * haven't already. */ |
ae791a2d2 futex: Cleanup st... |
3279 |
res = fixup_owner(uaddr2, &q, !ret); |
52400ba94 futex: add requeu... |
3280 3281 |
/* * If fixup_owner() returned an error, proprogate that. If it |
56ec1607b futex: Correct fu... |
3282 |
* acquired the lock, clear -ETIMEDOUT or -EINTR. |
52400ba94 futex: add requeu... |
3283 3284 3285 |
*/ if (res) ret = (res < 0) ? res : 0; |
c236c8e95 futex: Fix potent... |
3286 3287 3288 3289 3290 |
/* * If fixup_pi_state_owner() faulted and was unable to handle * the fault, unlock the rt_mutex and return the fault to * userspace. */ |
16ffa12d7 futex: Pull rt_mu... |
3291 3292 3293 3294 |
if (ret && rt_mutex_owner(&q.pi_state->pi_mutex) == current) { pi_state = q.pi_state; get_pi_state(pi_state); } |
c236c8e95 futex: Fix potent... |
3295 |
|
52400ba94 futex: add requeu... |
3296 3297 3298 |
/* Unqueue and drop the lock. */ unqueue_me_pi(&q); } |
16ffa12d7 futex: Pull rt_mu... |
3299 3300 3301 3302 |
if (pi_state) { rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); } |
c236c8e95 futex: Fix potent... |
3303 |
if (ret == -EINTR) { |
52400ba94 futex: add requeu... |
3304 |
/* |
cc6db4e60 futex: Correct fu... |
3305 3306 3307 3308 3309 |
* We've already been requeued, but cannot restart by calling * futex_lock_pi() directly. We could restart this syscall, but * it would detect that the user space "val" changed and return * -EWOULDBLOCK. Save the overhead of the restart and return * -EWOULDBLOCK directly. |
52400ba94 futex: add requeu... |
3310 |
*/ |
2070887fd futex: fix restar... |
3311 |
ret = -EWOULDBLOCK; |
52400ba94 futex: add requeu... |
3312 3313 3314 |
} out_put_keys: |
ae791a2d2 futex: Cleanup st... |
3315 |
put_futex_key(&q.key); |
c8b15a706 futex: cleanup er... |
3316 |
out_key2: |
ae791a2d2 futex: Cleanup st... |
3317 |
put_futex_key(&key2); |
52400ba94 futex: add requeu... |
3318 3319 3320 3321 3322 3323 3324 3325 |
out: if (to) { hrtimer_cancel(&to->timer); destroy_hrtimer_on_stack(&to->timer); } return ret; } |
0771dfefc [PATCH] lightweig... |
3326 3327 3328 3329 3330 3331 3332 |
/* * Support for robust futexes: the kernel cleans up held futexes at * thread exit time. * * Implementation: user-space maintains a per-thread list of locks it * is holding. Upon do_exit(), the kernel carefully walks this list, * and marks all locks that are owned by this thread with the |
c87e2837b [PATCH] pi-futex:... |
3333 |
* FUTEX_OWNER_DIED bit, and wakes up a waiter (if any). The list is |
0771dfefc [PATCH] lightweig... |
3334 3335 3336 3337 3338 3339 3340 3341 |
* always manipulated with the lock held, so the list is private and * per-thread. Userspace also maintains a per-thread 'list_op_pending' * field, to allow the kernel to clean up if the thread dies after * acquiring the lock, but just before it could have added itself to * the list. There can only be one such pending lock. */ /** |
d96ee56ce futex: Make funct... |
3342 3343 3344 |
* sys_set_robust_list() - Set the robust-futex list head of a task * @head: pointer to the list-head * @len: length of the list-head, as userspace expects |
0771dfefc [PATCH] lightweig... |
3345 |
*/ |
836f92adf [CVE-2009-0029] S... |
3346 3347 |
SYSCALL_DEFINE2(set_robust_list, struct robust_list_head __user *, head, size_t, len) |
0771dfefc [PATCH] lightweig... |
3348 |
{ |
a0c1e9073 futex: runtime en... |
3349 3350 |
if (!futex_cmpxchg_enabled) return -ENOSYS; |
0771dfefc [PATCH] lightweig... |
3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 |
/* * The kernel knows only one size for now: */ if (unlikely(len != sizeof(*head))) return -EINVAL; current->robust_list = head; return 0; } /** |
d96ee56ce futex: Make funct... |
3363 3364 3365 3366 |
* sys_get_robust_list() - Get the robust-futex list head of a task * @pid: pid of the process [zero for current task] * @head_ptr: pointer to a list-head pointer, the kernel fills it in * @len_ptr: pointer to a length field, the kernel fills in the header size |
0771dfefc [PATCH] lightweig... |
3367 |
*/ |
836f92adf [CVE-2009-0029] S... |
3368 3369 3370 |
SYSCALL_DEFINE3(get_robust_list, int, pid, struct robust_list_head __user * __user *, head_ptr, size_t __user *, len_ptr) |
0771dfefc [PATCH] lightweig... |
3371 |
{ |
ba46df984 [PATCH] __user an... |
3372 |
struct robust_list_head __user *head; |
0771dfefc [PATCH] lightweig... |
3373 |
unsigned long ret; |
bdbb776f8 futex: Do not lea... |
3374 |
struct task_struct *p; |
0771dfefc [PATCH] lightweig... |
3375 |
|
a0c1e9073 futex: runtime en... |
3376 3377 |
if (!futex_cmpxchg_enabled) return -ENOSYS; |
bdbb776f8 futex: Do not lea... |
3378 3379 3380 |
rcu_read_lock(); ret = -ESRCH; |
0771dfefc [PATCH] lightweig... |
3381 |
if (!pid) |
bdbb776f8 futex: Do not lea... |
3382 |
p = current; |
0771dfefc [PATCH] lightweig... |
3383 |
else { |
228ebcbe6 Uninline find_tas... |
3384 |
p = find_task_by_vpid(pid); |
0771dfefc [PATCH] lightweig... |
3385 3386 |
if (!p) goto err_unlock; |
0771dfefc [PATCH] lightweig... |
3387 |
} |
bdbb776f8 futex: Do not lea... |
3388 |
ret = -EPERM; |
caaee6234 ptrace: use fsuid... |
3389 |
if (!ptrace_may_access(p, PTRACE_MODE_READ_REALCREDS)) |
bdbb776f8 futex: Do not lea... |
3390 3391 3392 3393 |
goto err_unlock; head = p->robust_list; rcu_read_unlock(); |
0771dfefc [PATCH] lightweig... |
3394 3395 3396 3397 3398 |
if (put_user(sizeof(*head), len_ptr)) return -EFAULT; return put_user(head, head_ptr); err_unlock: |
aaa2a97eb [PATCH] sys_get_r... |
3399 |
rcu_read_unlock(); |
0771dfefc [PATCH] lightweig... |
3400 3401 3402 |
return ret; } |
82ca3ab31 futex: Prevent ro... |
3403 3404 3405 |
/* Constants for the pending_op argument of handle_futex_death */ #define HANDLE_DEATH_PENDING true #define HANDLE_DEATH_LIST false |
0771dfefc [PATCH] lightweig... |
3406 3407 3408 3409 |
/* * Process a futex-list entry, check whether it's owned by the * dying task, and do notification if so: */ |
82ca3ab31 futex: Prevent ro... |
3410 3411 |
static int handle_futex_death(u32 __user *uaddr, struct task_struct *curr, bool pi, bool pending_op) |
0771dfefc [PATCH] lightweig... |
3412 |
{ |
7cfdaf38d futex: uninitiali... |
3413 |
u32 uval, uninitialized_var(nval), mval; |
6b4f4bc9c locking/futex: Al... |
3414 |
int err; |
0771dfefc [PATCH] lightweig... |
3415 |
|
5a07168d8 futex: Ensure tha... |
3416 3417 3418 |
/* Futex address must be 32bit aligned */ if ((((unsigned long)uaddr) % sizeof(*uaddr)) != 0) return -1; |
8f17d3a50 [PATCH] lightweig... |
3419 3420 |
retry: if (get_user(uval, uaddr)) |
0771dfefc [PATCH] lightweig... |
3421 |
return -1; |
82ca3ab31 futex: Prevent ro... |
3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 |
/* * Special case for regular (non PI) futexes. The unlock path in * user space has two race scenarios: * * 1. The unlock path releases the user space futex value and * before it can execute the futex() syscall to wake up * waiters it is killed. * * 2. A woken up waiter is killed before it can acquire the * futex in user space. * * In both cases the TID validation below prevents a wakeup of * potential waiters which can cause these waiters to block * forever. * * In both cases the following conditions are met: * * 1) task->robust_list->list_op_pending != NULL * @pending_op == true * 2) User space futex value == 0 * 3) Regular futex: @pi == false * * If these conditions are met, it is safe to attempt waking up a * potential waiter without touching the user space futex value and * trying to set the OWNER_DIED bit. The user space futex value is * uncontended and the rest of the user space mutex state is * consistent, so a woken waiter will just take over the * uncontended futex. Setting the OWNER_DIED bit would create * inconsistent state and malfunction of the user space owner died * handling. */ if (pending_op && !pi && !uval) { futex_wake(uaddr, 1, 1, FUTEX_BITSET_MATCH_ANY); return 0; } |
6b4f4bc9c locking/futex: Al... |
3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 |
if ((uval & FUTEX_TID_MASK) != task_pid_vnr(curr)) return 0; /* * Ok, this dying thread is truly holding a futex * of interest. Set the OWNER_DIED bit atomically * via cmpxchg, and if the value had FUTEX_WAITERS * set, wake up a waiter (if any). (We have to do a * futex_wake() even if OWNER_DIED is already set - * to handle the rare but possible case of recursive * thread-death.) The rest of the cleanup is done in * userspace. */ mval = (uval & FUTEX_WAITERS) | FUTEX_OWNER_DIED; /* * We are not holding a lock here, but we want to have * the pagefault_disable/enable() protection because * we want to handle the fault gracefully. If the * access fails we try to fault in the futex with R/W * verification via get_user_pages. get_user() above * does not guarantee R/W access. If that fails we * give up and leave the futex locked. */ if ((err = cmpxchg_futex_value_locked(&nval, uaddr, uval, mval))) { switch (err) { case -EFAULT: |
6e0aa9f8a futex: Deobfuscat... |
3484 3485 3486 |
if (fault_in_user_writeable(uaddr)) return -1; goto retry; |
6b4f4bc9c locking/futex: Al... |
3487 3488 3489 |
case -EAGAIN: cond_resched(); |
8f17d3a50 [PATCH] lightweig... |
3490 |
goto retry; |
0771dfefc [PATCH] lightweig... |
3491 |
|
6b4f4bc9c locking/futex: Al... |
3492 3493 3494 3495 |
default: WARN_ON_ONCE(1); return err; } |
0771dfefc [PATCH] lightweig... |
3496 |
} |
6b4f4bc9c locking/futex: Al... |
3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 |
if (nval != uval) goto retry; /* * Wake robust non-PI futexes here. The wakeup of * PI futexes happens in exit_pi_state(): */ if (!pi && (uval & FUTEX_WAITERS)) futex_wake(uaddr, 1, 1, FUTEX_BITSET_MATCH_ANY); |
0771dfefc [PATCH] lightweig... |
3507 3508 3509 3510 |
return 0; } /* |
e3f2ddeac [PATCH] pi-futex:... |
3511 3512 3513 |
* Fetch a robust-list pointer. Bit 0 signals PI futexes: */ static inline int fetch_robust_entry(struct robust_list __user **entry, |
ba46df984 [PATCH] __user an... |
3514 |
struct robust_list __user * __user *head, |
1dcc41bb0 futex: Change 3rd... |
3515 |
unsigned int *pi) |
e3f2ddeac [PATCH] pi-futex:... |
3516 3517 |
{ unsigned long uentry; |
ba46df984 [PATCH] __user an... |
3518 |
if (get_user(uentry, (unsigned long __user *)head)) |
e3f2ddeac [PATCH] pi-futex:... |
3519 |
return -EFAULT; |
ba46df984 [PATCH] __user an... |
3520 |
*entry = (void __user *)(uentry & ~1UL); |
e3f2ddeac [PATCH] pi-futex:... |
3521 3522 3523 3524 3525 3526 |
*pi = uentry & 1; return 0; } /* |
0771dfefc [PATCH] lightweig... |
3527 3528 3529 3530 3531 |
* Walk curr->robust_list (very carefully, it's a userspace list!) * and mark any locks found there dead, and notify any waiters. * * We silently return on any sign of list-walking problem. */ |
8012f98f9 futex: Move futex... |
3532 |
static void exit_robust_list(struct task_struct *curr) |
0771dfefc [PATCH] lightweig... |
3533 3534 |
{ struct robust_list_head __user *head = curr->robust_list; |
9f96cb1e8 robust futex thre... |
3535 |
struct robust_list __user *entry, *next_entry, *pending; |
4c115e951 futex: Address co... |
3536 3537 |
unsigned int limit = ROBUST_LIST_LIMIT, pi, pip; unsigned int uninitialized_var(next_pi); |
0771dfefc [PATCH] lightweig... |
3538 |
unsigned long futex_offset; |
9f96cb1e8 robust futex thre... |
3539 |
int rc; |
0771dfefc [PATCH] lightweig... |
3540 |
|
a0c1e9073 futex: runtime en... |
3541 3542 |
if (!futex_cmpxchg_enabled) return; |
0771dfefc [PATCH] lightweig... |
3543 3544 3545 3546 |
/* * Fetch the list head (which was registered earlier, via * sys_set_robust_list()): */ |
e3f2ddeac [PATCH] pi-futex:... |
3547 |
if (fetch_robust_entry(&entry, &head->list.next, &pi)) |
0771dfefc [PATCH] lightweig... |
3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 |
return; /* * Fetch the relative futex offset: */ if (get_user(futex_offset, &head->futex_offset)) return; /* * Fetch any possibly pending lock-add first, and handle it * if it exists: */ |
e3f2ddeac [PATCH] pi-futex:... |
3558 |
if (fetch_robust_entry(&pending, &head->list_op_pending, &pip)) |
0771dfefc [PATCH] lightweig... |
3559 |
return; |
e3f2ddeac [PATCH] pi-futex:... |
3560 |
|
9f96cb1e8 robust futex thre... |
3561 |
next_entry = NULL; /* avoid warning with gcc */ |
0771dfefc [PATCH] lightweig... |
3562 3563 |
while (entry != &head->list) { /* |
9f96cb1e8 robust futex thre... |
3564 3565 3566 3567 3568 |
* Fetch the next entry in the list before calling * handle_futex_death: */ rc = fetch_robust_entry(&next_entry, &entry->next, &next_pi); /* |
0771dfefc [PATCH] lightweig... |
3569 |
* A pending lock might already be on the list, so |
c87e2837b [PATCH] pi-futex:... |
3570 |
* don't process it twice: |
0771dfefc [PATCH] lightweig... |
3571 |
*/ |
82ca3ab31 futex: Prevent ro... |
3572 |
if (entry != pending) { |
ba46df984 [PATCH] __user an... |
3573 |
if (handle_futex_death((void __user *)entry + futex_offset, |
82ca3ab31 futex: Prevent ro... |
3574 |
curr, pi, HANDLE_DEATH_LIST)) |
0771dfefc [PATCH] lightweig... |
3575 |
return; |
82ca3ab31 futex: Prevent ro... |
3576 |
} |
9f96cb1e8 robust futex thre... |
3577 |
if (rc) |
0771dfefc [PATCH] lightweig... |
3578 |
return; |
9f96cb1e8 robust futex thre... |
3579 3580 |
entry = next_entry; pi = next_pi; |
0771dfefc [PATCH] lightweig... |
3581 3582 3583 3584 3585 3586 3587 3588 |
/* * Avoid excessively long or circular lists: */ if (!--limit) break; cond_resched(); } |
9f96cb1e8 robust futex thre... |
3589 |
|
82ca3ab31 futex: Prevent ro... |
3590 |
if (pending) { |
9f96cb1e8 robust futex thre... |
3591 |
handle_futex_death((void __user *)pending + futex_offset, |
82ca3ab31 futex: Prevent ro... |
3592 3593 |
curr, pip, HANDLE_DEATH_PENDING); } |
0771dfefc [PATCH] lightweig... |
3594 |
} |
172b09ddc futex: Provide st... |
3595 |
static void futex_cleanup(struct task_struct *tsk) |
8012f98f9 futex: Move futex... |
3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 |
{ if (unlikely(tsk->robust_list)) { exit_robust_list(tsk); tsk->robust_list = NULL; } #ifdef CONFIG_COMPAT if (unlikely(tsk->compat_robust_list)) { compat_exit_robust_list(tsk); tsk->compat_robust_list = NULL; } #endif if (unlikely(!list_empty(&tsk->pi_state_list))) exit_pi_state_list(tsk); } |
b2f4e1067 futex: Mark the b... |
3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 |
/** * futex_exit_recursive - Set the tasks futex state to FUTEX_STATE_DEAD * @tsk: task to set the state on * * Set the futex exit state of the task lockless. The futex waiter code * observes that state when a task is exiting and loops until the task has * actually finished the futex cleanup. The worst case for this is that the * waiter runs through the wait loop until the state becomes visible. * * This is called from the recursive fault handling path in do_exit(). * * This is best effort. Either the futex exit code has run already or * not. If the OWNER_DIED bit has been set on the futex then the waiter can * take it over. If not, the problem is pushed back to user space. If the * futex exit code did not run yet, then an already queued waiter might * block forever, but there is nothing which can be done about that. */ void futex_exit_recursive(struct task_struct *tsk) { |
d3ba1e8d5 futex: Add mutex ... |
3631 3632 3633 |
/* If the state is FUTEX_STATE_EXITING then futex_exit_mutex is held */ if (tsk->futex_state == FUTEX_STATE_EXITING) mutex_unlock(&tsk->futex_exit_mutex); |
b2f4e1067 futex: Mark the b... |
3634 3635 |
tsk->futex_state = FUTEX_STATE_DEAD; } |
172b09ddc futex: Provide st... |
3636 |
static void futex_cleanup_begin(struct task_struct *tsk) |
1bcee2337 futex: Split fute... |
3637 |
{ |
b2f4e1067 futex: Mark the b... |
3638 |
/* |
d3ba1e8d5 futex: Add mutex ... |
3639 3640 3641 3642 3643 3644 3645 3646 |
* Prevent various race issues against a concurrent incoming waiter * including live locks by forcing the waiter to block on * tsk->futex_exit_mutex when it observes FUTEX_STATE_EXITING in * attach_to_pi_owner(). */ mutex_lock(&tsk->futex_exit_mutex); /* |
398902dbc futex: Sanitize e... |
3647 3648 3649 3650 3651 3652 3653 3654 3655 |
* Switch the state to FUTEX_STATE_EXITING under tsk->pi_lock. * * This ensures that all subsequent checks of tsk->futex_state in * attach_to_pi_owner() must observe FUTEX_STATE_EXITING with * tsk->pi_lock held. * * It guarantees also that a pi_state which was queued right before * the state change under tsk->pi_lock by a concurrent waiter must * be observed in exit_pi_state_list(). |
b2f4e1067 futex: Mark the b... |
3656 3657 |
*/ raw_spin_lock_irq(&tsk->pi_lock); |
398902dbc futex: Sanitize e... |
3658 |
tsk->futex_state = FUTEX_STATE_EXITING; |
b2f4e1067 futex: Mark the b... |
3659 |
raw_spin_unlock_irq(&tsk->pi_lock); |
172b09ddc futex: Provide st... |
3660 |
} |
b2f4e1067 futex: Mark the b... |
3661 |
|
172b09ddc futex: Provide st... |
3662 3663 3664 3665 3666 3667 3668 |
static void futex_cleanup_end(struct task_struct *tsk, int state) { /* * Lockless store. The only side effect is that an observer might * take another loop until it becomes visible. */ tsk->futex_state = state; |
d3ba1e8d5 futex: Add mutex ... |
3669 3670 3671 3672 3673 |
/* * Drop the exit protection. This unblocks waiters which observed * FUTEX_STATE_EXITING to reevaluate the state. */ mutex_unlock(&tsk->futex_exit_mutex); |
172b09ddc futex: Provide st... |
3674 |
} |
b2f4e1067 futex: Mark the b... |
3675 |
|
172b09ddc futex: Provide st... |
3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 |
void futex_exec_release(struct task_struct *tsk) { /* * The state handling is done for consistency, but in the case of * exec() there is no way to prevent futher damage as the PID stays * the same. But for the unlikely and arguably buggy case that a * futex is held on exec(), this provides at least as much state * consistency protection which is possible. */ futex_cleanup_begin(tsk); futex_cleanup(tsk); /* * Reset the state to FUTEX_STATE_OK. The task is alive and about * exec a new binary. */ futex_cleanup_end(tsk, FUTEX_STATE_OK); } void futex_exit_release(struct task_struct *tsk) { futex_cleanup_begin(tsk); futex_cleanup(tsk); futex_cleanup_end(tsk, FUTEX_STATE_DEAD); |
1bcee2337 futex: Split fute... |
3699 |
} |
c19384b5b Make futex_wait()... |
3700 |
long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, |
e2970f2fb [PATCH] pi-futex:... |
3701 |
u32 __user *uaddr2, u32 val2, u32 val3) |
1da177e4c Linux-2.6.12-rc2 |
3702 |
{ |
81b40539e futex: Simplify r... |
3703 |
int cmd = op & FUTEX_CMD_MASK; |
b41277dc7 futex: Replace fs... |
3704 |
unsigned int flags = 0; |
34f01cc1f FUTEX: new PRIVAT... |
3705 3706 |
if (!(op & FUTEX_PRIVATE_FLAG)) |
b41277dc7 futex: Replace fs... |
3707 |
flags |= FLAGS_SHARED; |
1da177e4c Linux-2.6.12-rc2 |
3708 |
|
b41277dc7 futex: Replace fs... |
3709 3710 |
if (op & FUTEX_CLOCK_REALTIME) { flags |= FLAGS_CLOCKRT; |
337f13046 futex: Allow FUTE... |
3711 3712 |
if (cmd != FUTEX_WAIT && cmd != FUTEX_WAIT_BITSET && \ cmd != FUTEX_WAIT_REQUEUE_PI) |
b41277dc7 futex: Replace fs... |
3713 3714 |
return -ENOSYS; } |
1da177e4c Linux-2.6.12-rc2 |
3715 |
|
34f01cc1f FUTEX: new PRIVAT... |
3716 |
switch (cmd) { |
59263b513 futex: Cover all ... |
3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 |
case FUTEX_LOCK_PI: case FUTEX_UNLOCK_PI: case FUTEX_TRYLOCK_PI: case FUTEX_WAIT_REQUEUE_PI: case FUTEX_CMP_REQUEUE_PI: if (!futex_cmpxchg_enabled) return -ENOSYS; } switch (cmd) { |
1da177e4c Linux-2.6.12-rc2 |
3727 |
case FUTEX_WAIT: |
cd689985c futex: Add bitset... |
3728 |
val3 = FUTEX_BITSET_MATCH_ANY; |
b639186ff futex: Mark expec... |
3729 |
/* fall through */ |
cd689985c futex: Add bitset... |
3730 |
case FUTEX_WAIT_BITSET: |
81b40539e futex: Simplify r... |
3731 |
return futex_wait(uaddr, flags, val, timeout, val3); |
1da177e4c Linux-2.6.12-rc2 |
3732 |
case FUTEX_WAKE: |
cd689985c futex: Add bitset... |
3733 |
val3 = FUTEX_BITSET_MATCH_ANY; |
b639186ff futex: Mark expec... |
3734 |
/* fall through */ |
cd689985c futex: Add bitset... |
3735 |
case FUTEX_WAKE_BITSET: |
81b40539e futex: Simplify r... |
3736 |
return futex_wake(uaddr, flags, val, val3); |
1da177e4c Linux-2.6.12-rc2 |
3737 |
case FUTEX_REQUEUE: |
81b40539e futex: Simplify r... |
3738 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, NULL, 0); |
1da177e4c Linux-2.6.12-rc2 |
3739 |
case FUTEX_CMP_REQUEUE: |
81b40539e futex: Simplify r... |
3740 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 0); |
4732efbeb [PATCH] FUTEX_WAK... |
3741 |
case FUTEX_WAKE_OP: |
81b40539e futex: Simplify r... |
3742 |
return futex_wake_op(uaddr, flags, uaddr2, val, val2, val3); |
c87e2837b [PATCH] pi-futex:... |
3743 |
case FUTEX_LOCK_PI: |
996636dda futex: Fix argume... |
3744 |
return futex_lock_pi(uaddr, flags, timeout, 0); |
c87e2837b [PATCH] pi-futex:... |
3745 |
case FUTEX_UNLOCK_PI: |
81b40539e futex: Simplify r... |
3746 |
return futex_unlock_pi(uaddr, flags); |
c87e2837b [PATCH] pi-futex:... |
3747 |
case FUTEX_TRYLOCK_PI: |
996636dda futex: Fix argume... |
3748 |
return futex_lock_pi(uaddr, flags, NULL, 1); |
52400ba94 futex: add requeu... |
3749 3750 |
case FUTEX_WAIT_REQUEUE_PI: val3 = FUTEX_BITSET_MATCH_ANY; |
81b40539e futex: Simplify r... |
3751 3752 |
return futex_wait_requeue_pi(uaddr, flags, val, timeout, val3, uaddr2); |
52400ba94 futex: add requeu... |
3753 |
case FUTEX_CMP_REQUEUE_PI: |
81b40539e futex: Simplify r... |
3754 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 1); |
1da177e4c Linux-2.6.12-rc2 |
3755 |
} |
81b40539e futex: Simplify r... |
3756 |
return -ENOSYS; |
1da177e4c Linux-2.6.12-rc2 |
3757 |
} |
17da2bd90 [CVE-2009-0029] S... |
3758 |
SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, |
bec2f7cbb y2038: futex: Add... |
3759 |
struct __kernel_timespec __user *, utime, u32 __user *, uaddr2, |
17da2bd90 [CVE-2009-0029] S... |
3760 |
u32, val3) |
1da177e4c Linux-2.6.12-rc2 |
3761 |
{ |
bec2f7cbb y2038: futex: Add... |
3762 |
struct timespec64 ts; |
c19384b5b Make futex_wait()... |
3763 |
ktime_t t, *tp = NULL; |
e2970f2fb [PATCH] pi-futex:... |
3764 |
u32 val2 = 0; |
34f01cc1f FUTEX: new PRIVAT... |
3765 |
int cmd = op & FUTEX_CMD_MASK; |
1da177e4c Linux-2.6.12-rc2 |
3766 |
|
cd689985c futex: Add bitset... |
3767 |
if (utime && (cmd == FUTEX_WAIT || cmd == FUTEX_LOCK_PI || |
52400ba94 futex: add requeu... |
3768 3769 |
cmd == FUTEX_WAIT_BITSET || cmd == FUTEX_WAIT_REQUEUE_PI)) { |
ab51fbab3 futex: Fault/erro... |
3770 3771 |
if (unlikely(should_fail_futex(!(op & FUTEX_PRIVATE_FLAG)))) return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3772 |
if (get_timespec64(&ts, utime)) |
1da177e4c Linux-2.6.12-rc2 |
3773 |
return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3774 |
if (!timespec64_valid(&ts)) |
9741ef964 [PATCH] futex: ch... |
3775 |
return -EINVAL; |
c19384b5b Make futex_wait()... |
3776 |
|
bec2f7cbb y2038: futex: Add... |
3777 |
t = timespec64_to_ktime(ts); |
34f01cc1f FUTEX: new PRIVAT... |
3778 |
if (cmd == FUTEX_WAIT) |
5a7780e72 hrtimer: check re... |
3779 |
t = ktime_add_safe(ktime_get(), t); |
c19384b5b Make futex_wait()... |
3780 |
tp = &t; |
1da177e4c Linux-2.6.12-rc2 |
3781 3782 |
} /* |
52400ba94 futex: add requeu... |
3783 |
* requeue parameter in 'utime' if cmd == FUTEX_*_REQUEUE_*. |
f54f09861 futex: pass nr_wa... |
3784 |
* number of waiters to wake in 'utime' if cmd == FUTEX_WAKE_OP. |
1da177e4c Linux-2.6.12-rc2 |
3785 |
*/ |
f54f09861 futex: pass nr_wa... |
3786 |
if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE || |
ba9c22f2c futex: remove FUT... |
3787 |
cmd == FUTEX_CMP_REQUEUE_PI || cmd == FUTEX_WAKE_OP) |
e2970f2fb [PATCH] pi-futex:... |
3788 |
val2 = (u32) (unsigned long) utime; |
1da177e4c Linux-2.6.12-rc2 |
3789 |
|
c19384b5b Make futex_wait()... |
3790 |
return do_futex(uaddr, op, val, tp, uaddr2, val2, val3); |
1da177e4c Linux-2.6.12-rc2 |
3791 |
} |
04e7712f4 y2038: futex: Mov... |
3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 |
#ifdef CONFIG_COMPAT /* * Fetch a robust-list pointer. Bit 0 signals PI futexes: */ static inline int compat_fetch_robust_entry(compat_uptr_t *uentry, struct robust_list __user **entry, compat_uptr_t __user *head, unsigned int *pi) { if (get_user(*uentry, head)) return -EFAULT; *entry = compat_ptr((*uentry) & ~1); *pi = (unsigned int)(*uentry) & 1; return 0; } static void __user *futex_uaddr(struct robust_list __user *entry, compat_long_t futex_offset) { compat_uptr_t base = ptr_to_compat(entry); void __user *uaddr = compat_ptr(base + futex_offset); return uaddr; } /* * Walk curr->robust_list (very carefully, it's a userspace list!) * and mark any locks found there dead, and notify any waiters. * * We silently return on any sign of list-walking problem. */ |
8012f98f9 futex: Move futex... |
3824 |
static void compat_exit_robust_list(struct task_struct *curr) |
04e7712f4 y2038: futex: Mov... |
3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 |
{ struct compat_robust_list_head __user *head = curr->compat_robust_list; struct robust_list __user *entry, *next_entry, *pending; unsigned int limit = ROBUST_LIST_LIMIT, pi, pip; unsigned int uninitialized_var(next_pi); compat_uptr_t uentry, next_uentry, upending; compat_long_t futex_offset; int rc; if (!futex_cmpxchg_enabled) return; /* * Fetch the list head (which was registered earlier, via * sys_set_robust_list()): */ if (compat_fetch_robust_entry(&uentry, &entry, &head->list.next, &pi)) return; /* * Fetch the relative futex offset: */ if (get_user(futex_offset, &head->futex_offset)) return; /* * Fetch any possibly pending lock-add first, and handle it * if it exists: */ if (compat_fetch_robust_entry(&upending, &pending, &head->list_op_pending, &pip)) return; next_entry = NULL; /* avoid warning with gcc */ while (entry != (struct robust_list __user *) &head->list) { /* * Fetch the next entry in the list before calling * handle_futex_death: */ rc = compat_fetch_robust_entry(&next_uentry, &next_entry, (compat_uptr_t __user *)&entry->next, &next_pi); /* * A pending lock might already be on the list, so * dont process it twice: */ if (entry != pending) { void __user *uaddr = futex_uaddr(entry, futex_offset); |
82ca3ab31 futex: Prevent ro... |
3870 3871 |
if (handle_futex_death(uaddr, curr, pi, HANDLE_DEATH_LIST)) |
04e7712f4 y2038: futex: Mov... |
3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 |
return; } if (rc) return; uentry = next_uentry; entry = next_entry; pi = next_pi; /* * Avoid excessively long or circular lists: */ if (!--limit) break; cond_resched(); } if (pending) { void __user *uaddr = futex_uaddr(pending, futex_offset); |
82ca3ab31 futex: Prevent ro... |
3889 |
handle_futex_death(uaddr, curr, pip, HANDLE_DEATH_PENDING); |
04e7712f4 y2038: futex: Mov... |
3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 |
} } COMPAT_SYSCALL_DEFINE2(set_robust_list, struct compat_robust_list_head __user *, head, compat_size_t, len) { if (!futex_cmpxchg_enabled) return -ENOSYS; if (unlikely(len != sizeof(*head))) return -EINVAL; current->compat_robust_list = head; return 0; } COMPAT_SYSCALL_DEFINE3(get_robust_list, int, pid, compat_uptr_t __user *, head_ptr, compat_size_t __user *, len_ptr) { struct compat_robust_list_head __user *head; unsigned long ret; struct task_struct *p; if (!futex_cmpxchg_enabled) return -ENOSYS; rcu_read_lock(); ret = -ESRCH; if (!pid) p = current; else { p = find_task_by_vpid(pid); if (!p) goto err_unlock; } ret = -EPERM; if (!ptrace_may_access(p, PTRACE_MODE_READ_REALCREDS)) goto err_unlock; head = p->compat_robust_list; rcu_read_unlock(); if (put_user(sizeof(*head), len_ptr)) return -EFAULT; return put_user(ptr_to_compat(head), head_ptr); err_unlock: rcu_read_unlock(); return ret; } |
bec2f7cbb y2038: futex: Add... |
3946 |
#endif /* CONFIG_COMPAT */ |
04e7712f4 y2038: futex: Mov... |
3947 |
|
bec2f7cbb y2038: futex: Add... |
3948 |
#ifdef CONFIG_COMPAT_32BIT_TIME |
8dabe7245 y2038: syscalls: ... |
3949 |
SYSCALL_DEFINE6(futex_time32, u32 __user *, uaddr, int, op, u32, val, |
04e7712f4 y2038: futex: Mov... |
3950 3951 3952 |
struct old_timespec32 __user *, utime, u32 __user *, uaddr2, u32, val3) { |
bec2f7cbb y2038: futex: Add... |
3953 |
struct timespec64 ts; |
04e7712f4 y2038: futex: Mov... |
3954 3955 3956 3957 3958 3959 3960 |
ktime_t t, *tp = NULL; int val2 = 0; int cmd = op & FUTEX_CMD_MASK; if (utime && (cmd == FUTEX_WAIT || cmd == FUTEX_LOCK_PI || cmd == FUTEX_WAIT_BITSET || cmd == FUTEX_WAIT_REQUEUE_PI)) { |
bec2f7cbb y2038: futex: Add... |
3961 |
if (get_old_timespec32(&ts, utime)) |
04e7712f4 y2038: futex: Mov... |
3962 |
return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3963 |
if (!timespec64_valid(&ts)) |
04e7712f4 y2038: futex: Mov... |
3964 |
return -EINVAL; |
bec2f7cbb y2038: futex: Add... |
3965 |
t = timespec64_to_ktime(ts); |
04e7712f4 y2038: futex: Mov... |
3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 |
if (cmd == FUTEX_WAIT) t = ktime_add_safe(ktime_get(), t); tp = &t; } if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE || cmd == FUTEX_CMP_REQUEUE_PI || cmd == FUTEX_WAKE_OP) val2 = (int) (unsigned long) utime; return do_futex(uaddr, op, val, tp, uaddr2, val2, val3); } |
bec2f7cbb y2038: futex: Add... |
3976 |
#endif /* CONFIG_COMPAT_32BIT_TIME */ |
04e7712f4 y2038: futex: Mov... |
3977 |
|
03b8c7b62 futex: Allow arch... |
3978 |
static void __init futex_detect_cmpxchg(void) |
1da177e4c Linux-2.6.12-rc2 |
3979 |
{ |
03b8c7b62 futex: Allow arch... |
3980 |
#ifndef CONFIG_HAVE_FUTEX_CMPXCHG |
a0c1e9073 futex: runtime en... |
3981 |
u32 curval; |
03b8c7b62 futex: Allow arch... |
3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 |
/* * This will fail and we want it. Some arch implementations do * runtime detection of the futex_atomic_cmpxchg_inatomic() * functionality. We want to know that before we call in any * of the complex code paths. Also we want to prevent * registration of robust lists in that case. NULL is * guaranteed to fault and we get -EFAULT on functional * implementation, the non-functional ones will return * -ENOSYS. */ if (cmpxchg_futex_value_locked(&curval, NULL, 0, 0) == -EFAULT) futex_cmpxchg_enabled = 1; #endif } static int __init futex_init(void) { |
63b1a8169 futexes: Fix fute... |
4000 |
unsigned int futex_shift; |
a52b89ebb futexes: Increase... |
4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 |
unsigned long i; #if CONFIG_BASE_SMALL futex_hashsize = 16; #else futex_hashsize = roundup_pow_of_two(256 * num_possible_cpus()); #endif futex_queues = alloc_large_system_hash("futex", sizeof(*futex_queues), futex_hashsize, 0, futex_hashsize < 256 ? HASH_SMALL : 0, |
63b1a8169 futexes: Fix fute... |
4012 4013 4014 |
&futex_shift, NULL, futex_hashsize, futex_hashsize); futex_hashsize = 1UL << futex_shift; |
03b8c7b62 futex: Allow arch... |
4015 4016 |
futex_detect_cmpxchg(); |
a0c1e9073 futex: runtime en... |
4017 |
|
a52b89ebb futexes: Increase... |
4018 |
for (i = 0; i < futex_hashsize; i++) { |
11d4616bd futex: revert bac... |
4019 |
atomic_set(&futex_queues[i].waiters, 0); |
732375c6a plist: Remove the... |
4020 |
plist_head_init(&futex_queues[i].chain); |
3e4ab747e futex: fix init o... |
4021 4022 |
spin_lock_init(&futex_queues[i].lock); } |
1da177e4c Linux-2.6.12-rc2 |
4023 4024 |
return 0; } |
25f71d1c3 futex: Move futex... |
4025 |
core_initcall(futex_init); |