Blame view
kernel/futex.c
110 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 |
#include <linux/jhash.h> |
1da177e4c Linux-2.6.12-rc2 |
36 37 |
#include <linux/pagemap.h> #include <linux/syscalls.h> |
13d60f4b6 futex: Take hugep... |
38 |
#include <linux/hugetlb.h> |
88c8004fd futex: Use freeza... |
39 |
#include <linux/freezer.h> |
57c8a661d mm: remove includ... |
40 |
#include <linux/memblock.h> |
ab51fbab3 futex: Fault/erro... |
41 |
#include <linux/fault-inject.h> |
c2f7d08cc futex: Adjust abs... |
42 |
#include <linux/time_namespace.h> |
b488893a3 pid namespaces: c... |
43 |
|
4732efbeb [PATCH] FUTEX_WAK... |
44 |
#include <asm/futex.h> |
1da177e4c Linux-2.6.12-rc2 |
45 |
|
1696a8bee locking: Move the... |
46 |
#include "locking/rtmutex_common.h" |
c87e2837b [PATCH] pi-futex:... |
47 |
|
99b60ce69 futexes: Document... |
48 |
/* |
d7e8af1af futex: update doc... |
49 50 51 52 |
* READ this before attempting to hack on futexes! * * Basic futex operation and ordering guarantees * ============================================= |
99b60ce69 futexes: Document... |
53 54 55 56 |
* * 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... |
57 58 59 |
* 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... |
60 61 |
* * The waker side modifies the user space value of the futex and calls |
b0c29f79e futexes: Avoid ta... |
62 63 64 |
* 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... |
65 |
* |
b0c29f79e futexes: Avoid ta... |
66 67 68 69 70 |
* 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... |
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 |
* * 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... |
91 |
* |
b0c29f79e futexes: Avoid ta... |
92 93 94 95 96 |
* 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... |
97 98 99 |
* val = *futex; * sys_futex(WAIT, futex, val); * futex_wait(futex, val); |
b0c29f79e futexes: Avoid ta... |
100 |
* |
d7e8af1af futex: update doc... |
101 |
* waiters++; (a) |
8ad7b378d futex: Rename bar... |
102 103 104 105 106 107 108 109 110 111 |
* 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... |
112 |
* if (uval == val) |
b0c29f79e futexes: Avoid ta... |
113 |
* queue(); |
99b60ce69 futexes: Document... |
114 |
* unlock(hash_bucket(futex)); |
b0c29f79e futexes: Avoid ta... |
115 116 |
* schedule(); if (waiters) * lock(hash_bucket(futex)); |
d7e8af1af futex: update doc... |
117 118 |
* else wake_waiters(futex); * waiters--; (b) unlock(hash_bucket(futex)); |
b0c29f79e futexes: Avoid ta... |
119 |
* |
d7e8af1af futex: update doc... |
120 121 |
* Where (A) orders the waiters increment and the futex value read through * atomic operations (see hb_waiters_inc) and where (B) orders the write |
4b39f99c2 futex: Remove {ge... |
122 |
* to futex and the waiters read (see hb_waiters_pending()). |
b0c29f79e futexes: Avoid ta... |
123 124 125 126 127 128 129 130 131 132 133 134 |
* * 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... |
135 136 137 138 139 140 141 142 143 144 145 |
* * 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... |
146 |
*/ |
04e7712f4 y2038: futex: Mov... |
147 148 149 150 |
#ifdef CONFIG_HAVE_FUTEX_CMPXCHG #define futex_cmpxchg_enabled 1 #else static int __read_mostly futex_cmpxchg_enabled; |
03b8c7b62 futex: Allow arch... |
151 |
#endif |
a0c1e9073 futex: runtime en... |
152 |
|
1da177e4c Linux-2.6.12-rc2 |
153 |
/* |
b41277dc7 futex: Replace fs... |
154 155 156 |
* Futex flags used to encode options to functions and preserve them across * restarts. */ |
784bdf3bb futex: Assume all... |
157 158 159 160 161 162 163 164 165 |
#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... |
166 167 168 169 |
#define FLAGS_CLOCKRT 0x02 #define FLAGS_HAS_TIMEOUT 0x04 /* |
c87e2837b [PATCH] pi-futex:... |
170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 |
* 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... |
185 |
refcount_t refcount; |
c87e2837b [PATCH] pi-futex:... |
186 187 |
union futex_key key; |
3859a271a randstruct: Mark ... |
188 |
} __randomize_layout; |
c87e2837b [PATCH] pi-futex:... |
189 |
|
d8d88fbb1 futex: Correct fu... |
190 191 |
/** * struct futex_q - The hashed futex queue entry, one per waiting task |
fb62db2ba futex: Fix kernel... |
192 |
* @list: priority-sorted list of tasks waiting on this futex |
d8d88fbb1 futex: Correct fu... |
193 194 195 196 197 198 199 200 |
* @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... |
201 |
* We use this hashed waitqueue, instead of a normal wait_queue_entry_t, so |
1da177e4c Linux-2.6.12-rc2 |
202 203 204 |
* 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... |
205 |
* It is considered woken when plist_node_empty(&q->list) || q->lock_ptr == 0. |
fb62db2ba futex: Fix kernel... |
206 |
* The order of wakeup is always to make the first condition true, then |
d8d88fbb1 futex: Correct fu... |
207 208 209 210 |
* 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 |
211 212 |
*/ struct futex_q { |
ec92d0829 futex priority ba... |
213 |
struct plist_node list; |
1da177e4c Linux-2.6.12-rc2 |
214 |
|
d8d88fbb1 futex: Correct fu... |
215 |
struct task_struct *task; |
1da177e4c Linux-2.6.12-rc2 |
216 |
spinlock_t *lock_ptr; |
1da177e4c Linux-2.6.12-rc2 |
217 |
union futex_key key; |
c87e2837b [PATCH] pi-futex:... |
218 |
struct futex_pi_state *pi_state; |
52400ba94 futex: add requeu... |
219 |
struct rt_mutex_waiter *rt_waiter; |
84bc4af59 futex: Detect mis... |
220 |
union futex_key *requeue_pi_key; |
cd689985c futex: Add bitset... |
221 |
u32 bitset; |
3859a271a randstruct: Mark ... |
222 |
} __randomize_layout; |
1da177e4c Linux-2.6.12-rc2 |
223 |
|
5bdb05f91 futex: Add futex_... |
224 225 226 227 228 |
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 |
229 |
/* |
b2d0994b1 futex: update fut... |
230 231 232 |
* 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 |
233 234 |
*/ struct futex_hash_bucket { |
11d4616bd futex: revert bac... |
235 |
atomic_t waiters; |
ec92d0829 futex priority ba... |
236 237 |
spinlock_t lock; struct plist_head chain; |
a52b89ebb futexes: Increase... |
238 |
} ____cacheline_aligned_in_smp; |
1da177e4c Linux-2.6.12-rc2 |
239 |
|
ac742d371 futex: Force hot ... |
240 241 242 243 244 245 246 247 248 249 250 |
/* * 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... |
251 |
|
1da177e4c Linux-2.6.12-rc2 |
252 |
|
ab51fbab3 futex: Fault/erro... |
253 254 255 256 257 258 259 |
/* * Fault injections for futexes. */ #ifdef CONFIG_FAIL_FUTEX static struct { struct fault_attr attr; |
621a5f7ad debugfs: Pass boo... |
260 |
bool ignore_private; |
ab51fbab3 futex: Fault/erro... |
261 262 |
} fail_futex = { .attr = FAULT_ATTR_INITIALIZER, |
621a5f7ad debugfs: Pass boo... |
263 |
.ignore_private = false, |
ab51fbab3 futex: Fault/erro... |
264 265 266 267 268 269 270 |
}; 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... |
271 |
static bool should_fail_futex(bool fshared) |
ab51fbab3 futex: Fault/erro... |
272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 |
{ 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... |
290 291 |
debugfs_create_bool("ignore-private", mode, dir, &fail_futex.ignore_private); |
ab51fbab3 futex: Fault/erro... |
292 293 294 295 296 297 298 299 300 301 302 303 304 |
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 */ |
ba31c1a48 futex: Move futex... |
305 306 307 308 309 |
#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 |
11d4616bd futex: revert bac... |
310 311 312 313 |
/* * Reflects a new waiter being added to the waitqueue. */ static inline void hb_waiters_inc(struct futex_hash_bucket *hb) |
b0c29f79e futexes: Avoid ta... |
314 315 |
{ #ifdef CONFIG_SMP |
11d4616bd futex: revert bac... |
316 |
atomic_inc(&hb->waiters); |
b0c29f79e futexes: Avoid ta... |
317 |
/* |
11d4616bd futex: revert bac... |
318 |
* Full barrier (A), see the ordering comment above. |
b0c29f79e futexes: Avoid ta... |
319 |
*/ |
4e857c58e arch: Mass conver... |
320 |
smp_mb__after_atomic(); |
11d4616bd futex: revert bac... |
321 322 323 324 325 326 327 328 329 330 331 332 333 |
#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... |
334 |
|
11d4616bd futex: revert bac... |
335 336 337 |
static inline int hb_waiters_pending(struct futex_hash_bucket *hb) { #ifdef CONFIG_SMP |
4b39f99c2 futex: Remove {ge... |
338 339 340 341 |
/* * Full barrier (B), see the ordering comment above. */ smp_mb(); |
11d4616bd futex: revert bac... |
342 |
return atomic_read(&hb->waiters); |
b0c29f79e futexes: Avoid ta... |
343 |
#else |
11d4616bd futex: revert bac... |
344 |
return 1; |
b0c29f79e futexes: Avoid ta... |
345 346 |
#endif } |
e8b61b3f2 futex: Add some m... |
347 348 349 350 351 352 |
/** * 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 |
353 354 355 |
*/ static struct futex_hash_bucket *hash_futex(union futex_key *key) { |
8d6774365 futex: Unbreak fu... |
356 |
u32 hash = jhash2((u32 *)key, offsetof(typeof(*key), both.offset) / 4, |
1da177e4c Linux-2.6.12-rc2 |
357 |
key->both.offset); |
8d6774365 futex: Unbreak fu... |
358 |
|
a52b89ebb futexes: Increase... |
359 |
return &futex_queues[hash & (futex_hashsize - 1)]; |
1da177e4c Linux-2.6.12-rc2 |
360 |
} |
e8b61b3f2 futex: Add some m... |
361 362 363 364 365 366 |
/** * match_futex - Check whether two futex keys are equal * @key1: Pointer to key1 * @key2: Pointer to key2 * |
1da177e4c Linux-2.6.12-rc2 |
367 368 369 370 |
* 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 ... |
371 372 |
return (key1 && key2 && key1->both.word == key2->both.word |
1da177e4c Linux-2.6.12-rc2 |
373 374 375 |
&& key1->both.ptr == key2->both.ptr && key1->both.offset == key2->both.offset); } |
96d4f267e Remove 'type' arg... |
376 377 378 379 |
enum futex_access { FUTEX_READ, FUTEX_WRITE }; |
34f01cc1f FUTEX: new PRIVAT... |
380 |
/** |
5ca584d93 futex: Consolidat... |
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 |
* 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... |
396 397 398 |
hrtimer_init_sleeper_on_stack(timeout, (flags & FLAGS_CLOCKRT) ? CLOCK_REALTIME : CLOCK_MONOTONIC, HRTIMER_MODE_ABS); |
5ca584d93 futex: Consolidat... |
399 400 401 402 403 404 405 406 |
/* * 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; } |
8019ad13e futex: Fix inode ... |
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 |
/* * 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... |
446 |
/** |
d96ee56ce futex: Make funct... |
447 448 |
* get_futex_key() - Get parameters which are the keys for a futex * @uaddr: virtual address of the futex |
926130859 futex: Consistent... |
449 |
* @fshared: false for a PROCESS_PRIVATE futex, true for PROCESS_SHARED |
d96ee56ce futex: Make funct... |
450 |
* @key: address where result is stored. |
96d4f267e Remove 'type' arg... |
451 452 |
* @rw: mapping needs to be read/write (values: FUTEX_READ, * FUTEX_WRITE) |
34f01cc1f FUTEX: new PRIVAT... |
453 |
* |
6c23cbbd5 futex: fix kernel... |
454 455 |
* Return: a negative error code or 0 * |
7b4ff1adb mutex, futex: adj... |
456 |
* The key words are stored in @key on success. |
1da177e4c Linux-2.6.12-rc2 |
457 |
* |
8019ad13e futex: Fix inode ... |
458 |
* For shared mappings (when @fshared), the key is: |
03c109d66 futex: get rid of... |
459 |
* |
8019ad13e futex: Fix inode ... |
460 |
* ( inode->i_sequence, page->index, offset_within_page ) |
03c109d66 futex: get rid of... |
461 |
* |
8019ad13e futex: Fix inode ... |
462 463 464 |
* [ also see get_inode_sequence_number() ] * * For private mappings (or when !@fshared), the key is: |
03c109d66 futex: get rid of... |
465 |
* |
8019ad13e futex: Fix inode ... |
466 467 468 469 |
* ( 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 |
470 |
* |
b2d0994b1 futex: update fut... |
471 |
* lock_page() might sleep, the caller should not hold a spinlock. |
1da177e4c Linux-2.6.12-rc2 |
472 |
*/ |
926130859 futex: Consistent... |
473 474 |
static int get_futex_key(u32 __user *uaddr, bool fshared, union futex_key *key, enum futex_access rw) |
1da177e4c Linux-2.6.12-rc2 |
475 |
{ |
e2970f2fb [PATCH] pi-futex:... |
476 |
unsigned long address = (unsigned long)uaddr; |
1da177e4c Linux-2.6.12-rc2 |
477 |
struct mm_struct *mm = current->mm; |
077fa7aed futex: Calculate ... |
478 |
struct page *page, *tail; |
14d27abd1 futex, thp: remov... |
479 |
struct address_space *mapping; |
9ea71503a futex: Fix regres... |
480 |
int err, ro = 0; |
1da177e4c Linux-2.6.12-rc2 |
481 482 483 484 |
/* * The futex address must be "naturally" aligned. */ |
e2970f2fb [PATCH] pi-futex:... |
485 |
key->both.offset = address % PAGE_SIZE; |
34f01cc1f FUTEX: new PRIVAT... |
486 |
if (unlikely((address % sizeof(u32)) != 0)) |
1da177e4c Linux-2.6.12-rc2 |
487 |
return -EINVAL; |
e2970f2fb [PATCH] pi-futex:... |
488 |
address -= key->both.offset; |
1da177e4c Linux-2.6.12-rc2 |
489 |
|
96d4f267e Remove 'type' arg... |
490 |
if (unlikely(!access_ok(uaddr, sizeof(u32)))) |
5cdec2d83 futex: move user ... |
491 |
return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
492 493 |
if (unlikely(should_fail_futex(fshared))) return -EFAULT; |
1da177e4c Linux-2.6.12-rc2 |
494 |
/* |
34f01cc1f FUTEX: new PRIVAT... |
495 496 497 498 499 500 501 |
* 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... |
502 503 504 505 |
key->private.mm = mm; key->private.address = address; return 0; } |
1da177e4c Linux-2.6.12-rc2 |
506 |
|
38d47c1b7 futex: rely on ge... |
507 |
again: |
ab51fbab3 futex: Fault/erro... |
508 |
/* Ignore any VERIFY_READ mapping (futex common case) */ |
926130859 futex: Consistent... |
509 |
if (unlikely(should_fail_futex(true))) |
ab51fbab3 futex: Fault/erro... |
510 |
return -EFAULT; |
73b0140bf mm/gup: change GU... |
511 |
err = get_user_pages_fast(address, 1, FOLL_WRITE, &page); |
9ea71503a futex: Fix regres... |
512 513 514 515 |
/* * If write access is not required (eg. FUTEX_WAIT), try * and get read-only access. */ |
96d4f267e Remove 'type' arg... |
516 |
if (err == -EFAULT && rw == FUTEX_READ) { |
9ea71503a futex: Fix regres... |
517 518 519 |
err = get_user_pages_fast(address, 1, 0, &page); ro = 1; } |
38d47c1b7 futex: rely on ge... |
520 521 |
if (err < 0) return err; |
9ea71503a futex: Fix regres... |
522 523 |
else err = 0; |
38d47c1b7 futex: rely on ge... |
524 |
|
65d8fc777 futex: Remove req... |
525 526 527 528 529 530 531 532 533 534 |
/* * 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 ... |
535 536 537 538 539 540 541 |
* * 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... |
542 |
*/ |
077fa7aed futex: Calculate ... |
543 |
tail = page; |
65d8fc777 futex: Remove req... |
544 545 |
page = compound_head(page); mapping = READ_ONCE(page->mapping); |
e6780f724 futex: Fix uninte... |
546 |
/* |
14d27abd1 futex, thp: remov... |
547 |
* If page->mapping is NULL, then it cannot be a PageAnon |
e6780f724 futex: Fix uninte... |
548 549 550 551 552 553 554 555 556 557 558 |
* 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... |
559 |
* an unlikely race, but we do need to retry for page->mapping. |
e6780f724 futex: Fix uninte... |
560 |
*/ |
65d8fc777 futex: Remove req... |
561 562 563 564 565 566 567 568 569 570 |
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... |
571 572 |
unlock_page(page); put_page(page); |
65d8fc777 futex: Remove req... |
573 |
|
e6780f724 futex: Fix uninte... |
574 575 |
if (shmem_swizzled) goto again; |
65d8fc777 futex: Remove req... |
576 |
|
e6780f724 futex: Fix uninte... |
577 |
return -EFAULT; |
38d47c1b7 futex: rely on ge... |
578 |
} |
1da177e4c Linux-2.6.12-rc2 |
579 580 581 582 |
/* * Private mappings are handled in a simple way. * |
65d8fc777 futex: Remove req... |
583 584 585 |
* 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 |
586 587 |
* 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... |
588 |
* the object not the particular process. |
1da177e4c Linux-2.6.12-rc2 |
589 |
*/ |
14d27abd1 futex, thp: remov... |
590 |
if (PageAnon(page)) { |
9ea71503a futex: Fix regres... |
591 592 593 594 |
/* * A RO anonymous page will never change and thus doesn't make * sense for futex operations. */ |
926130859 futex: Consistent... |
595 |
if (unlikely(should_fail_futex(true)) || ro) { |
9ea71503a futex: Fix regres... |
596 597 598 |
err = -EFAULT; goto out; } |
38d47c1b7 futex: rely on ge... |
599 |
key->both.offset |= FUT_OFF_MMSHARED; /* ref taken on mm */ |
1da177e4c Linux-2.6.12-rc2 |
600 |
key->private.mm = mm; |
e2970f2fb [PATCH] pi-futex:... |
601 |
key->private.address = address; |
65d8fc777 futex: Remove req... |
602 |
|
38d47c1b7 futex: rely on ge... |
603 |
} else { |
65d8fc777 futex: Remove req... |
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 |
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... |
633 |
key->both.offset |= FUT_OFF_INODE; /* inode-based key */ |
8019ad13e futex: Fix inode ... |
634 |
key->shared.i_seq = get_inode_sequence_number(inode); |
077fa7aed futex: Calculate ... |
635 |
key->shared.pgoff = basepage_index(tail); |
65d8fc777 futex: Remove req... |
636 |
rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
637 |
} |
9ea71503a futex: Fix regres... |
638 |
out: |
14d27abd1 futex, thp: remov... |
639 |
put_page(page); |
9ea71503a futex: Fix regres... |
640 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
641 |
} |
d96ee56ce futex: Make funct... |
642 643 |
/** * fault_in_user_writeable() - Fault in user address and verify RW access |
d0725992c futex: Fix the wr... |
644 645 646 647 648 |
* @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... |
649 |
* We have no generic implementation of a non-destructive write to the |
d0725992c futex: Fix the wr... |
650 651 652 653 654 655 |
* 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_... |
656 657 |
struct mm_struct *mm = current->mm; int ret; |
d8ed45c5d mmap locking API:... |
658 |
mmap_read_lock(mm); |
64019a2e4 mm/gup: remove ta... |
659 |
ret = fixup_user_fault(mm, (unsigned long)uaddr, |
4a9e1cda2 mm: bring in addi... |
660 |
FAULT_FLAG_WRITE, NULL); |
d8ed45c5d mmap locking API:... |
661 |
mmap_read_unlock(mm); |
722d01723 futex: Take mmap_... |
662 |
|
d0725992c futex: Fix the wr... |
663 664 |
return ret < 0 ? ret : 0; } |
4b1c486b3 futex: add helper... |
665 666 |
/** * futex_top_waiter() - Return the highest priority waiter on a futex |
d96ee56ce futex: Make funct... |
667 668 |
* @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... |
669 670 671 672 673 674 675 676 677 678 679 680 681 682 |
* * 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... |
683 684 |
static int cmpxchg_futex_value_locked(u32 *curval, u32 __user *uaddr, u32 uval, u32 newval) |
36cf3b5c3 FUTEX: Tidy up th... |
685 |
{ |
37a9d912b futex: Sanitize c... |
686 |
int ret; |
36cf3b5c3 FUTEX: Tidy up th... |
687 688 |
pagefault_disable(); |
37a9d912b futex: Sanitize c... |
689 |
ret = futex_atomic_cmpxchg_inatomic(curval, uaddr, uval, newval); |
36cf3b5c3 FUTEX: Tidy up th... |
690 |
pagefault_enable(); |
37a9d912b futex: Sanitize c... |
691 |
return ret; |
36cf3b5c3 FUTEX: Tidy up th... |
692 693 694 |
} static int get_futex_value_locked(u32 *dest, u32 __user *from) |
1da177e4c Linux-2.6.12-rc2 |
695 696 |
{ int ret; |
a866374ae [PATCH] mm: pagef... |
697 |
pagefault_disable(); |
bd28b1459 x86: remove more ... |
698 |
ret = __get_user(*dest, from); |
a866374ae [PATCH] mm: pagef... |
699 |
pagefault_enable(); |
1da177e4c Linux-2.6.12-rc2 |
700 701 702 |
return ret ? -EFAULT : 0; } |
c87e2837b [PATCH] pi-futex:... |
703 704 705 706 707 708 709 710 711 712 |
/* * 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... |
713 |
pi_state = kzalloc(sizeof(*pi_state), GFP_KERNEL); |
c87e2837b [PATCH] pi-futex:... |
714 715 716 |
if (!pi_state) return -ENOMEM; |
c87e2837b [PATCH] pi-futex:... |
717 718 719 |
INIT_LIST_HEAD(&pi_state->list); /* pi_mutex gets initialized later */ pi_state->owner = NULL; |
49262de22 futex: Convert fu... |
720 |
refcount_set(&pi_state->refcount, 1); |
38d47c1b7 futex: rely on ge... |
721 |
pi_state->key = FUTEX_KEY_INIT; |
c87e2837b [PATCH] pi-futex:... |
722 723 724 725 726 |
current->pi_state_cache = pi_state; return 0; } |
bf92cf3a5 futex: Cleanup re... |
727 |
static struct futex_pi_state *alloc_pi_state(void) |
c87e2837b [PATCH] pi-futex:... |
728 729 730 731 732 733 734 735 |
{ 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... |
736 737 |
static void get_pi_state(struct futex_pi_state *pi_state) { |
49262de22 futex: Convert fu... |
738 |
WARN_ON_ONCE(!refcount_inc_not_zero(&pi_state->refcount)); |
bf92cf3a5 futex: Cleanup re... |
739 |
} |
30a6b8031 futex: Fix a race... |
740 |
/* |
29e9ee5d4 futex: Rename fre... |
741 742 |
* Drops a reference to the pi_state object and frees or caches it * when the last reference is gone. |
30a6b8031 futex: Fix a race... |
743 |
*/ |
29e9ee5d4 futex: Rename fre... |
744 |
static void put_pi_state(struct futex_pi_state *pi_state) |
c87e2837b [PATCH] pi-futex:... |
745 |
{ |
30a6b8031 futex: Fix a race... |
746 747 |
if (!pi_state) return; |
49262de22 futex: Convert fu... |
748 |
if (!refcount_dec_and_test(&pi_state->refcount)) |
c87e2837b [PATCH] pi-futex:... |
749 750 751 752 753 754 755 |
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... |
756 |
struct task_struct *owner; |
1e106aa35 futex: Don't enab... |
757 |
unsigned long flags; |
c87e2837b [PATCH] pi-futex:... |
758 |
|
1e106aa35 futex: Don't enab... |
759 |
raw_spin_lock_irqsave(&pi_state->pi_mutex.wait_lock, flags); |
c74aef2d0 futex: Fix pi_sta... |
760 761 762 763 764 765 766 |
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); |
1e106aa35 futex: Don't enab... |
767 |
raw_spin_unlock_irqrestore(&pi_state->pi_mutex.wait_lock, flags); |
c87e2837b [PATCH] pi-futex:... |
768 |
} |
c74aef2d0 futex: Fix pi_sta... |
769 |
if (current->pi_state_cache) { |
c87e2837b [PATCH] pi-futex:... |
770 |
kfree(pi_state); |
c74aef2d0 futex: Fix pi_sta... |
771 |
} else { |
c87e2837b [PATCH] pi-futex:... |
772 773 774 775 776 777 |
/* * 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... |
778 |
refcount_set(&pi_state->refcount, 1); |
c87e2837b [PATCH] pi-futex:... |
779 780 781 |
current->pi_state_cache = pi_state; } } |
bc2eecd7e futex: Allow for ... |
782 |
#ifdef CONFIG_FUTEX_PI |
c87e2837b [PATCH] pi-futex:... |
783 784 785 786 787 |
/* * 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.) */ |
ba31c1a48 futex: Move futex... |
788 |
static void exit_pi_state_list(struct task_struct *curr) |
c87e2837b [PATCH] pi-futex:... |
789 |
{ |
c87e2837b [PATCH] pi-futex:... |
790 791 |
struct list_head *next, *head = &curr->pi_state_list; struct futex_pi_state *pi_state; |
627371d73 [PATCH] pi-futex:... |
792 |
struct futex_hash_bucket *hb; |
38d47c1b7 futex: rely on ge... |
793 |
union futex_key key = FUTEX_KEY_INIT; |
c87e2837b [PATCH] pi-futex:... |
794 |
|
a0c1e9073 futex: runtime en... |
795 796 |
if (!futex_cmpxchg_enabled) return; |
c87e2837b [PATCH] pi-futex:... |
797 798 799 |
/* * We are a ZOMBIE and nobody can enqueue itself on * pi_state_list anymore, but we have to be careful |
627371d73 [PATCH] pi-futex:... |
800 |
* versus waiters unqueueing themselves: |
c87e2837b [PATCH] pi-futex:... |
801 |
*/ |
1d6154825 sched: Convert pi... |
802 |
raw_spin_lock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
803 |
while (!list_empty(head)) { |
c87e2837b [PATCH] pi-futex:... |
804 805 806 |
next = head->next; pi_state = list_entry(next, struct futex_pi_state, list); key = pi_state->key; |
627371d73 [PATCH] pi-futex:... |
807 |
hb = hash_futex(&key); |
153fbd122 futex: Fix more p... |
808 809 810 811 812 813 814 815 816 817 818 |
/* * 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... |
819 |
if (!refcount_inc_not_zero(&pi_state->refcount)) { |
153fbd122 futex: Fix more p... |
820 821 822 823 824 |
raw_spin_unlock_irq(&curr->pi_lock); cpu_relax(); raw_spin_lock_irq(&curr->pi_lock); continue; } |
1d6154825 sched: Convert pi... |
825 |
raw_spin_unlock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
826 |
|
c87e2837b [PATCH] pi-futex:... |
827 |
spin_lock(&hb->lock); |
c74aef2d0 futex: Fix pi_sta... |
828 829 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); raw_spin_lock(&curr->pi_lock); |
627371d73 [PATCH] pi-futex:... |
830 831 832 833 |
/* * We dropped the pi-lock, so re-check whether this * task still owns the PI-state: */ |
c87e2837b [PATCH] pi-futex:... |
834 |
if (head->next != next) { |
153fbd122 futex: Fix more p... |
835 |
/* retain curr->pi_lock for the loop invariant */ |
c74aef2d0 futex: Fix pi_sta... |
836 |
raw_spin_unlock(&pi_state->pi_mutex.wait_lock); |
c87e2837b [PATCH] pi-futex:... |
837 |
spin_unlock(&hb->lock); |
153fbd122 futex: Fix more p... |
838 |
put_pi_state(pi_state); |
c87e2837b [PATCH] pi-futex:... |
839 840 |
continue; } |
c87e2837b [PATCH] pi-futex:... |
841 |
WARN_ON(pi_state->owner != curr); |
627371d73 [PATCH] pi-futex:... |
842 843 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
c87e2837b [PATCH] pi-futex:... |
844 |
pi_state->owner = NULL; |
c87e2837b [PATCH] pi-futex:... |
845 |
|
153fbd122 futex: Fix more p... |
846 |
raw_spin_unlock(&curr->pi_lock); |
c74aef2d0 futex: Fix pi_sta... |
847 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
c87e2837b [PATCH] pi-futex:... |
848 |
spin_unlock(&hb->lock); |
16ffa12d7 futex: Pull rt_mu... |
849 850 |
rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); |
1d6154825 sched: Convert pi... |
851 |
raw_spin_lock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
852 |
} |
1d6154825 sched: Convert pi... |
853 |
raw_spin_unlock_irq(&curr->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
854 |
} |
ba31c1a48 futex: Move futex... |
855 856 |
#else static inline void exit_pi_state_list(struct task_struct *curr) { } |
bc2eecd7e futex: Allow for ... |
857 |
#endif |
54a217887 futex: Make looku... |
858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 |
/* * 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 |
7b7b8a2c9 kernel/: fix repe... |
880 |
* came here due to a stale FUTEX_WAITERS/FUTEX_OWNER_DIED bit. |
54a217887 futex: Make looku... |
881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 |
* * [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... |
906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 |
* * * 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... |
939 |
*/ |
e60cbc5ce futex: Split out ... |
940 941 942 943 944 945 |
/* * 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... |
946 947 |
static int attach_to_pi_state(u32 __user *uaddr, u32 uval, struct futex_pi_state *pi_state, |
e60cbc5ce futex: Split out ... |
948 |
struct futex_pi_state **ps) |
c87e2837b [PATCH] pi-futex:... |
949 |
{ |
778e9a9c3 pi-futex: fix exi... |
950 |
pid_t pid = uval & FUTEX_TID_MASK; |
94ffac5d8 futex: Fix small ... |
951 952 |
u32 uval2; int ret; |
c87e2837b [PATCH] pi-futex:... |
953 |
|
e60cbc5ce futex: Split out ... |
954 955 956 957 958 |
/* * Userspace might have messed up non-PI and PI futexes [3] */ if (unlikely(!pi_state)) return -EINVAL; |
06a9ec291 [PATCH] pi-futex:... |
959 |
|
734009e96 futex: Change loc... |
960 961 962 963 964 965 |
/* * 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... |
966 967 968 969 970 |
* * 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... |
971 |
*/ |
49262de22 futex: Convert fu... |
972 |
WARN_ON(!refcount_read(&pi_state->refcount)); |
59647b6ac futex: Handle fut... |
973 |
|
e60cbc5ce futex: Split out ... |
974 |
/* |
734009e96 futex: Change loc... |
975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 |
* 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 ... |
993 994 995 |
* Handle the owner died case: */ if (uval & FUTEX_OWNER_DIED) { |
bd1dbcc67 futex: Use futex_... |
996 |
/* |
e60cbc5ce futex: Split out ... |
997 998 999 |
* 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_... |
1000 |
*/ |
e60cbc5ce futex: Split out ... |
1001 |
if (!pi_state->owner) { |
59647b6ac futex: Handle fut... |
1002 |
/* |
e60cbc5ce futex: Split out ... |
1003 1004 |
* No pi state owner, but the user space TID * is not 0. Inconsistent state. [5] |
59647b6ac futex: Handle fut... |
1005 |
*/ |
e60cbc5ce futex: Split out ... |
1006 |
if (pid) |
734009e96 futex: Change loc... |
1007 |
goto out_einval; |
bd1dbcc67 futex: Use futex_... |
1008 |
/* |
e60cbc5ce futex: Split out ... |
1009 |
* Take a ref on the state and return success. [4] |
866293ee5 futex: Add anothe... |
1010 |
*/ |
734009e96 futex: Change loc... |
1011 |
goto out_attach; |
c87e2837b [PATCH] pi-futex:... |
1012 |
} |
bd1dbcc67 futex: Use futex_... |
1013 1014 |
/* |
e60cbc5ce futex: Split out ... |
1015 1016 1017 1018 1019 1020 1021 1022 |
* 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... |
1023 |
goto out_attach; |
e60cbc5ce futex: Split out ... |
1024 1025 1026 1027 |
} else { /* * If the owner died bit is not set, then the pi_state * must have an owner. [7] |
bd1dbcc67 futex: Use futex_... |
1028 |
*/ |
e60cbc5ce futex: Split out ... |
1029 |
if (!pi_state->owner) |
734009e96 futex: Change loc... |
1030 |
goto out_einval; |
c87e2837b [PATCH] pi-futex:... |
1031 1032 1033 |
} /* |
e60cbc5ce futex: Split out ... |
1034 1035 1036 1037 1038 |
* 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... |
1039 1040 1041 |
goto out_einval; out_attach: |
bf92cf3a5 futex: Cleanup re... |
1042 |
get_pi_state(pi_state); |
734009e96 futex: Change loc... |
1043 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
e60cbc5ce futex: Split out ... |
1044 1045 |
*ps = pi_state; return 0; |
734009e96 futex: Change loc... |
1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 |
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 ... |
1062 |
} |
3ef240eaf futex: Prevent ex... |
1063 1064 |
/** * wait_for_owner_exiting - Block until the owner has exited |
51bfb1d11 futex: Fix kernel... |
1065 |
* @ret: owner's current futex lock status |
3ef240eaf futex: Prevent ex... |
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 |
* @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 ... |
1093 1094 1095 1096 1097 1098 |
static int handle_exit_race(u32 __user *uaddr, u32 uval, struct task_struct *tsk) { u32 uval2; /* |
ac31c7ff8 futex: Provide di... |
1099 1100 |
* If the futex exit state is not yet FUTEX_STATE_DEAD, tell the * caller that the alleged owner is busy. |
da791a667 futex: Cure exit ... |
1101 |
*/ |
3d4775df0 futex: Replace PF... |
1102 |
if (tsk && tsk->futex_state != FUTEX_STATE_DEAD) |
ac31c7ff8 futex: Provide di... |
1103 |
return -EBUSY; |
da791a667 futex: Cure exit ... |
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 |
/* * 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(); |
3d4775df0 futex: Replace PF... |
1121 1122 1123 |
* tsk->futex_state = } else { * FUTEX_STATE_DEAD; if (tsk->futex_state != * FUTEX_STATE_DEAD) |
da791a667 futex: Cure exit ... |
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 |
* 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 ... |
1148 1149 1150 1151 |
/* * Lookup the task for the TID provided from user space and attach to * it after doing proper sanity checks. */ |
da791a667 futex: Cure exit ... |
1152 |
static int attach_to_pi_owner(u32 __user *uaddr, u32 uval, union futex_key *key, |
3ef240eaf futex: Prevent ex... |
1153 1154 |
struct futex_pi_state **ps, struct task_struct **exiting) |
e60cbc5ce futex: Split out ... |
1155 |
{ |
e60cbc5ce futex: Split out ... |
1156 |
pid_t pid = uval & FUTEX_TID_MASK; |
04e1b2e52 futex: Split out ... |
1157 1158 |
struct futex_pi_state *pi_state; struct task_struct *p; |
e60cbc5ce futex: Split out ... |
1159 1160 |
/* |
e3f2ddeac [PATCH] pi-futex:... |
1161 |
* We are the first waiter - try to look up the real owner and attach |
54a217887 futex: Make looku... |
1162 |
* the new pi_state to it, but bail out when TID = 0 [1] |
da791a667 futex: Cure exit ... |
1163 1164 1165 |
* * 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:... |
1166 |
*/ |
778e9a9c3 pi-futex: fix exi... |
1167 |
if (!pid) |
da791a667 futex: Cure exit ... |
1168 |
return -EAGAIN; |
2ee082608 pids: introduce f... |
1169 |
p = find_get_task_by_vpid(pid); |
7a0ea09ad futex: futex_find... |
1170 |
if (!p) |
da791a667 futex: Cure exit ... |
1171 |
return handle_exit_race(uaddr, uval, NULL); |
778e9a9c3 pi-futex: fix exi... |
1172 |
|
a21294644 locking/futex: Ch... |
1173 |
if (unlikely(p->flags & PF_KTHREAD)) { |
f0d71b3dc futex: Prevent at... |
1174 1175 1176 |
put_task_struct(p); return -EPERM; } |
778e9a9c3 pi-futex: fix exi... |
1177 |
/* |
3d4775df0 futex: Replace PF... |
1178 1179 1180 |
* 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... |
1181 |
*/ |
1d6154825 sched: Convert pi... |
1182 |
raw_spin_lock_irq(&p->pi_lock); |
3d4775df0 futex: Replace PF... |
1183 |
if (unlikely(p->futex_state != FUTEX_STATE_OK)) { |
778e9a9c3 pi-futex: fix exi... |
1184 |
/* |
3d4775df0 futex: Replace PF... |
1185 1186 1187 |
* 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... |
1188 |
*/ |
da791a667 futex: Cure exit ... |
1189 |
int ret = handle_exit_race(uaddr, uval, p); |
778e9a9c3 pi-futex: fix exi... |
1190 |
|
1d6154825 sched: Convert pi... |
1191 |
raw_spin_unlock_irq(&p->pi_lock); |
3ef240eaf futex: Prevent ex... |
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 |
/* * 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... |
1205 1206 |
return ret; } |
c87e2837b [PATCH] pi-futex:... |
1207 |
|
54a217887 futex: Make looku... |
1208 1209 |
/* * No existing pi state. First waiter. [2] |
734009e96 futex: Change loc... |
1210 1211 1212 |
* * This creates pi_state, we have hb->lock held, this means nothing can * observe this state, wait_lock is irrelevant. |
54a217887 futex: Make looku... |
1213 |
*/ |
c87e2837b [PATCH] pi-futex:... |
1214 1215 1216 |
pi_state = alloc_pi_state(); /* |
04e1b2e52 futex: Split out ... |
1217 |
* Initialize the pi_mutex in locked state and make @p |
c87e2837b [PATCH] pi-futex:... |
1218 1219 1220 1221 1222 |
* 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 ... |
1223 |
pi_state->key = *key; |
c87e2837b [PATCH] pi-futex:... |
1224 |
|
627371d73 [PATCH] pi-futex:... |
1225 |
WARN_ON(!list_empty(&pi_state->list)); |
c87e2837b [PATCH] pi-futex:... |
1226 |
list_add(&pi_state->list, &p->pi_state_list); |
c74aef2d0 futex: Fix pi_sta... |
1227 1228 1229 1230 |
/* * 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:... |
1231 |
pi_state->owner = p; |
1d6154825 sched: Convert pi... |
1232 |
raw_spin_unlock_irq(&p->pi_lock); |
c87e2837b [PATCH] pi-futex:... |
1233 1234 |
put_task_struct(p); |
d0aa7a70b futex_requeue_pi ... |
1235 |
*ps = pi_state; |
c87e2837b [PATCH] pi-futex:... |
1236 1237 1238 |
return 0; } |
734009e96 futex: Change loc... |
1239 1240 |
static int lookup_pi_state(u32 __user *uaddr, u32 uval, struct futex_hash_bucket *hb, |
3ef240eaf futex: Prevent ex... |
1241 1242 |
union futex_key *key, struct futex_pi_state **ps, struct task_struct **exiting) |
04e1b2e52 futex: Split out ... |
1243 |
{ |
499f5aca2 futex: Cleanup va... |
1244 |
struct futex_q *top_waiter = futex_top_waiter(hb, key); |
04e1b2e52 futex: Split out ... |
1245 1246 1247 1248 1249 |
/* * If there is a waiter on that futex, validate it and * attach to the pi_state when the validation succeeds. */ |
499f5aca2 futex: Cleanup va... |
1250 |
if (top_waiter) |
734009e96 futex: Change loc... |
1251 |
return attach_to_pi_state(uaddr, uval, top_waiter->pi_state, ps); |
04e1b2e52 futex: Split out ... |
1252 1253 1254 1255 1256 |
/* * We are the first waiter - try to look up the owner based on * @uval and attach to it. */ |
3ef240eaf futex: Prevent ex... |
1257 |
return attach_to_pi_owner(uaddr, uval, key, ps, exiting); |
04e1b2e52 futex: Split out ... |
1258 |
} |
af54d6a1c futex: Simplify f... |
1259 1260 |
static int lock_pi_update_atomic(u32 __user *uaddr, u32 uval, u32 newval) { |
6b4f4bc9c locking/futex: Al... |
1261 |
int err; |
3f649ab72 treewide: Remove ... |
1262 |
u32 curval; |
af54d6a1c futex: Simplify f... |
1263 |
|
ab51fbab3 futex: Fault/erro... |
1264 1265 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
6b4f4bc9c locking/futex: Al... |
1266 1267 1268 |
err = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (unlikely(err)) return err; |
af54d6a1c futex: Simplify f... |
1269 |
|
734009e96 futex: Change loc... |
1270 |
/* If user space value changed, let the caller retry */ |
af54d6a1c futex: Simplify f... |
1271 1272 |
return curval != uval ? -EAGAIN : 0; } |
1a52084d0 futex: split out ... |
1273 |
/** |
d96ee56ce futex: Make funct... |
1274 |
* futex_lock_pi_atomic() - Atomic work required to acquire a pi aware futex |
bab5bc9e8 futex: fixup unlo... |
1275 1276 1277 1278 1279 1280 1281 |
* @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. |
3ef240eaf futex: Prevent ex... |
1282 1283 |
* @exiting: Pointer to store the task pointer of the owner task * which is in the middle of exiting |
bab5bc9e8 futex: fixup unlo... |
1284 |
* @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) |
1a52084d0 futex: split out ... |
1285 |
* |
6c23cbbd5 futex: fix kernel... |
1286 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1287 1288 1289 |
* - 0 - ready to wait; * - 1 - acquired the lock; * - <0 - error |
1a52084d0 futex: split out ... |
1290 1291 |
* * The hb->lock and futex_key refs shall be held by the caller. |
3ef240eaf futex: Prevent ex... |
1292 1293 1294 1295 |
* * @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 ... |
1296 1297 1298 1299 |
*/ static int futex_lock_pi_atomic(u32 __user *uaddr, struct futex_hash_bucket *hb, union futex_key *key, struct futex_pi_state **ps, |
3ef240eaf futex: Prevent ex... |
1300 1301 1302 |
struct task_struct *task, struct task_struct **exiting, int set_waiters) |
1a52084d0 futex: split out ... |
1303 |
{ |
af54d6a1c futex: Simplify f... |
1304 |
u32 uval, newval, vpid = task_pid_vnr(task); |
499f5aca2 futex: Cleanup va... |
1305 |
struct futex_q *top_waiter; |
af54d6a1c futex: Simplify f... |
1306 |
int ret; |
1a52084d0 futex: split out ... |
1307 1308 |
/* |
af54d6a1c futex: Simplify f... |
1309 1310 |
* Read the user space value first so we can validate a few * things before proceeding further. |
1a52084d0 futex: split out ... |
1311 |
*/ |
af54d6a1c futex: Simplify f... |
1312 |
if (get_futex_value_locked(&uval, uaddr)) |
1a52084d0 futex: split out ... |
1313 |
return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
1314 1315 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
1a52084d0 futex: split out ... |
1316 1317 1318 |
/* * Detect deadlocks. */ |
af54d6a1c futex: Simplify f... |
1319 |
if ((unlikely((uval & FUTEX_TID_MASK) == vpid))) |
1a52084d0 futex: split out ... |
1320 |
return -EDEADLK; |
ab51fbab3 futex: Fault/erro... |
1321 1322 |
if ((unlikely(should_fail_futex(true)))) return -EDEADLK; |
1a52084d0 futex: split out ... |
1323 |
/* |
af54d6a1c futex: Simplify f... |
1324 1325 |
* Lookup existing state first. If it exists, try to attach to * its pi_state. |
1a52084d0 futex: split out ... |
1326 |
*/ |
499f5aca2 futex: Cleanup va... |
1327 1328 |
top_waiter = futex_top_waiter(hb, key); if (top_waiter) |
734009e96 futex: Change loc... |
1329 |
return attach_to_pi_state(uaddr, uval, top_waiter->pi_state, ps); |
1a52084d0 futex: split out ... |
1330 1331 |
/* |
af54d6a1c futex: Simplify f... |
1332 1333 1334 1335 |
* 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 ... |
1336 |
*/ |
af54d6a1c futex: Simplify f... |
1337 |
if (!(uval & FUTEX_TID_MASK)) { |
59fa62451 futex: Handle fut... |
1338 |
/* |
af54d6a1c futex: Simplify f... |
1339 1340 |
* 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... |
1341 |
*/ |
af54d6a1c futex: Simplify f... |
1342 1343 |
newval = uval & FUTEX_OWNER_DIED; newval |= vpid; |
1a52084d0 futex: split out ... |
1344 |
|
af54d6a1c futex: Simplify f... |
1345 1346 1347 1348 1349 1350 1351 1352 |
/* 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 ... |
1353 1354 |
/* |
af54d6a1c futex: Simplify f... |
1355 1356 1357 |
* 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 ... |
1358 |
*/ |
af54d6a1c futex: Simplify f... |
1359 1360 1361 1362 |
newval = uval | FUTEX_WAITERS; ret = lock_pi_update_atomic(uaddr, uval, newval); if (ret) return ret; |
1a52084d0 futex: split out ... |
1363 |
/* |
af54d6a1c futex: Simplify f... |
1364 1365 1366 |
* 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 ... |
1367 |
*/ |
3ef240eaf futex: Prevent ex... |
1368 |
return attach_to_pi_owner(uaddr, newval, key, ps, exiting); |
1a52084d0 futex: split out ... |
1369 |
} |
2e12978a9 futex,plist: Pass... |
1370 1371 1372 1373 1374 1375 1376 1377 1378 |
/** * __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... |
1379 |
if (WARN_ON_SMP(!q->lock_ptr) || WARN_ON(plist_node_empty(&q->list))) |
2e12978a9 futex,plist: Pass... |
1380 |
return; |
4de1a293a futex: Replace sp... |
1381 |
lockdep_assert_held(q->lock_ptr); |
2e12978a9 futex,plist: Pass... |
1382 1383 1384 |
hb = container_of(q->lock_ptr, struct futex_hash_bucket, lock); plist_del(&q->list, &hb->chain); |
11d4616bd futex: revert bac... |
1385 |
hb_waiters_dec(hb); |
2e12978a9 futex,plist: Pass... |
1386 |
} |
c87e2837b [PATCH] pi-futex:... |
1387 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1388 |
* The hash bucket lock must be held when this is called. |
1d0dcb3ad futex: Implement ... |
1389 1390 1391 |
* 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 |
1392 |
*/ |
1d0dcb3ad futex: Implement ... |
1393 |
static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q) |
1da177e4c Linux-2.6.12-rc2 |
1394 |
{ |
f1a11e057 futex: remove the... |
1395 |
struct task_struct *p = q->task; |
aa10990e0 futex: avoid wake... |
1396 1397 1398 |
if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex ")) return; |
b061c38be futex: Fix (possi... |
1399 |
get_task_struct(p); |
2e12978a9 futex,plist: Pass... |
1400 |
__unqueue_futex(q); |
1da177e4c Linux-2.6.12-rc2 |
1401 |
/* |
38fcd06e9 futex: Clarify ma... |
1402 1403 1404 1405 1406 |
* 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 |
1407 |
*/ |
1b367ece0 futex: Use smp_st... |
1408 |
smp_store_release(&q->lock_ptr, NULL); |
b061c38be futex: Fix (possi... |
1409 1410 1411 |
/* * Queue the task for later wakeup for after we've released |
751459043 futex: Drop lefto... |
1412 |
* the hb->lock. |
b061c38be futex: Fix (possi... |
1413 |
*/ |
07879c6a3 sched/wake_q: Red... |
1414 |
wake_q_add_safe(wake_q, p); |
1da177e4c Linux-2.6.12-rc2 |
1415 |
} |
16ffa12d7 futex: Pull rt_mu... |
1416 1417 1418 1419 |
/* * 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:... |
1420 |
{ |
3f649ab72 treewide: Remove ... |
1421 |
u32 curval, newval; |
16ffa12d7 futex: Pull rt_mu... |
1422 |
struct task_struct *new_owner; |
aa2bfe553 rtmutex: Clean up |
1423 |
bool postunlock = false; |
194a6b5b9 sched/wake_q: Ren... |
1424 |
DEFINE_WAKE_Q(wake_q); |
13fbca4c6 futex: Always cle... |
1425 |
int ret = 0; |
c87e2837b [PATCH] pi-futex:... |
1426 |
|
c87e2837b [PATCH] pi-futex:... |
1427 |
new_owner = rt_mutex_next_owner(&pi_state->pi_mutex); |
bebe5b514 futex: Futex_unlo... |
1428 |
if (WARN_ON_ONCE(!new_owner)) { |
16ffa12d7 futex: Pull rt_mu... |
1429 |
/* |
bebe5b514 futex: Futex_unlo... |
1430 |
* As per the comment in futex_unlock_pi() this should not happen. |
16ffa12d7 futex: Pull rt_mu... |
1431 1432 1433 1434 1435 1436 1437 1438 |
* * 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... |
1439 |
} |
c87e2837b [PATCH] pi-futex:... |
1440 1441 |
/* |
16ffa12d7 futex: Pull rt_mu... |
1442 1443 1444 |
* 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:... |
1445 |
*/ |
13fbca4c6 futex: Always cle... |
1446 |
newval = FUTEX_WAITERS | task_pid_vnr(new_owner); |
e3f2ddeac [PATCH] pi-futex:... |
1447 |
|
921c7ebd1 futex: Fix incorr... |
1448 |
if (unlikely(should_fail_futex(true))) { |
ab51fbab3 futex: Fault/erro... |
1449 |
ret = -EFAULT; |
921c7ebd1 futex: Fix incorr... |
1450 1451 |
goto out_unlock; } |
ab51fbab3 futex: Fault/erro... |
1452 |
|
6b4f4bc9c locking/futex: Al... |
1453 1454 |
ret = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (!ret && (curval != uval)) { |
89e9e66ba futex: Handle unl... |
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 |
/* * 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... |
1466 |
|
16ffa12d7 futex: Pull rt_mu... |
1467 1468 |
if (ret) goto out_unlock; |
c87e2837b [PATCH] pi-futex:... |
1469 |
|
94ffac5d8 futex: Fix small ... |
1470 1471 1472 1473 |
/* * 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... |
1474 |
raw_spin_lock(&pi_state->owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1475 1476 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
b4abf9104 rtmutex: Make wai... |
1477 |
raw_spin_unlock(&pi_state->owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1478 |
|
b4abf9104 rtmutex: Make wai... |
1479 |
raw_spin_lock(&new_owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1480 |
WARN_ON(!list_empty(&pi_state->list)); |
c87e2837b [PATCH] pi-futex:... |
1481 1482 |
list_add(&pi_state->list, &new_owner->pi_state_list); pi_state->owner = new_owner; |
b4abf9104 rtmutex: Make wai... |
1483 |
raw_spin_unlock(&new_owner->pi_lock); |
627371d73 [PATCH] pi-futex:... |
1484 |
|
aa2bfe553 rtmutex: Clean up |
1485 |
postunlock = __rt_mutex_futex_unlock(&pi_state->pi_mutex, &wake_q); |
5293c2efd futex,rt_mutex: P... |
1486 |
|
16ffa12d7 futex: Pull rt_mu... |
1487 |
out_unlock: |
5293c2efd futex,rt_mutex: P... |
1488 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
5293c2efd futex,rt_mutex: P... |
1489 |
|
aa2bfe553 rtmutex: Clean up |
1490 1491 |
if (postunlock) rt_mutex_postunlock(&wake_q); |
c87e2837b [PATCH] pi-futex:... |
1492 |
|
16ffa12d7 futex: Pull rt_mu... |
1493 |
return ret; |
c87e2837b [PATCH] pi-futex:... |
1494 |
} |
1da177e4c Linux-2.6.12-rc2 |
1495 |
/* |
8b8f319fc [PATCH] lockdep: ... |
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 |
* 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... |
1510 1511 1512 |
static inline void double_unlock_hb(struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2) { |
f061d3515 futex: remove the... |
1513 |
spin_unlock(&hb1->lock); |
88f502fed futex: remove the... |
1514 1515 |
if (hb1 != hb2) spin_unlock(&hb2->lock); |
5eb3dc62f futex: add double... |
1516 |
} |
8b8f319fc [PATCH] lockdep: ... |
1517 |
/* |
b2d0994b1 futex: update fut... |
1518 |
* Wake up waiters matching bitset queued on this futex (uaddr). |
1da177e4c Linux-2.6.12-rc2 |
1519 |
*/ |
b41277dc7 futex: Replace fs... |
1520 1521 |
static int futex_wake(u32 __user *uaddr, unsigned int flags, int nr_wake, u32 bitset) |
1da177e4c Linux-2.6.12-rc2 |
1522 |
{ |
e2970f2fb [PATCH] pi-futex:... |
1523 |
struct futex_hash_bucket *hb; |
1da177e4c Linux-2.6.12-rc2 |
1524 |
struct futex_q *this, *next; |
38d47c1b7 futex: rely on ge... |
1525 |
union futex_key key = FUTEX_KEY_INIT; |
1da177e4c Linux-2.6.12-rc2 |
1526 |
int ret; |
194a6b5b9 sched/wake_q: Ren... |
1527 |
DEFINE_WAKE_Q(wake_q); |
1da177e4c Linux-2.6.12-rc2 |
1528 |
|
cd689985c futex: Add bitset... |
1529 1530 |
if (!bitset) return -EINVAL; |
96d4f267e Remove 'type' arg... |
1531 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &key, FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1532 |
if (unlikely(ret != 0)) |
d7c5ed73b futex: Remove nee... |
1533 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1534 |
|
e2970f2fb [PATCH] pi-futex:... |
1535 |
hb = hash_futex(&key); |
b0c29f79e futexes: Avoid ta... |
1536 1537 1538 |
/* Make sure we really have tasks to wakeup */ if (!hb_waiters_pending(hb)) |
d7c5ed73b futex: Remove nee... |
1539 |
return ret; |
b0c29f79e futexes: Avoid ta... |
1540 |
|
e2970f2fb [PATCH] pi-futex:... |
1541 |
spin_lock(&hb->lock); |
1da177e4c Linux-2.6.12-rc2 |
1542 |
|
0d00c7b20 futexes: Clean up... |
1543 |
plist_for_each_entry_safe(this, next, &hb->chain, list) { |
1da177e4c Linux-2.6.12-rc2 |
1544 |
if (match_futex (&this->key, &key)) { |
52400ba94 futex: add requeu... |
1545 |
if (this->pi_state || this->rt_waiter) { |
ed6f7b10e [PATCH] pi-futex:... |
1546 1547 1548 |
ret = -EINVAL; break; } |
cd689985c futex: Add bitset... |
1549 1550 1551 1552 |
/* Check if one of the bits is set in both bitsets */ if (!(this->bitset & bitset)) continue; |
1d0dcb3ad futex: Implement ... |
1553 |
mark_wake_futex(&wake_q, this); |
1da177e4c Linux-2.6.12-rc2 |
1554 1555 1556 1557 |
if (++ret >= nr_wake) break; } } |
e2970f2fb [PATCH] pi-futex:... |
1558 |
spin_unlock(&hb->lock); |
1d0dcb3ad futex: Implement ... |
1559 |
wake_up_q(&wake_q); |
1da177e4c Linux-2.6.12-rc2 |
1560 1561 |
return ret; } |
30d6e0a41 futex: Remove dup... |
1562 1563 1564 1565 |
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... |
1566 1567 |
int oparg = sign_extend32((encoded_op & 0x00fff000) >> 12, 11); int cmparg = sign_extend32(encoded_op & 0x00000fff, 11); |
30d6e0a41 futex: Remove dup... |
1568 1569 1570 |
int oldval, ret; if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28)) { |
e78c38f6b futex: futex_wake... |
1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 |
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... |
1582 1583 |
oparg = 1 << oparg; } |
a08971e94 futex: arch_futex... |
1584 |
pagefault_disable(); |
30d6e0a41 futex: Remove dup... |
1585 |
ret = arch_futex_atomic_op_inuser(op, oparg, &oldval, uaddr); |
a08971e94 futex: arch_futex... |
1586 |
pagefault_enable(); |
30d6e0a41 futex: Remove dup... |
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 |
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 |
1607 |
/* |
4732efbeb [PATCH] FUTEX_WAK... |
1608 1609 1610 |
* Wake up all waiters hashed on the physical page that is mapped * to this virtual address: */ |
e2970f2fb [PATCH] pi-futex:... |
1611 |
static int |
b41277dc7 futex: Replace fs... |
1612 |
futex_wake_op(u32 __user *uaddr1, unsigned int flags, u32 __user *uaddr2, |
e2970f2fb [PATCH] pi-futex:... |
1613 |
int nr_wake, int nr_wake2, int op) |
4732efbeb [PATCH] FUTEX_WAK... |
1614 |
{ |
38d47c1b7 futex: rely on ge... |
1615 |
union futex_key key1 = FUTEX_KEY_INIT, key2 = FUTEX_KEY_INIT; |
e2970f2fb [PATCH] pi-futex:... |
1616 |
struct futex_hash_bucket *hb1, *hb2; |
4732efbeb [PATCH] FUTEX_WAK... |
1617 |
struct futex_q *this, *next; |
e4dc5b7a3 futex: clean up f... |
1618 |
int ret, op_ret; |
194a6b5b9 sched/wake_q: Ren... |
1619 |
DEFINE_WAKE_Q(wake_q); |
4732efbeb [PATCH] FUTEX_WAK... |
1620 |
|
e4dc5b7a3 futex: clean up f... |
1621 |
retry: |
96d4f267e Remove 'type' arg... |
1622 |
ret = get_futex_key(uaddr1, flags & FLAGS_SHARED, &key1, FUTEX_READ); |
4732efbeb [PATCH] FUTEX_WAK... |
1623 |
if (unlikely(ret != 0)) |
d7c5ed73b futex: Remove nee... |
1624 |
return ret; |
96d4f267e Remove 'type' arg... |
1625 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, FUTEX_WRITE); |
4732efbeb [PATCH] FUTEX_WAK... |
1626 |
if (unlikely(ret != 0)) |
d7c5ed73b futex: Remove nee... |
1627 |
return ret; |
4732efbeb [PATCH] FUTEX_WAK... |
1628 |
|
e2970f2fb [PATCH] pi-futex:... |
1629 1630 |
hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); |
4732efbeb [PATCH] FUTEX_WAK... |
1631 |
|
e4dc5b7a3 futex: clean up f... |
1632 |
retry_private: |
eaaea8036 futex: Fix lockin... |
1633 |
double_lock_hb(hb1, hb2); |
e2970f2fb [PATCH] pi-futex:... |
1634 |
op_ret = futex_atomic_op_inuser(op, uaddr2); |
4732efbeb [PATCH] FUTEX_WAK... |
1635 |
if (unlikely(op_ret < 0)) { |
5eb3dc62f futex: add double... |
1636 |
double_unlock_hb(hb1, hb2); |
4732efbeb [PATCH] FUTEX_WAK... |
1637 |
|
6b4f4bc9c locking/futex: Al... |
1638 1639 1640 1641 1642 1643 |
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... |
1644 |
ret = op_ret; |
d7c5ed73b futex: Remove nee... |
1645 |
return ret; |
796f8d9b9 [PATCH] FUTEX_WAK... |
1646 |
} |
6b4f4bc9c locking/futex: Al... |
1647 1648 1649 |
if (op_ret == -EFAULT) { ret = fault_in_user_writeable(uaddr2); if (ret) |
d7c5ed73b futex: Remove nee... |
1650 |
return ret; |
6b4f4bc9c locking/futex: Al... |
1651 |
} |
4732efbeb [PATCH] FUTEX_WAK... |
1652 |
|
6b4f4bc9c locking/futex: Al... |
1653 1654 |
if (!(flags & FLAGS_SHARED)) { cond_resched(); |
e4dc5b7a3 futex: clean up f... |
1655 |
goto retry_private; |
6b4f4bc9c locking/futex: Al... |
1656 |
} |
e4dc5b7a3 futex: clean up f... |
1657 |
|
6b4f4bc9c locking/futex: Al... |
1658 |
cond_resched(); |
e4dc5b7a3 futex: clean up f... |
1659 |
goto retry; |
4732efbeb [PATCH] FUTEX_WAK... |
1660 |
} |
0d00c7b20 futexes: Clean up... |
1661 |
plist_for_each_entry_safe(this, next, &hb1->chain, list) { |
4732efbeb [PATCH] FUTEX_WAK... |
1662 |
if (match_futex (&this->key, &key1)) { |
aa10990e0 futex: avoid wake... |
1663 1664 1665 1666 |
if (this->pi_state || this->rt_waiter) { ret = -EINVAL; goto out_unlock; } |
1d0dcb3ad futex: Implement ... |
1667 |
mark_wake_futex(&wake_q, this); |
4732efbeb [PATCH] FUTEX_WAK... |
1668 1669 1670 1671 1672 1673 |
if (++ret >= nr_wake) break; } } if (op_ret > 0) { |
4732efbeb [PATCH] FUTEX_WAK... |
1674 |
op_ret = 0; |
0d00c7b20 futexes: Clean up... |
1675 |
plist_for_each_entry_safe(this, next, &hb2->chain, list) { |
4732efbeb [PATCH] FUTEX_WAK... |
1676 |
if (match_futex (&this->key, &key2)) { |
aa10990e0 futex: avoid wake... |
1677 1678 1679 1680 |
if (this->pi_state || this->rt_waiter) { ret = -EINVAL; goto out_unlock; } |
1d0dcb3ad futex: Implement ... |
1681 |
mark_wake_futex(&wake_q, this); |
4732efbeb [PATCH] FUTEX_WAK... |
1682 1683 1684 1685 1686 1687 |
if (++op_ret >= nr_wake2) break; } } ret += op_ret; } |
aa10990e0 futex: avoid wake... |
1688 |
out_unlock: |
5eb3dc62f futex: add double... |
1689 |
double_unlock_hb(hb1, hb2); |
1d0dcb3ad futex: Implement ... |
1690 |
wake_up_q(&wake_q); |
4732efbeb [PATCH] FUTEX_WAK... |
1691 1692 |
return ret; } |
9121e4783 futex: distangle ... |
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 |
/** * 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... |
1711 |
hb_waiters_dec(hb1); |
11d4616bd futex: revert bac... |
1712 |
hb_waiters_inc(hb2); |
fe1bce9e2 futex: Acknowledg... |
1713 |
plist_add(&q->list, &hb2->chain); |
9121e4783 futex: distangle ... |
1714 |
q->lock_ptr = &hb2->lock; |
9121e4783 futex: distangle ... |
1715 |
} |
9121e4783 futex: distangle ... |
1716 1717 |
q->key = *key2; } |
52400ba94 futex: add requeu... |
1718 1719 |
/** * requeue_pi_wake_futex() - Wake a task that acquired the lock during requeue |
d96ee56ce futex: Make funct... |
1720 1721 1722 |
* @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... |
1723 1724 1725 1726 1727 |
* * 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... |
1728 1729 1730 |
* 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... |
1731 1732 |
*/ static inline |
beda2c7ea futex: Update fut... |
1733 1734 |
void requeue_pi_wake_futex(struct futex_q *q, union futex_key *key, struct futex_hash_bucket *hb) |
52400ba94 futex: add requeu... |
1735 |
{ |
52400ba94 futex: add requeu... |
1736 |
q->key = *key; |
2e12978a9 futex,plist: Pass... |
1737 |
__unqueue_futex(q); |
52400ba94 futex: add requeu... |
1738 1739 1740 |
WARN_ON(!q->rt_waiter); q->rt_waiter = NULL; |
beda2c7ea futex: Update fut... |
1741 |
q->lock_ptr = &hb->lock; |
beda2c7ea futex: Update fut... |
1742 |
|
f1a11e057 futex: remove the... |
1743 |
wake_up_state(q->task, TASK_NORMAL); |
52400ba94 futex: add requeu... |
1744 1745 1746 1747 |
} /** * futex_proxy_trylock_atomic() - Attempt an atomic lock for the top waiter |
bab5bc9e8 futex: fixup unlo... |
1748 1749 1750 1751 1752 1753 |
* @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 |
3ef240eaf futex: Prevent ex... |
1754 1755 |
* @exiting: Pointer to store the task pointer of the owner task * which is in the middle of exiting |
bab5bc9e8 futex: fixup unlo... |
1756 |
* @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) |
52400ba94 futex: add requeu... |
1757 1758 |
* * Try and get the lock on behalf of the top waiter if we can do it atomically. |
bab5bc9e8 futex: fixup unlo... |
1759 1760 1761 |
* 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... |
1762 |
* |
3ef240eaf futex: Prevent ex... |
1763 1764 1765 1766 |
* @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... |
1767 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1768 1769 1770 |
* - 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... |
1771 |
*/ |
3ef240eaf futex: Prevent ex... |
1772 1773 1774 1775 1776 |
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... |
1777 |
{ |
bab5bc9e8 futex: fixup unlo... |
1778 |
struct futex_q *top_waiter = NULL; |
52400ba94 futex: add requeu... |
1779 |
u32 curval; |
866293ee5 futex: Add anothe... |
1780 |
int ret, vpid; |
52400ba94 futex: add requeu... |
1781 1782 1783 |
if (get_futex_value_locked(&curval, pifutex)) return -EFAULT; |
ab51fbab3 futex: Fault/erro... |
1784 1785 |
if (unlikely(should_fail_futex(true))) return -EFAULT; |
bab5bc9e8 futex: fixup unlo... |
1786 1787 1788 1789 1790 1791 1792 1793 |
/* * 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... |
1794 1795 1796 1797 1798 |
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... |
1799 1800 1801 |
/* Ensure we requeue to the expected futex. */ if (!match_futex(top_waiter->requeue_pi_key, key2)) return -EINVAL; |
52400ba94 futex: add requeu... |
1802 |
/* |
bab5bc9e8 futex: fixup unlo... |
1803 1804 1805 |
* 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... |
1806 |
*/ |
866293ee5 futex: Add anothe... |
1807 |
vpid = task_pid_vnr(top_waiter->task); |
bab5bc9e8 futex: fixup unlo... |
1808 |
ret = futex_lock_pi_atomic(pifutex, hb2, key2, ps, top_waiter->task, |
3ef240eaf futex: Prevent ex... |
1809 |
exiting, set_waiters); |
866293ee5 futex: Add anothe... |
1810 |
if (ret == 1) { |
beda2c7ea futex: Update fut... |
1811 |
requeue_pi_wake_futex(top_waiter, key2, hb2); |
866293ee5 futex: Add anothe... |
1812 1813 |
return vpid; } |
52400ba94 futex: add requeu... |
1814 1815 1816 1817 1818 |
return ret; } /** * futex_requeue() - Requeue waiters from uaddr1 to uaddr2 |
fb62db2ba futex: Fix kernel... |
1819 |
* @uaddr1: source futex user address |
b41277dc7 futex: Replace fs... |
1820 |
* @flags: futex flags (FLAGS_SHARED, etc.) |
fb62db2ba futex: Fix kernel... |
1821 1822 1823 1824 1825 |
* @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... |
1826 |
* pi futex (pi to pi requeue is not supported) |
52400ba94 futex: add requeu... |
1827 1828 1829 1830 |
* * 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... |
1831 |
* Return: |
7b4ff1adb mutex, futex: adj... |
1832 1833 |
* - >=0 - on success, the number of tasks requeued or woken; * - <0 - on error |
1da177e4c Linux-2.6.12-rc2 |
1834 |
*/ |
b41277dc7 futex: Replace fs... |
1835 1836 1837 |
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 |
1838 |
{ |
38d47c1b7 futex: rely on ge... |
1839 |
union futex_key key1 = FUTEX_KEY_INIT, key2 = FUTEX_KEY_INIT; |
4b39f99c2 futex: Remove {ge... |
1840 |
int task_count = 0, ret; |
52400ba94 futex: add requeu... |
1841 |
struct futex_pi_state *pi_state = NULL; |
e2970f2fb [PATCH] pi-futex:... |
1842 |
struct futex_hash_bucket *hb1, *hb2; |
1da177e4c Linux-2.6.12-rc2 |
1843 |
struct futex_q *this, *next; |
194a6b5b9 sched/wake_q: Ren... |
1844 |
DEFINE_WAKE_Q(wake_q); |
52400ba94 futex: add requeu... |
1845 |
|
fbe0e839d futex: Prevent ov... |
1846 1847 |
if (nr_wake < 0 || nr_requeue < 0) return -EINVAL; |
bc2eecd7e futex: Allow for ... |
1848 1849 1850 1851 1852 1853 1854 1855 |
/* * 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... |
1856 1857 |
if (requeue_pi) { /* |
e9c243a5a futex-prevent-req... |
1858 1859 1860 1861 1862 1863 1864 |
* 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... |
1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 |
* 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 |
1883 |
|
42d35d48c futex: make futex... |
1884 |
retry: |
96d4f267e Remove 'type' arg... |
1885 |
ret = get_futex_key(uaddr1, flags & FLAGS_SHARED, &key1, FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1886 |
if (unlikely(ret != 0)) |
d7c5ed73b futex: Remove nee... |
1887 |
return ret; |
9ea71503a futex: Fix regres... |
1888 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, |
96d4f267e Remove 'type' arg... |
1889 |
requeue_pi ? FUTEX_WRITE : FUTEX_READ); |
1da177e4c Linux-2.6.12-rc2 |
1890 |
if (unlikely(ret != 0)) |
d7c5ed73b futex: Remove nee... |
1891 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1892 |
|
e9c243a5a futex-prevent-req... |
1893 1894 1895 1896 |
/* * The check above which compares uaddrs is not sufficient for * shared futexes. We need to compare the keys: */ |
d7c5ed73b futex: Remove nee... |
1897 1898 |
if (requeue_pi && match_futex(&key1, &key2)) return -EINVAL; |
e9c243a5a futex-prevent-req... |
1899 |
|
e2970f2fb [PATCH] pi-futex:... |
1900 1901 |
hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); |
1da177e4c Linux-2.6.12-rc2 |
1902 |
|
e4dc5b7a3 futex: clean up f... |
1903 |
retry_private: |
69cd9eba3 futex: avoid race... |
1904 |
hb_waiters_inc(hb2); |
8b8f319fc [PATCH] lockdep: ... |
1905 |
double_lock_hb(hb1, hb2); |
1da177e4c Linux-2.6.12-rc2 |
1906 |
|
e2970f2fb [PATCH] pi-futex:... |
1907 1908 |
if (likely(cmpval != NULL)) { u32 curval; |
1da177e4c Linux-2.6.12-rc2 |
1909 |
|
e2970f2fb [PATCH] pi-futex:... |
1910 |
ret = get_futex_value_locked(&curval, uaddr1); |
1da177e4c Linux-2.6.12-rc2 |
1911 1912 |
if (unlikely(ret)) { |
5eb3dc62f futex: add double... |
1913 |
double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
1914 |
hb_waiters_dec(hb2); |
1da177e4c Linux-2.6.12-rc2 |
1915 |
|
e2970f2fb [PATCH] pi-futex:... |
1916 |
ret = get_user(curval, uaddr1); |
e4dc5b7a3 futex: clean up f... |
1917 |
if (ret) |
d7c5ed73b futex: Remove nee... |
1918 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1919 |
|
b41277dc7 futex: Replace fs... |
1920 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
1921 |
goto retry_private; |
1da177e4c Linux-2.6.12-rc2 |
1922 |
|
e4dc5b7a3 futex: clean up f... |
1923 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
1924 |
} |
e2970f2fb [PATCH] pi-futex:... |
1925 |
if (curval != *cmpval) { |
1da177e4c Linux-2.6.12-rc2 |
1926 1927 1928 1929 |
ret = -EAGAIN; goto out_unlock; } } |
52400ba94 futex: add requeu... |
1930 |
if (requeue_pi && (task_count - nr_wake < nr_requeue)) { |
3ef240eaf futex: Prevent ex... |
1931 |
struct task_struct *exiting = NULL; |
bab5bc9e8 futex: fixup unlo... |
1932 1933 1934 1935 1936 1937 |
/* * 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... |
1938 |
ret = futex_proxy_trylock_atomic(uaddr2, hb1, hb2, &key1, |
3ef240eaf futex: Prevent ex... |
1939 1940 |
&key2, &pi_state, &exiting, nr_requeue); |
52400ba94 futex: add requeu... |
1941 1942 1943 1944 1945 |
/* * 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... |
1946 1947 |
* reference to it. If the lock was taken, ret contains the * vpid of the top waiter task. |
ecb38b78f futex: Document p... |
1948 1949 |
* 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... |
1950 |
*/ |
866293ee5 futex: Add anothe... |
1951 |
if (ret > 0) { |
52400ba94 futex: add requeu... |
1952 1953 |
WARN_ON(pi_state); task_count++; |
866293ee5 futex: Add anothe... |
1954 |
/* |
ecb38b78f futex: Document p... |
1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 |
* 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... |
1965 |
*/ |
3ef240eaf futex: Prevent ex... |
1966 1967 |
ret = lookup_pi_state(uaddr2, ret, hb2, &key2, &pi_state, &exiting); |
52400ba94 futex: add requeu... |
1968 1969 1970 1971 |
} switch (ret) { case 0: |
ecb38b78f futex: Document p... |
1972 |
/* We hold a reference on the pi state. */ |
52400ba94 futex: add requeu... |
1973 |
break; |
4959f2de1 futex: Remove poi... |
1974 1975 |
/* If the above failed, then pi_state is NULL */ |
52400ba94 futex: add requeu... |
1976 1977 |
case -EFAULT: double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
1978 |
hb_waiters_dec(hb2); |
d0725992c futex: Fix the wr... |
1979 |
ret = fault_in_user_writeable(uaddr2); |
52400ba94 futex: add requeu... |
1980 1981 |
if (!ret) goto retry; |
d7c5ed73b futex: Remove nee... |
1982 |
return ret; |
ac31c7ff8 futex: Provide di... |
1983 |
case -EBUSY: |
52400ba94 futex: add requeu... |
1984 |
case -EAGAIN: |
af54d6a1c futex: Simplify f... |
1985 1986 |
/* * Two reasons for this: |
ac31c7ff8 futex: Provide di... |
1987 |
* - EBUSY: Owner is exiting and we just wait for the |
af54d6a1c futex: Simplify f... |
1988 |
* exit to complete. |
ac31c7ff8 futex: Provide di... |
1989 |
* - EAGAIN: The user space value changed. |
af54d6a1c futex: Simplify f... |
1990 |
*/ |
52400ba94 futex: add requeu... |
1991 |
double_unlock_hb(hb1, hb2); |
69cd9eba3 futex: avoid race... |
1992 |
hb_waiters_dec(hb2); |
3ef240eaf futex: Prevent ex... |
1993 1994 1995 1996 1997 1998 |
/* * 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... |
1999 2000 2001 2002 2003 2004 |
cond_resched(); goto retry; default: goto out_unlock; } } |
0d00c7b20 futexes: Clean up... |
2005 |
plist_for_each_entry_safe(this, next, &hb1->chain, list) { |
52400ba94 futex: add requeu... |
2006 2007 2008 2009 |
if (task_count - nr_wake >= nr_requeue) break; if (!match_futex(&this->key, &key1)) |
1da177e4c Linux-2.6.12-rc2 |
2010 |
continue; |
52400ba94 futex: add requeu... |
2011 |
|
392741e0a futex: Fix handli... |
2012 2013 2014 |
/* * 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... |
2015 2016 2017 |
* * We should never be requeueing a futex_q with a pi_state, * which is awaiting a futex_unlock_pi(). |
392741e0a futex: Fix handli... |
2018 2019 |
*/ if ((requeue_pi && !this->rt_waiter) || |
aa10990e0 futex: avoid wake... |
2020 2021 |
(!requeue_pi && this->rt_waiter) || this->pi_state) { |
392741e0a futex: Fix handli... |
2022 2023 2024 |
ret = -EINVAL; break; } |
52400ba94 futex: add requeu... |
2025 2026 2027 2028 2029 2030 2031 |
/* * 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 ... |
2032 |
mark_wake_futex(&wake_q, this); |
52400ba94 futex: add requeu... |
2033 2034 |
continue; } |
1da177e4c Linux-2.6.12-rc2 |
2035 |
|
84bc4af59 futex: Detect mis... |
2036 2037 2038 2039 2040 |
/* 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... |
2041 2042 2043 2044 2045 |
/* * 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... |
2046 2047 2048 2049 2050 |
/* * 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... |
2051 |
get_pi_state(pi_state); |
52400ba94 futex: add requeu... |
2052 2053 2054 |
this->pi_state = pi_state; ret = rt_mutex_start_proxy_lock(&pi_state->pi_mutex, this->rt_waiter, |
c051b21f7 rtmutex: Confine ... |
2055 |
this->task); |
52400ba94 futex: add requeu... |
2056 |
if (ret == 1) { |
ecb38b78f futex: Document p... |
2057 2058 2059 2060 2061 2062 2063 2064 |
/* * 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... |
2065 |
requeue_pi_wake_futex(this, &key2, hb2); |
52400ba94 futex: add requeu... |
2066 2067 |
continue; } else if (ret) { |
ecb38b78f futex: Document p... |
2068 2069 2070 2071 2072 2073 2074 2075 |
/* * 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... |
2076 |
this->pi_state = NULL; |
29e9ee5d4 futex: Rename fre... |
2077 |
put_pi_state(pi_state); |
885c2cb77 futex: Cleanup th... |
2078 2079 2080 2081 2082 |
/* * We stop queueing more waiters and let user * space deal with the mess. */ break; |
52400ba94 futex: add requeu... |
2083 |
} |
1da177e4c Linux-2.6.12-rc2 |
2084 |
} |
52400ba94 futex: add requeu... |
2085 |
requeue_futex(this, hb1, hb2, &key2); |
1da177e4c Linux-2.6.12-rc2 |
2086 |
} |
ecb38b78f futex: Document p... |
2087 2088 2089 2090 2091 |
/* * 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... |
2092 |
put_pi_state(pi_state); |
885c2cb77 futex: Cleanup th... |
2093 2094 |
out_unlock: |
5eb3dc62f futex: add double... |
2095 |
double_unlock_hb(hb1, hb2); |
1d0dcb3ad futex: Implement ... |
2096 |
wake_up_q(&wake_q); |
69cd9eba3 futex: avoid race... |
2097 |
hb_waiters_dec(hb2); |
52400ba94 futex: add requeu... |
2098 |
return ret ? ret : task_count; |
1da177e4c Linux-2.6.12-rc2 |
2099 2100 2101 |
} /* The key must be already stored in q->key. */ |
82af7aca5 Removal of FUTEX_FD |
2102 |
static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) |
15e408cd6 futex: Add lock c... |
2103 |
__acquires(&hb->lock) |
1da177e4c Linux-2.6.12-rc2 |
2104 |
{ |
e2970f2fb [PATCH] pi-futex:... |
2105 |
struct futex_hash_bucket *hb; |
1da177e4c Linux-2.6.12-rc2 |
2106 |
|
e2970f2fb [PATCH] pi-futex:... |
2107 |
hb = hash_futex(&q->key); |
11d4616bd futex: revert bac... |
2108 2109 2110 2111 2112 2113 2114 2115 2116 |
/* * 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... |
2117 |
hb_waiters_inc(hb); /* implies smp_mb(); (A) */ |
11d4616bd futex: revert bac... |
2118 |
|
e2970f2fb [PATCH] pi-futex:... |
2119 |
q->lock_ptr = &hb->lock; |
1da177e4c Linux-2.6.12-rc2 |
2120 |
|
6f568ebe2 futex: Fix barrie... |
2121 |
spin_lock(&hb->lock); |
e2970f2fb [PATCH] pi-futex:... |
2122 |
return hb; |
1da177e4c Linux-2.6.12-rc2 |
2123 |
} |
d40d65c8d futex: Correct qu... |
2124 |
static inline void |
0d00c7b20 futexes: Clean up... |
2125 |
queue_unlock(struct futex_hash_bucket *hb) |
15e408cd6 futex: Add lock c... |
2126 |
__releases(&hb->lock) |
d40d65c8d futex: Correct qu... |
2127 2128 |
{ spin_unlock(&hb->lock); |
11d4616bd futex: revert bac... |
2129 |
hb_waiters_dec(hb); |
d40d65c8d futex: Correct qu... |
2130 |
} |
cfafcd117 futex: Rework fut... |
2131 |
static inline void __queue_me(struct futex_q *q, struct futex_hash_bucket *hb) |
1da177e4c Linux-2.6.12-rc2 |
2132 |
{ |
ec92d0829 futex priority ba... |
2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 |
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... |
2146 |
plist_add(&q->list, &hb->chain); |
c87e2837b [PATCH] pi-futex:... |
2147 |
q->task = current; |
cfafcd117 futex: Rework fut... |
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 |
} /** * 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:... |
2166 |
spin_unlock(&hb->lock); |
1da177e4c Linux-2.6.12-rc2 |
2167 |
} |
d40d65c8d futex: Correct qu... |
2168 2169 2170 2171 2172 2173 2174 |
/** * 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... |
2175 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2176 2177 |
* - 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 |
2178 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
2179 2180 |
static int unqueue_me(struct futex_q *q) { |
1da177e4c Linux-2.6.12-rc2 |
2181 |
spinlock_t *lock_ptr; |
e2970f2fb [PATCH] pi-futex:... |
2182 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
2183 2184 |
/* In the common case we don't take the spinlock, which is nice. */ |
42d35d48c futex: make futex... |
2185 |
retry: |
29b75eb2d futex: Replace ba... |
2186 2187 2188 2189 2190 2191 |
/* * 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... |
2192 |
if (lock_ptr != NULL) { |
1da177e4c Linux-2.6.12-rc2 |
2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 |
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... |
2211 |
__unqueue_futex(q); |
c87e2837b [PATCH] pi-futex:... |
2212 2213 |
BUG_ON(q->pi_state); |
1da177e4c Linux-2.6.12-rc2 |
2214 2215 2216 |
spin_unlock(lock_ptr); ret = 1; } |
1da177e4c Linux-2.6.12-rc2 |
2217 2218 |
return ret; } |
c87e2837b [PATCH] pi-futex:... |
2219 2220 |
/* * PI futexes can not be requeued and must remove themself from the |
d0aa7a70b futex_requeue_pi ... |
2221 2222 |
* hash bucket. The hash bucket lock (i.e. lock_ptr) is held on entry * and dropped here. |
c87e2837b [PATCH] pi-futex:... |
2223 |
*/ |
d0aa7a70b futex_requeue_pi ... |
2224 |
static void unqueue_me_pi(struct futex_q *q) |
15e408cd6 futex: Add lock c... |
2225 |
__releases(q->lock_ptr) |
c87e2837b [PATCH] pi-futex:... |
2226 |
{ |
2e12978a9 futex,plist: Pass... |
2227 |
__unqueue_futex(q); |
c87e2837b [PATCH] pi-futex:... |
2228 2229 |
BUG_ON(!q->pi_state); |
29e9ee5d4 futex: Rename fre... |
2230 |
put_pi_state(q->pi_state); |
c87e2837b [PATCH] pi-futex:... |
2231 |
q->pi_state = NULL; |
d0aa7a70b futex_requeue_pi ... |
2232 |
spin_unlock(q->lock_ptr); |
c87e2837b [PATCH] pi-futex:... |
2233 |
} |
778e9a9c3 pi-futex: fix exi... |
2234 |
static int fixup_pi_state_owner(u32 __user *uaddr, struct futex_q *q, |
c1e2f0eaf futex: Avoid viol... |
2235 |
struct task_struct *argowner) |
d0aa7a70b futex_requeue_pi ... |
2236 |
{ |
d0aa7a70b futex_requeue_pi ... |
2237 |
struct futex_pi_state *pi_state = q->pi_state; |
3f649ab72 treewide: Remove ... |
2238 |
u32 uval, curval, newval; |
c1e2f0eaf futex: Avoid viol... |
2239 2240 |
struct task_struct *oldowner, *newowner; u32 newtid; |
6b4f4bc9c locking/futex: Al... |
2241 |
int ret, err = 0; |
d0aa7a70b futex_requeue_pi ... |
2242 |
|
c1e2f0eaf futex: Avoid viol... |
2243 |
lockdep_assert_held(q->lock_ptr); |
734009e96 futex: Change loc... |
2244 2245 2246 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); oldowner = pi_state->owner; |
1b7558e45 futexes: fix faul... |
2247 2248 |
/* |
c1e2f0eaf futex: Avoid viol... |
2249 |
* We are here because either: |
16ffa12d7 futex: Pull rt_mu... |
2250 |
* |
c1e2f0eaf futex: Avoid viol... |
2251 2252 2253 2254 2255 2256 2257 2258 2259 |
* - 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... |
2260 |
* This must be atomic as we have to preserve the owner died bit here. |
1b7558e45 futexes: fix faul... |
2261 |
* |
b2d0994b1 futex: update fut... |
2262 2263 2264 |
* 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... |
2265 |
* |
734009e96 futex: Change loc... |
2266 2267 2268 2269 |
* 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... |
2270 2271 |
*/ retry: |
c1e2f0eaf futex: Avoid viol... |
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 |
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; } /* |
9f5d1c336 futex: Handle tra... |
2289 2290 |
* The trylock just failed, so either there is an owner or * there is a higher priority waiter than this one. |
c1e2f0eaf futex: Avoid viol... |
2291 2292 |
*/ newowner = rt_mutex_owner(&pi_state->pi_mutex); |
9f5d1c336 futex: Handle tra... |
2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 |
/* * If the higher priority waiter has not yet taken over the * rtmutex then newowner is NULL. We can't return here with * that state because it's inconsistent vs. the user space * state. So drop the locks and try again. It's a valid * situation and not any different from the other retry * conditions. */ if (unlikely(!newowner)) { err = -EAGAIN; goto handle_err; } |
c1e2f0eaf futex: Avoid viol... |
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 |
} 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_... |
2319 2320 2321 |
/* Owner died? */ if (!pi_state->owner) newtid |= FUTEX_OWNER_DIED; |
c1e2f0eaf futex: Avoid viol... |
2322 |
|
6b4f4bc9c locking/futex: Al... |
2323 2324 2325 |
err = get_futex_value_locked(&uval, uaddr); if (err) goto handle_err; |
1b7558e45 futexes: fix faul... |
2326 |
|
16ffa12d7 futex: Pull rt_mu... |
2327 |
for (;;) { |
1b7558e45 futexes: fix faul... |
2328 |
newval = (uval & FUTEX_OWNER_DIED) | newtid; |
6b4f4bc9c locking/futex: Al... |
2329 2330 2331 |
err = cmpxchg_futex_value_locked(&curval, uaddr, uval, newval); if (err) goto handle_err; |
1b7558e45 futexes: fix faul... |
2332 2333 2334 2335 2336 2337 2338 2339 2340 |
if (curval == uval) break; uval = curval; } /* * We fixed up user space. Now we need to fix the pi_state * itself. */ |
d0aa7a70b futex_requeue_pi ... |
2341 |
if (pi_state->owner != NULL) { |
734009e96 futex: Change loc... |
2342 |
raw_spin_lock(&pi_state->owner->pi_lock); |
d0aa7a70b futex_requeue_pi ... |
2343 2344 |
WARN_ON(list_empty(&pi_state->list)); list_del_init(&pi_state->list); |
734009e96 futex: Change loc... |
2345 |
raw_spin_unlock(&pi_state->owner->pi_lock); |
1b7558e45 futexes: fix faul... |
2346 |
} |
d0aa7a70b futex_requeue_pi ... |
2347 |
|
cdf71a10c futex: Prevent st... |
2348 |
pi_state->owner = newowner; |
d0aa7a70b futex_requeue_pi ... |
2349 |
|
734009e96 futex: Change loc... |
2350 |
raw_spin_lock(&newowner->pi_lock); |
d0aa7a70b futex_requeue_pi ... |
2351 |
WARN_ON(!list_empty(&pi_state->list)); |
cdf71a10c futex: Prevent st... |
2352 |
list_add(&pi_state->list, &newowner->pi_state_list); |
734009e96 futex: Change loc... |
2353 2354 |
raw_spin_unlock(&newowner->pi_lock); raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
1b7558e45 futexes: fix faul... |
2355 |
return 0; |
d0aa7a70b futex_requeue_pi ... |
2356 |
|
d0aa7a70b futex_requeue_pi ... |
2357 |
/* |
6b4f4bc9c locking/futex: Al... |
2358 2359 2360 2361 2362 2363 2364 |
* 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... |
2365 2366 2367 2368 |
* * 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 ... |
2369 |
*/ |
6b4f4bc9c locking/futex: Al... |
2370 |
handle_err: |
734009e96 futex: Change loc... |
2371 |
raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); |
1b7558e45 futexes: fix faul... |
2372 |
spin_unlock(q->lock_ptr); |
778e9a9c3 pi-futex: fix exi... |
2373 |
|
6b4f4bc9c locking/futex: Al... |
2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 |
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... |
2389 |
|
1b7558e45 futexes: fix faul... |
2390 |
spin_lock(q->lock_ptr); |
734009e96 futex: Change loc... |
2391 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); |
778e9a9c3 pi-futex: fix exi... |
2392 |
|
1b7558e45 futexes: fix faul... |
2393 2394 2395 |
/* * Check if someone else fixed it for us: */ |
734009e96 futex: Change loc... |
2396 2397 2398 2399 |
if (pi_state->owner != oldowner) { ret = 0; goto out_unlock; } |
1b7558e45 futexes: fix faul... |
2400 2401 |
if (ret) |
734009e96 futex: Change loc... |
2402 |
goto out_unlock; |
1b7558e45 futexes: fix faul... |
2403 2404 |
goto retry; |
734009e96 futex: Change loc... |
2405 2406 2407 2408 |
out_unlock: raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); return ret; |
d0aa7a70b futex_requeue_pi ... |
2409 |
} |
72c1bbf30 futex: restartabl... |
2410 |
static long futex_wait_restart(struct restart_block *restart); |
36cf3b5c3 FUTEX: Tidy up th... |
2411 |
|
ca5f9524d futex: separate f... |
2412 |
/** |
dd9739980 futex: split out ... |
2413 2414 |
* fixup_owner() - Post lock pi_state and corner case management * @uaddr: user address of the futex |
dd9739980 futex: split out ... |
2415 2416 2417 2418 2419 2420 2421 |
* @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... |
2422 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2423 2424 2425 |
* - 1 - success, lock taken; * - 0 - success, lock not taken; * - <0 - on error (-EFAULT) |
dd9739980 futex: split out ... |
2426 |
*/ |
ae791a2d2 futex: Cleanup st... |
2427 |
static int fixup_owner(u32 __user *uaddr, struct futex_q *q, int locked) |
dd9739980 futex: split out ... |
2428 |
{ |
dd9739980 futex: split out ... |
2429 2430 2431 2432 2433 2434 |
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... |
2435 |
* |
c1e2f0eaf futex: Avoid viol... |
2436 2437 2438 |
* 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 ... |
2439 2440 |
*/ if (q->pi_state->owner != current) |
ae791a2d2 futex: Cleanup st... |
2441 |
ret = fixup_pi_state_owner(uaddr, q, current); |
d7c5ed73b futex: Remove nee... |
2442 |
return ret ? ret : locked; |
dd9739980 futex: split out ... |
2443 2444 2445 |
} /* |
c1e2f0eaf futex: Avoid viol... |
2446 2447 2448 2449 2450 2451 2452 2453 2454 |
* 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); |
d7c5ed73b futex: Remove nee... |
2455 |
return ret; |
c1e2f0eaf futex: Avoid viol... |
2456 2457 2458 |
} /* |
dd9739980 futex: split out ... |
2459 |
* Paranoia check. If we did not take the lock, then we should not be |
8161239a8 rtmutex: Simplify... |
2460 |
* the owner of the rt_mutex. |
dd9739980 futex: split out ... |
2461 |
*/ |
73d786bd0 futex: Rework inc... |
2462 |
if (rt_mutex_owner(&q->pi_state->pi_mutex) == current) { |
dd9739980 futex: split out ... |
2463 2464 2465 2466 2467 |
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... |
2468 |
} |
dd9739980 futex: split out ... |
2469 |
|
d7c5ed73b futex: Remove nee... |
2470 |
return ret; |
dd9739980 futex: split out ... |
2471 2472 2473 |
} /** |
ca5f9524d futex: separate f... |
2474 2475 2476 2477 |
* 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... |
2478 2479 |
*/ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, |
f1a11e057 futex: remove the... |
2480 |
struct hrtimer_sleeper *timeout) |
ca5f9524d futex: separate f... |
2481 |
{ |
9beba3c54 futex: Add memory... |
2482 2483 |
/* * The task state is guaranteed to be set before another task can |
b92b8b35a locking/arch: Ren... |
2484 |
* wake it. set_current_state() is implemented using smp_store_mb() and |
9beba3c54 futex: Add memory... |
2485 2486 2487 |
* queue_me() calls spin_unlock() upon completion, both serializing * access to the hash list and forcing another memory barrier. */ |
f1a11e057 futex: remove the... |
2488 |
set_current_state(TASK_INTERRUPTIBLE); |
0729e1961 futex: Fix wakeup... |
2489 |
queue_me(q, hb); |
ca5f9524d futex: separate f... |
2490 2491 |
/* Arm the timer */ |
2e4b0d3fe futex: Remove bog... |
2492 |
if (timeout) |
9dd8813ed hrtimer/treewide:... |
2493 |
hrtimer_sleeper_start_expires(timeout, HRTIMER_MODE_ABS); |
ca5f9524d futex: separate f... |
2494 2495 |
/* |
0729e1961 futex: Fix wakeup... |
2496 2497 |
* 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... |
2498 2499 2500 2501 2502 2503 2504 2505 |
*/ 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... |
2506 |
freezable_schedule(); |
ca5f9524d futex: separate f... |
2507 2508 2509 |
} __set_current_state(TASK_RUNNING); } |
f801073f8 futex: split out ... |
2510 2511 2512 2513 |
/** * futex_wait_setup() - Prepare to wait on a futex * @uaddr: the futex userspace address * @val: the expected value |
b41277dc7 futex: Replace fs... |
2514 |
* @flags: futex flags (FLAGS_SHARED, etc.) |
f801073f8 futex: split out ... |
2515 2516 2517 2518 2519 2520 2521 2522 |
* @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... |
2523 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2524 2525 |
* - 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 ... |
2526 |
*/ |
b41277dc7 futex: Replace fs... |
2527 |
static int futex_wait_setup(u32 __user *uaddr, u32 val, unsigned int flags, |
f801073f8 futex: split out ... |
2528 |
struct futex_q *q, struct futex_hash_bucket **hb) |
1da177e4c Linux-2.6.12-rc2 |
2529 |
{ |
e2970f2fb [PATCH] pi-futex:... |
2530 2531 |
u32 uval; int ret; |
1da177e4c Linux-2.6.12-rc2 |
2532 |
|
1da177e4c Linux-2.6.12-rc2 |
2533 |
/* |
b2d0994b1 futex: update fut... |
2534 |
* Access the page AFTER the hash-bucket is locked. |
1da177e4c Linux-2.6.12-rc2 |
2535 2536 2537 2538 2539 2540 2541 |
* 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... |
2542 2543 |
* 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 |
2544 2545 |
* cond(var) false, which would violate the guarantee. * |
8fe8f545c futex: Update fut... |
2546 2547 2548 2549 |
* 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 |
2550 |
*/ |
f801073f8 futex: split out ... |
2551 |
retry: |
96d4f267e Remove 'type' arg... |
2552 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &q->key, FUTEX_READ); |
f801073f8 futex: split out ... |
2553 |
if (unlikely(ret != 0)) |
a5a2a0c7f futex: fix futex_... |
2554 |
return ret; |
f801073f8 futex: split out ... |
2555 2556 2557 |
retry_private: *hb = queue_lock(q); |
e2970f2fb [PATCH] pi-futex:... |
2558 |
ret = get_futex_value_locked(&uval, uaddr); |
1da177e4c Linux-2.6.12-rc2 |
2559 |
|
f801073f8 futex: split out ... |
2560 |
if (ret) { |
0d00c7b20 futexes: Clean up... |
2561 |
queue_unlock(*hb); |
1da177e4c Linux-2.6.12-rc2 |
2562 |
|
e2970f2fb [PATCH] pi-futex:... |
2563 |
ret = get_user(uval, uaddr); |
e4dc5b7a3 futex: clean up f... |
2564 |
if (ret) |
d7c5ed73b futex: Remove nee... |
2565 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
2566 |
|
b41277dc7 futex: Replace fs... |
2567 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
2568 |
goto retry_private; |
e4dc5b7a3 futex: clean up f... |
2569 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
2570 |
} |
ca5f9524d futex: separate f... |
2571 |
|
f801073f8 futex: split out ... |
2572 |
if (uval != val) { |
0d00c7b20 futexes: Clean up... |
2573 |
queue_unlock(*hb); |
f801073f8 futex: split out ... |
2574 |
ret = -EWOULDBLOCK; |
2fff78c78 futex: fix refere... |
2575 |
} |
1da177e4c Linux-2.6.12-rc2 |
2576 |
|
f801073f8 futex: split out ... |
2577 2578 |
return ret; } |
b41277dc7 futex: Replace fs... |
2579 2580 |
static int futex_wait(u32 __user *uaddr, unsigned int flags, u32 val, ktime_t *abs_time, u32 bitset) |
f801073f8 futex: split out ... |
2581 |
{ |
5ca584d93 futex: Consolidat... |
2582 |
struct hrtimer_sleeper timeout, *to; |
f801073f8 futex: split out ... |
2583 2584 |
struct restart_block *restart; struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
2585 |
struct futex_q q = futex_q_init; |
f801073f8 futex: split out ... |
2586 2587 2588 2589 |
int ret; if (!bitset) return -EINVAL; |
f801073f8 futex: split out ... |
2590 |
q.bitset = bitset; |
5ca584d93 futex: Consolidat... |
2591 2592 |
to = futex_setup_timer(abs_time, &timeout, flags, current->timer_slack_ns); |
d58e6576b futex: Handle spu... |
2593 |
retry: |
7ada876a8 futex: Fix errors... |
2594 2595 2596 2597 |
/* * Prepare to wait on uaddr. On success, holds hb lock and increments * q.key refs. */ |
b41277dc7 futex: Replace fs... |
2598 |
ret = futex_wait_setup(uaddr, val, flags, &q, &hb); |
f801073f8 futex: split out ... |
2599 2600 |
if (ret) goto out; |
ca5f9524d futex: separate f... |
2601 |
/* queue_me and wait for wakeup, timeout, or a signal. */ |
f1a11e057 futex: remove the... |
2602 |
futex_wait_queue_me(hb, &q, to); |
1da177e4c Linux-2.6.12-rc2 |
2603 2604 |
/* If we were woken (and unqueued), we succeeded, whatever. */ |
2fff78c78 futex: fix refere... |
2605 |
ret = 0; |
7ada876a8 futex: Fix errors... |
2606 |
/* unqueue_me() drops q.key ref */ |
1da177e4c Linux-2.6.12-rc2 |
2607 |
if (!unqueue_me(&q)) |
7ada876a8 futex: Fix errors... |
2608 |
goto out; |
2fff78c78 futex: fix refere... |
2609 |
ret = -ETIMEDOUT; |
ca5f9524d futex: separate f... |
2610 |
if (to && !to->task) |
7ada876a8 futex: Fix errors... |
2611 |
goto out; |
72c1bbf30 futex: restartabl... |
2612 |
|
e2970f2fb [PATCH] pi-futex:... |
2613 |
/* |
d58e6576b futex: Handle spu... |
2614 2615 |
* We expect signal_pending(current), but we might be the * victim of a spurious wakeup as well. |
e2970f2fb [PATCH] pi-futex:... |
2616 |
*/ |
7ada876a8 futex: Fix errors... |
2617 |
if (!signal_pending(current)) |
d58e6576b futex: Handle spu... |
2618 |
goto retry; |
d58e6576b futex: Handle spu... |
2619 |
|
2fff78c78 futex: fix refere... |
2620 |
ret = -ERESTARTSYS; |
c19384b5b Make futex_wait()... |
2621 |
if (!abs_time) |
7ada876a8 futex: Fix errors... |
2622 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
2623 |
|
f56141e3e all arches, signa... |
2624 |
restart = ¤t->restart_block; |
2fff78c78 futex: fix refere... |
2625 |
restart->fn = futex_wait_restart; |
a3c74c525 futex: Mark resta... |
2626 |
restart->futex.uaddr = uaddr; |
2fff78c78 futex: fix refere... |
2627 |
restart->futex.val = val; |
2456e8553 ktime: Get rid of... |
2628 |
restart->futex.time = *abs_time; |
2fff78c78 futex: fix refere... |
2629 |
restart->futex.bitset = bitset; |
0cd9c6494 futex: Set FLAGS_... |
2630 |
restart->futex.flags = flags | FLAGS_HAS_TIMEOUT; |
42d35d48c futex: make futex... |
2631 |
|
2fff78c78 futex: fix refere... |
2632 |
ret = -ERESTART_RESTARTBLOCK; |
42d35d48c futex: make futex... |
2633 |
out: |
ca5f9524d futex: separate f... |
2634 2635 2636 2637 |
if (to) { hrtimer_cancel(&to->timer); destroy_hrtimer_on_stack(&to->timer); } |
c87e2837b [PATCH] pi-futex:... |
2638 2639 |
return ret; } |
72c1bbf30 futex: restartabl... |
2640 2641 2642 |
static long futex_wait_restart(struct restart_block *restart) { |
a3c74c525 futex: Mark resta... |
2643 |
u32 __user *uaddr = restart->futex.uaddr; |
a72188d8a futex: add FUTEX_... |
2644 |
ktime_t t, *tp = NULL; |
72c1bbf30 futex: restartabl... |
2645 |
|
a72188d8a futex: add FUTEX_... |
2646 |
if (restart->futex.flags & FLAGS_HAS_TIMEOUT) { |
2456e8553 ktime: Get rid of... |
2647 |
t = restart->futex.time; |
a72188d8a futex: add FUTEX_... |
2648 2649 |
tp = &t; } |
72c1bbf30 futex: restartabl... |
2650 |
restart->fn = do_no_restart_syscall; |
b41277dc7 futex: Replace fs... |
2651 2652 2653 |
return (long)futex_wait(uaddr, restart->futex.flags, restart->futex.val, tp, restart->futex.bitset); |
72c1bbf30 futex: restartabl... |
2654 |
} |
c87e2837b [PATCH] pi-futex:... |
2655 2656 2657 |
/* * 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... |
2658 2659 2660 2661 2662 |
* 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:... |
2663 |
*/ |
996636dda futex: Fix argume... |
2664 |
static int futex_lock_pi(u32 __user *uaddr, unsigned int flags, |
b41277dc7 futex: Replace fs... |
2665 |
ktime_t *time, int trylock) |
c87e2837b [PATCH] pi-futex:... |
2666 |
{ |
5ca584d93 futex: Consolidat... |
2667 |
struct hrtimer_sleeper timeout, *to; |
16ffa12d7 futex: Pull rt_mu... |
2668 |
struct futex_pi_state *pi_state = NULL; |
3ef240eaf futex: Prevent ex... |
2669 |
struct task_struct *exiting = NULL; |
cfafcd117 futex: Rework fut... |
2670 |
struct rt_mutex_waiter rt_waiter; |
c87e2837b [PATCH] pi-futex:... |
2671 |
struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
2672 |
struct futex_q q = futex_q_init; |
dd9739980 futex: split out ... |
2673 |
int res, ret; |
c87e2837b [PATCH] pi-futex:... |
2674 |
|
bc2eecd7e futex: Allow for ... |
2675 2676 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
c87e2837b [PATCH] pi-futex:... |
2677 2678 |
if (refill_pi_state_cache()) return -ENOMEM; |
5ca584d93 futex: Consolidat... |
2679 |
to = futex_setup_timer(time, &timeout, FLAGS_CLOCKRT, 0); |
c5780e976 [PATCH] Use the c... |
2680 |
|
42d35d48c futex: make futex... |
2681 |
retry: |
96d4f267e Remove 'type' arg... |
2682 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &q.key, FUTEX_WRITE); |
c87e2837b [PATCH] pi-futex:... |
2683 |
if (unlikely(ret != 0)) |
42d35d48c futex: make futex... |
2684 |
goto out; |
c87e2837b [PATCH] pi-futex:... |
2685 |
|
e4dc5b7a3 futex: clean up f... |
2686 |
retry_private: |
82af7aca5 Removal of FUTEX_FD |
2687 |
hb = queue_lock(&q); |
c87e2837b [PATCH] pi-futex:... |
2688 |
|
3ef240eaf futex: Prevent ex... |
2689 2690 |
ret = futex_lock_pi_atomic(uaddr, hb, &q.key, &q.pi_state, current, &exiting, 0); |
c87e2837b [PATCH] pi-futex:... |
2691 |
if (unlikely(ret)) { |
767f509ca futex: Enhance co... |
2692 2693 2694 2695 |
/* * Atomic work succeeded and we got the lock, * or failed. Either way, we do _not_ block. */ |
778e9a9c3 pi-futex: fix exi... |
2696 |
switch (ret) { |
1a52084d0 futex: split out ... |
2697 2698 2699 2700 2701 2702 |
case 1: /* We got the lock. */ ret = 0; goto out_unlock_put_key; case -EFAULT: goto uaddr_faulted; |
ac31c7ff8 futex: Provide di... |
2703 |
case -EBUSY: |
778e9a9c3 pi-futex: fix exi... |
2704 2705 |
case -EAGAIN: /* |
af54d6a1c futex: Simplify f... |
2706 |
* Two reasons for this: |
ac31c7ff8 futex: Provide di... |
2707 |
* - EBUSY: Task is exiting and we just wait for the |
af54d6a1c futex: Simplify f... |
2708 |
* exit to complete. |
ac31c7ff8 futex: Provide di... |
2709 |
* - EAGAIN: The user space value changed. |
778e9a9c3 pi-futex: fix exi... |
2710 |
*/ |
0d00c7b20 futexes: Clean up... |
2711 |
queue_unlock(hb); |
3ef240eaf futex: Prevent ex... |
2712 2713 2714 2715 2716 2717 |
/* * 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... |
2718 2719 |
cond_resched(); goto retry; |
778e9a9c3 pi-futex: fix exi... |
2720 |
default: |
42d35d48c futex: make futex... |
2721 |
goto out_unlock_put_key; |
c87e2837b [PATCH] pi-futex:... |
2722 |
} |
c87e2837b [PATCH] pi-futex:... |
2723 |
} |
cfafcd117 futex: Rework fut... |
2724 |
WARN_ON(!q.pi_state); |
c87e2837b [PATCH] pi-futex:... |
2725 2726 2727 |
/* * Only actually queue now that the atomic ops are done: */ |
cfafcd117 futex: Rework fut... |
2728 |
__queue_me(&q, hb); |
c87e2837b [PATCH] pi-futex:... |
2729 |
|
cfafcd117 futex: Rework fut... |
2730 |
if (trylock) { |
5293c2efd futex,rt_mutex: P... |
2731 |
ret = rt_mutex_futex_trylock(&q.pi_state->pi_mutex); |
c87e2837b [PATCH] pi-futex:... |
2732 2733 |
/* Fixup the trylock return value: */ ret = ret ? 0 : -EWOULDBLOCK; |
cfafcd117 futex: Rework fut... |
2734 |
goto no_block; |
c87e2837b [PATCH] pi-futex:... |
2735 |
} |
56222b212 futex: Drop hb->l... |
2736 |
rt_mutex_init_waiter(&rt_waiter); |
cfafcd117 futex: Rework fut... |
2737 |
/* |
56222b212 futex: Drop hb->l... |
2738 2739 2740 2741 2742 2743 2744 |
* 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... |
2745 2746 2747 2748 |
* 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... |
2749 |
*/ |
56222b212 futex: Drop hb->l... |
2750 2751 |
raw_spin_lock_irq(&q.pi_state->pi_mutex.wait_lock); spin_unlock(q.lock_ptr); |
1a1fb985f futex: Handle ear... |
2752 2753 2754 2755 2756 |
/* * __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... |
2757 2758 |
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... |
2759 2760 2761 |
if (ret) { if (ret == 1) ret = 0; |
1a1fb985f futex: Handle ear... |
2762 |
goto cleanup; |
cfafcd117 futex: Rework fut... |
2763 |
} |
cfafcd117 futex: Rework fut... |
2764 |
if (unlikely(to)) |
9dd8813ed hrtimer/treewide:... |
2765 |
hrtimer_sleeper_start_expires(to, HRTIMER_MODE_ABS); |
cfafcd117 futex: Rework fut... |
2766 2767 |
ret = rt_mutex_wait_proxy_lock(&q.pi_state->pi_mutex, to, &rt_waiter); |
1a1fb985f futex: Handle ear... |
2768 |
cleanup: |
a99e4e413 [PATCH] pi-futex:... |
2769 |
spin_lock(q.lock_ptr); |
dd9739980 futex: split out ... |
2770 |
/* |
1a1fb985f futex: Handle ear... |
2771 |
* If we failed to acquire the lock (deadlock/signal/timeout), we must |
cfafcd117 futex: Rework fut... |
2772 |
* first acquire the hb->lock before removing the lock from the |
1a1fb985f futex: Handle ear... |
2773 2774 |
* rt_mutex waitqueue, such that we can keep the hb and rt_mutex wait * lists consistent. |
56222b212 futex: Drop hb->l... |
2775 2776 2777 |
* * In particular; it is important that futex_unlock_pi() can not * observe this inconsistency. |
cfafcd117 futex: Rework fut... |
2778 2779 2780 2781 2782 2783 |
*/ if (ret && !rt_mutex_cleanup_proxy_lock(&q.pi_state->pi_mutex, &rt_waiter)) ret = 0; no_block: /* |
dd9739980 futex: split out ... |
2784 2785 2786 |
* Fixup the pi_state owner and possibly acquire the lock if we * haven't already. */ |
ae791a2d2 futex: Cleanup st... |
2787 |
res = fixup_owner(uaddr, &q, !ret); |
dd9739980 futex: split out ... |
2788 2789 2790 2791 2792 2793 |
/* * 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:... |
2794 |
|
e8f6386c0 futex: unlock bef... |
2795 |
/* |
dd9739980 futex: split out ... |
2796 2797 |
* If fixup_owner() faulted and was unable to handle the fault, unlock * it and return the fault to userspace. |
e8f6386c0 futex: unlock bef... |
2798 |
*/ |
16ffa12d7 futex: Pull rt_mu... |
2799 2800 2801 2802 |
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... |
2803 |
|
778e9a9c3 pi-futex: fix exi... |
2804 2805 |
/* Unqueue and drop the lock */ unqueue_me_pi(&q); |
c87e2837b [PATCH] pi-futex:... |
2806 |
|
16ffa12d7 futex: Pull rt_mu... |
2807 2808 2809 2810 |
if (pi_state) { rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); } |
9180bd467 futex: Remove put... |
2811 |
goto out; |
c87e2837b [PATCH] pi-futex:... |
2812 |
|
42d35d48c futex: make futex... |
2813 |
out_unlock_put_key: |
0d00c7b20 futexes: Clean up... |
2814 |
queue_unlock(hb); |
c87e2837b [PATCH] pi-futex:... |
2815 |
|
42d35d48c futex: make futex... |
2816 |
out: |
97181f9bd futex: Avoid free... |
2817 2818 |
if (to) { hrtimer_cancel(&to->timer); |
237fc6e7a add hrtimer speci... |
2819 |
destroy_hrtimer_on_stack(&to->timer); |
97181f9bd futex: Avoid free... |
2820 |
} |
dd9739980 futex: split out ... |
2821 |
return ret != -EINTR ? ret : -ERESTARTNOINTR; |
c87e2837b [PATCH] pi-futex:... |
2822 |
|
42d35d48c futex: make futex... |
2823 |
uaddr_faulted: |
0d00c7b20 futexes: Clean up... |
2824 |
queue_unlock(hb); |
778e9a9c3 pi-futex: fix exi... |
2825 |
|
d0725992c futex: Fix the wr... |
2826 |
ret = fault_in_user_writeable(uaddr); |
e4dc5b7a3 futex: clean up f... |
2827 |
if (ret) |
9180bd467 futex: Remove put... |
2828 |
goto out; |
c87e2837b [PATCH] pi-futex:... |
2829 |
|
b41277dc7 futex: Replace fs... |
2830 |
if (!(flags & FLAGS_SHARED)) |
e4dc5b7a3 futex: clean up f... |
2831 |
goto retry_private; |
e4dc5b7a3 futex: clean up f... |
2832 |
goto retry; |
c87e2837b [PATCH] pi-futex:... |
2833 2834 2835 |
} /* |
c87e2837b [PATCH] pi-futex:... |
2836 2837 2838 2839 |
* 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... |
2840 |
static int futex_unlock_pi(u32 __user *uaddr, unsigned int flags) |
c87e2837b [PATCH] pi-futex:... |
2841 |
{ |
3f649ab72 treewide: Remove ... |
2842 |
u32 curval, uval, vpid = task_pid_vnr(current); |
38d47c1b7 futex: rely on ge... |
2843 |
union futex_key key = FUTEX_KEY_INIT; |
ccf9e6a80 futex: Make unloc... |
2844 |
struct futex_hash_bucket *hb; |
499f5aca2 futex: Cleanup va... |
2845 |
struct futex_q *top_waiter; |
e4dc5b7a3 futex: clean up f... |
2846 |
int ret; |
c87e2837b [PATCH] pi-futex:... |
2847 |
|
bc2eecd7e futex: Allow for ... |
2848 2849 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
c87e2837b [PATCH] pi-futex:... |
2850 2851 2852 2853 2854 2855 |
retry: if (get_user(uval, uaddr)) return -EFAULT; /* * We release only a lock we actually own: */ |
c0c9ed150 futex: Avoid redu... |
2856 |
if ((uval & FUTEX_TID_MASK) != vpid) |
c87e2837b [PATCH] pi-futex:... |
2857 |
return -EPERM; |
c87e2837b [PATCH] pi-futex:... |
2858 |
|
96d4f267e Remove 'type' arg... |
2859 |
ret = get_futex_key(uaddr, flags & FLAGS_SHARED, &key, FUTEX_WRITE); |
ccf9e6a80 futex: Make unloc... |
2860 2861 |
if (ret) return ret; |
c87e2837b [PATCH] pi-futex:... |
2862 2863 2864 |
hb = hash_futex(&key); spin_lock(&hb->lock); |
c87e2837b [PATCH] pi-futex:... |
2865 |
/* |
ccf9e6a80 futex: Make unloc... |
2866 2867 2868 |
* 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:... |
2869 |
*/ |
499f5aca2 futex: Cleanup va... |
2870 2871 |
top_waiter = futex_top_waiter(hb, &key); if (top_waiter) { |
16ffa12d7 futex: Pull rt_mu... |
2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 |
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... |
2884 |
get_pi_state(pi_state); |
802ab58da futex: Lower the ... |
2885 |
/* |
bebe5b514 futex: Futex_unlo... |
2886 2887 2888 2889 |
* 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... |
2890 2891 2892 2893 |
* * 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... |
2894 |
*/ |
bebe5b514 futex: Futex_unlo... |
2895 |
raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); |
16ffa12d7 futex: Pull rt_mu... |
2896 |
spin_unlock(&hb->lock); |
c74aef2d0 futex: Fix pi_sta... |
2897 |
/* drops pi_state->pi_mutex.wait_lock */ |
16ffa12d7 futex: Pull rt_mu... |
2898 2899 2900 2901 2902 2903 |
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 ... |
2904 2905 2906 |
*/ if (!ret) goto out_putkey; |
c87e2837b [PATCH] pi-futex:... |
2907 |
/* |
ccf9e6a80 futex: Make unloc... |
2908 2909 |
* The atomic access to the futex value generated a * pagefault, so retry the user-access and the wakeup: |
c87e2837b [PATCH] pi-futex:... |
2910 2911 2912 |
*/ if (ret == -EFAULT) goto pi_faulted; |
802ab58da futex: Lower the ... |
2913 |
/* |
89e9e66ba futex: Handle unl... |
2914 2915 2916 |
* A unconditional UNLOCK_PI op raced against a waiter * setting the FUTEX_WAITERS bit. Try again. */ |
6b4f4bc9c locking/futex: Al... |
2917 2918 |
if (ret == -EAGAIN) goto pi_retry; |
89e9e66ba futex: Handle unl... |
2919 |
/* |
802ab58da futex: Lower the ... |
2920 2921 2922 |
* wake_futex_pi has detected invalid state. Tell user * space. */ |
16ffa12d7 futex: Pull rt_mu... |
2923 |
goto out_putkey; |
c87e2837b [PATCH] pi-futex:... |
2924 |
} |
ccf9e6a80 futex: Make unloc... |
2925 |
|
c87e2837b [PATCH] pi-futex:... |
2926 |
/* |
ccf9e6a80 futex: Make unloc... |
2927 2928 2929 2930 2931 |
* 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:... |
2932 |
*/ |
6b4f4bc9c locking/futex: Al... |
2933 |
if ((ret = cmpxchg_futex_value_locked(&curval, uaddr, uval, 0))) { |
16ffa12d7 futex: Pull rt_mu... |
2934 |
spin_unlock(&hb->lock); |
6b4f4bc9c locking/futex: Al... |
2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 |
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... |
2946 |
} |
c87e2837b [PATCH] pi-futex:... |
2947 |
|
ccf9e6a80 futex: Make unloc... |
2948 2949 2950 2951 |
/* * If uval has changed, let user space handle it. */ ret = (curval == uval) ? 0 : -EAGAIN; |
c87e2837b [PATCH] pi-futex:... |
2952 2953 |
out_unlock: spin_unlock(&hb->lock); |
802ab58da futex: Lower the ... |
2954 |
out_putkey: |
c87e2837b [PATCH] pi-futex:... |
2955 |
return ret; |
6b4f4bc9c locking/futex: Al... |
2956 |
pi_retry: |
6b4f4bc9c locking/futex: Al... |
2957 2958 |
cond_resched(); goto retry; |
c87e2837b [PATCH] pi-futex:... |
2959 |
pi_faulted: |
c87e2837b [PATCH] pi-futex:... |
2960 |
|
d0725992c futex: Fix the wr... |
2961 |
ret = fault_in_user_writeable(uaddr); |
b56863630 futex: clean up f... |
2962 |
if (!ret) |
c87e2837b [PATCH] pi-futex:... |
2963 |
goto retry; |
1da177e4c Linux-2.6.12-rc2 |
2964 2965 |
return ret; } |
52400ba94 futex: add requeu... |
2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 |
/** * 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... |
2978 |
* Return: |
7b4ff1adb mutex, futex: adj... |
2979 2980 |
* - 0 = no early wakeup detected; * - <0 = -ETIMEDOUT or -ERESTARTNOINTR |
52400ba94 futex: add requeu... |
2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 |
*/ 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... |
3002 |
plist_del(&q->list, &hb->chain); |
11d4616bd futex: revert bac... |
3003 |
hb_waiters_dec(hb); |
52400ba94 futex: add requeu... |
3004 |
|
d58e6576b futex: Handle spu... |
3005 |
/* Handle spurious wakeups gracefully */ |
11df6dddc futex: Fix spurio... |
3006 |
ret = -EWOULDBLOCK; |
52400ba94 futex: add requeu... |
3007 3008 |
if (timeout && !timeout->task) ret = -ETIMEDOUT; |
d58e6576b futex: Handle spu... |
3009 |
else if (signal_pending(current)) |
1c840c149 futex: fix restar... |
3010 |
ret = -ERESTARTNOINTR; |
52400ba94 futex: add requeu... |
3011 3012 3013 3014 3015 3016 |
} return ret; } /** * futex_wait_requeue_pi() - Wait on uaddr and take uaddr2 |
56ec1607b futex: Correct fu... |
3017 |
* @uaddr: the futex we initially wait on (non-pi) |
b41277dc7 futex: Replace fs... |
3018 |
* @flags: futex flags (FLAGS_SHARED, FLAGS_CLOCKRT, etc.), they must be |
ab51fbab3 futex: Fault/erro... |
3019 |
* the same type, no requeueing from private to shared, etc. |
52400ba94 futex: add requeu... |
3020 3021 |
* @val: the expected value of uaddr * @abs_time: absolute timeout |
56ec1607b futex: Correct fu... |
3022 |
* @bitset: 32 bit wakeup bitset set by userspace, defaults to all |
52400ba94 futex: add requeu... |
3023 3024 3025 |
* @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... |
3026 3027 3028 3029 3030 |
* 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... |
3031 3032 |
* * We call schedule in futex_wait_queue_me() when we enqueue and return there |
6c23cbbd5 futex: fix kernel... |
3033 |
* via the following-- |
52400ba94 futex: add requeu... |
3034 |
* 1) wakeup on uaddr2 after an atomic lock acquisition by futex_requeue() |
cc6db4e60 futex: Correct fu... |
3035 3036 3037 |
* 2) wakeup on uaddr2 after a requeue * 3) signal * 4) timeout |
52400ba94 futex: add requeu... |
3038 |
* |
cc6db4e60 futex: Correct fu... |
3039 |
* If 3, cleanup and return -ERESTARTNOINTR. |
52400ba94 futex: add requeu... |
3040 3041 3042 3043 3044 3045 3046 |
* * 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... |
3047 |
* If 6, return -EWOULDBLOCK (restarting the syscall would do the same). |
52400ba94 futex: add requeu... |
3048 3049 3050 |
* * If 4 or 7, we cleanup and return with -ETIMEDOUT. * |
6c23cbbd5 futex: fix kernel... |
3051 |
* Return: |
7b4ff1adb mutex, futex: adj... |
3052 3053 |
* - 0 - On success; * - <0 - On error |
52400ba94 futex: add requeu... |
3054 |
*/ |
b41277dc7 futex: Replace fs... |
3055 |
static int futex_wait_requeue_pi(u32 __user *uaddr, unsigned int flags, |
52400ba94 futex: add requeu... |
3056 |
u32 val, ktime_t *abs_time, u32 bitset, |
b41277dc7 futex: Replace fs... |
3057 |
u32 __user *uaddr2) |
52400ba94 futex: add requeu... |
3058 |
{ |
5ca584d93 futex: Consolidat... |
3059 |
struct hrtimer_sleeper timeout, *to; |
16ffa12d7 futex: Pull rt_mu... |
3060 |
struct futex_pi_state *pi_state = NULL; |
52400ba94 futex: add requeu... |
3061 |
struct rt_mutex_waiter rt_waiter; |
52400ba94 futex: add requeu... |
3062 |
struct futex_hash_bucket *hb; |
5bdb05f91 futex: Add futex_... |
3063 3064 |
union futex_key key2 = FUTEX_KEY_INIT; struct futex_q q = futex_q_init; |
52400ba94 futex: add requeu... |
3065 |
int res, ret; |
52400ba94 futex: add requeu... |
3066 |
|
bc2eecd7e futex: Allow for ... |
3067 3068 |
if (!IS_ENABLED(CONFIG_FUTEX_PI)) return -ENOSYS; |
6f7b0a2a5 futex: Forbid uad... |
3069 3070 |
if (uaddr == uaddr2) return -EINVAL; |
52400ba94 futex: add requeu... |
3071 3072 |
if (!bitset) return -EINVAL; |
5ca584d93 futex: Consolidat... |
3073 3074 |
to = futex_setup_timer(abs_time, &timeout, flags, current->timer_slack_ns); |
52400ba94 futex: add requeu... |
3075 3076 3077 3078 3079 |
/* * The waiter is allocated on our stack, manipulated by the requeue * code while we sleep on uaddr. */ |
50809358d futex,rt_mutex: I... |
3080 |
rt_mutex_init_waiter(&rt_waiter); |
52400ba94 futex: add requeu... |
3081 |
|
96d4f267e Remove 'type' arg... |
3082 |
ret = get_futex_key(uaddr2, flags & FLAGS_SHARED, &key2, FUTEX_WRITE); |
52400ba94 futex: add requeu... |
3083 3084 |
if (unlikely(ret != 0)) goto out; |
84bc4af59 futex: Detect mis... |
3085 3086 3087 |
q.bitset = bitset; q.rt_waiter = &rt_waiter; q.requeue_pi_key = &key2; |
7ada876a8 futex: Fix errors... |
3088 3089 3090 3091 |
/* * Prepare to wait on uaddr. On success, increments q.key (key1) ref * count. */ |
b41277dc7 futex: Replace fs... |
3092 |
ret = futex_wait_setup(uaddr, val, flags, &q, &hb); |
c8b15a706 futex: cleanup er... |
3093 |
if (ret) |
9180bd467 futex: Remove put... |
3094 |
goto out; |
52400ba94 futex: add requeu... |
3095 |
|
e9c243a5a futex-prevent-req... |
3096 3097 3098 3099 3100 |
/* * 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-... |
3101 |
queue_unlock(hb); |
e9c243a5a futex-prevent-req... |
3102 |
ret = -EINVAL; |
9180bd467 futex: Remove put... |
3103 |
goto out; |
e9c243a5a futex-prevent-req... |
3104 |
} |
52400ba94 futex: add requeu... |
3105 |
/* Queue the futex_q, drop the hb lock, wait for wakeup. */ |
f1a11e057 futex: remove the... |
3106 |
futex_wait_queue_me(hb, &q, to); |
52400ba94 futex: add requeu... |
3107 3108 3109 3110 3111 |
spin_lock(&hb->lock); ret = handle_early_requeue_pi_wakeup(hb, &q, &key2, to); spin_unlock(&hb->lock); if (ret) |
9180bd467 futex: Remove put... |
3112 |
goto out; |
52400ba94 futex: add requeu... |
3113 3114 3115 3116 3117 |
/* * 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... |
3118 3119 3120 |
* 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... |
3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 |
*/ /* 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... |
3131 |
ret = fixup_pi_state_owner(uaddr2, &q, current); |
16ffa12d7 futex: Pull rt_mu... |
3132 3133 3134 3135 |
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... |
3136 3137 3138 3139 |
/* * Drop the reference to the pi state which * the requeue_pi() code acquired for us. */ |
29e9ee5d4 futex: Rename fre... |
3140 |
put_pi_state(q.pi_state); |
52400ba94 futex: add requeu... |
3141 3142 3143 |
spin_unlock(q.lock_ptr); } } else { |
c236c8e95 futex: Fix potent... |
3144 |
struct rt_mutex *pi_mutex; |
52400ba94 futex: add requeu... |
3145 3146 3147 3148 3149 |
/* * 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... |
3150 |
WARN_ON(!q.pi_state); |
52400ba94 futex: add requeu... |
3151 |
pi_mutex = &q.pi_state->pi_mutex; |
38d589f2f futex,rt_mutex: R... |
3152 |
ret = rt_mutex_wait_proxy_lock(pi_mutex, to, &rt_waiter); |
52400ba94 futex: add requeu... |
3153 3154 |
spin_lock(q.lock_ptr); |
38d589f2f futex,rt_mutex: R... |
3155 3156 3157 3158 |
if (ret && !rt_mutex_cleanup_proxy_lock(pi_mutex, &rt_waiter)) ret = 0; debug_rt_mutex_free_waiter(&rt_waiter); |
52400ba94 futex: add requeu... |
3159 3160 3161 3162 |
/* * Fixup the pi_state owner and possibly acquire the lock if we * haven't already. */ |
ae791a2d2 futex: Cleanup st... |
3163 |
res = fixup_owner(uaddr2, &q, !ret); |
52400ba94 futex: add requeu... |
3164 3165 |
/* * If fixup_owner() returned an error, proprogate that. If it |
56ec1607b futex: Correct fu... |
3166 |
* acquired the lock, clear -ETIMEDOUT or -EINTR. |
52400ba94 futex: add requeu... |
3167 3168 3169 |
*/ if (res) ret = (res < 0) ? res : 0; |
c236c8e95 futex: Fix potent... |
3170 3171 3172 3173 3174 |
/* * 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... |
3175 3176 3177 3178 |
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... |
3179 |
|
52400ba94 futex: add requeu... |
3180 3181 3182 |
/* Unqueue and drop the lock. */ unqueue_me_pi(&q); } |
16ffa12d7 futex: Pull rt_mu... |
3183 3184 3185 3186 |
if (pi_state) { rt_mutex_futex_unlock(&pi_state->pi_mutex); put_pi_state(pi_state); } |
c236c8e95 futex: Fix potent... |
3187 |
if (ret == -EINTR) { |
52400ba94 futex: add requeu... |
3188 |
/* |
cc6db4e60 futex: Correct fu... |
3189 3190 3191 3192 3193 |
* 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... |
3194 |
*/ |
2070887fd futex: fix restar... |
3195 |
ret = -EWOULDBLOCK; |
52400ba94 futex: add requeu... |
3196 |
} |
52400ba94 futex: add requeu... |
3197 3198 3199 3200 3201 3202 3203 |
out: if (to) { hrtimer_cancel(&to->timer); destroy_hrtimer_on_stack(&to->timer); } return ret; } |
0771dfefc [PATCH] lightweig... |
3204 3205 3206 3207 3208 3209 3210 |
/* * 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:... |
3211 |
* FUTEX_OWNER_DIED bit, and wakes up a waiter (if any). The list is |
0771dfefc [PATCH] lightweig... |
3212 3213 3214 3215 3216 3217 3218 3219 |
* 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... |
3220 3221 3222 |
* 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... |
3223 |
*/ |
836f92adf [CVE-2009-0029] S... |
3224 3225 |
SYSCALL_DEFINE2(set_robust_list, struct robust_list_head __user *, head, size_t, len) |
0771dfefc [PATCH] lightweig... |
3226 |
{ |
a0c1e9073 futex: runtime en... |
3227 3228 |
if (!futex_cmpxchg_enabled) return -ENOSYS; |
0771dfefc [PATCH] lightweig... |
3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 |
/* * 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... |
3241 3242 3243 3244 |
* 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... |
3245 |
*/ |
836f92adf [CVE-2009-0029] S... |
3246 3247 3248 |
SYSCALL_DEFINE3(get_robust_list, int, pid, struct robust_list_head __user * __user *, head_ptr, size_t __user *, len_ptr) |
0771dfefc [PATCH] lightweig... |
3249 |
{ |
ba46df984 [PATCH] __user an... |
3250 |
struct robust_list_head __user *head; |
0771dfefc [PATCH] lightweig... |
3251 |
unsigned long ret; |
bdbb776f8 futex: Do not lea... |
3252 |
struct task_struct *p; |
0771dfefc [PATCH] lightweig... |
3253 |
|
a0c1e9073 futex: runtime en... |
3254 3255 |
if (!futex_cmpxchg_enabled) return -ENOSYS; |
bdbb776f8 futex: Do not lea... |
3256 3257 3258 |
rcu_read_lock(); ret = -ESRCH; |
0771dfefc [PATCH] lightweig... |
3259 |
if (!pid) |
bdbb776f8 futex: Do not lea... |
3260 |
p = current; |
0771dfefc [PATCH] lightweig... |
3261 |
else { |
228ebcbe6 Uninline find_tas... |
3262 |
p = find_task_by_vpid(pid); |
0771dfefc [PATCH] lightweig... |
3263 3264 |
if (!p) goto err_unlock; |
0771dfefc [PATCH] lightweig... |
3265 |
} |
bdbb776f8 futex: Do not lea... |
3266 |
ret = -EPERM; |
caaee6234 ptrace: use fsuid... |
3267 |
if (!ptrace_may_access(p, PTRACE_MODE_READ_REALCREDS)) |
bdbb776f8 futex: Do not lea... |
3268 3269 3270 3271 |
goto err_unlock; head = p->robust_list; rcu_read_unlock(); |
0771dfefc [PATCH] lightweig... |
3272 3273 3274 3275 3276 |
if (put_user(sizeof(*head), len_ptr)) return -EFAULT; return put_user(head, head_ptr); err_unlock: |
aaa2a97eb [PATCH] sys_get_r... |
3277 |
rcu_read_unlock(); |
0771dfefc [PATCH] lightweig... |
3278 3279 3280 |
return ret; } |
ca16d5bee futex: Prevent ro... |
3281 3282 3283 |
/* Constants for the pending_op argument of handle_futex_death */ #define HANDLE_DEATH_PENDING true #define HANDLE_DEATH_LIST false |
0771dfefc [PATCH] lightweig... |
3284 3285 3286 3287 |
/* * Process a futex-list entry, check whether it's owned by the * dying task, and do notification if so: */ |
ca16d5bee futex: Prevent ro... |
3288 3289 |
static int handle_futex_death(u32 __user *uaddr, struct task_struct *curr, bool pi, bool pending_op) |
0771dfefc [PATCH] lightweig... |
3290 |
{ |
3f649ab72 treewide: Remove ... |
3291 |
u32 uval, nval, mval; |
6b4f4bc9c locking/futex: Al... |
3292 |
int err; |
0771dfefc [PATCH] lightweig... |
3293 |
|
5a07168d8 futex: Ensure tha... |
3294 3295 3296 |
/* Futex address must be 32bit aligned */ if ((((unsigned long)uaddr) % sizeof(*uaddr)) != 0) return -1; |
8f17d3a50 [PATCH] lightweig... |
3297 3298 |
retry: if (get_user(uval, uaddr)) |
0771dfefc [PATCH] lightweig... |
3299 |
return -1; |
ca16d5bee futex: Prevent ro... |
3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 |
/* * 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... |
3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 |
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... |
3362 3363 3364 |
if (fault_in_user_writeable(uaddr)) return -1; goto retry; |
6b4f4bc9c locking/futex: Al... |
3365 3366 3367 |
case -EAGAIN: cond_resched(); |
8f17d3a50 [PATCH] lightweig... |
3368 |
goto retry; |
0771dfefc [PATCH] lightweig... |
3369 |
|
6b4f4bc9c locking/futex: Al... |
3370 3371 3372 3373 |
default: WARN_ON_ONCE(1); return err; } |
0771dfefc [PATCH] lightweig... |
3374 |
} |
6b4f4bc9c locking/futex: Al... |
3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 |
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... |
3385 3386 3387 3388 |
return 0; } /* |
e3f2ddeac [PATCH] pi-futex:... |
3389 3390 3391 |
* 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... |
3392 |
struct robust_list __user * __user *head, |
1dcc41bb0 futex: Change 3rd... |
3393 |
unsigned int *pi) |
e3f2ddeac [PATCH] pi-futex:... |
3394 3395 |
{ unsigned long uentry; |
ba46df984 [PATCH] __user an... |
3396 |
if (get_user(uentry, (unsigned long __user *)head)) |
e3f2ddeac [PATCH] pi-futex:... |
3397 |
return -EFAULT; |
ba46df984 [PATCH] __user an... |
3398 |
*entry = (void __user *)(uentry & ~1UL); |
e3f2ddeac [PATCH] pi-futex:... |
3399 3400 3401 3402 3403 3404 |
*pi = uentry & 1; return 0; } /* |
0771dfefc [PATCH] lightweig... |
3405 3406 3407 3408 3409 |
* 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. */ |
ba31c1a48 futex: Move futex... |
3410 |
static void exit_robust_list(struct task_struct *curr) |
0771dfefc [PATCH] lightweig... |
3411 3412 |
{ struct robust_list_head __user *head = curr->robust_list; |
9f96cb1e8 robust futex thre... |
3413 |
struct robust_list __user *entry, *next_entry, *pending; |
4c115e951 futex: Address co... |
3414 |
unsigned int limit = ROBUST_LIST_LIMIT, pi, pip; |
3f649ab72 treewide: Remove ... |
3415 |
unsigned int next_pi; |
0771dfefc [PATCH] lightweig... |
3416 |
unsigned long futex_offset; |
9f96cb1e8 robust futex thre... |
3417 |
int rc; |
0771dfefc [PATCH] lightweig... |
3418 |
|
a0c1e9073 futex: runtime en... |
3419 3420 |
if (!futex_cmpxchg_enabled) return; |
0771dfefc [PATCH] lightweig... |
3421 3422 3423 3424 |
/* * Fetch the list head (which was registered earlier, via * sys_set_robust_list()): */ |
e3f2ddeac [PATCH] pi-futex:... |
3425 |
if (fetch_robust_entry(&entry, &head->list.next, &pi)) |
0771dfefc [PATCH] lightweig... |
3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 |
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:... |
3436 |
if (fetch_robust_entry(&pending, &head->list_op_pending, &pip)) |
0771dfefc [PATCH] lightweig... |
3437 |
return; |
e3f2ddeac [PATCH] pi-futex:... |
3438 |
|
9f96cb1e8 robust futex thre... |
3439 |
next_entry = NULL; /* avoid warning with gcc */ |
0771dfefc [PATCH] lightweig... |
3440 3441 |
while (entry != &head->list) { /* |
9f96cb1e8 robust futex thre... |
3442 3443 3444 3445 3446 |
* 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... |
3447 |
* A pending lock might already be on the list, so |
c87e2837b [PATCH] pi-futex:... |
3448 |
* don't process it twice: |
0771dfefc [PATCH] lightweig... |
3449 |
*/ |
ca16d5bee futex: Prevent ro... |
3450 |
if (entry != pending) { |
ba46df984 [PATCH] __user an... |
3451 |
if (handle_futex_death((void __user *)entry + futex_offset, |
ca16d5bee futex: Prevent ro... |
3452 |
curr, pi, HANDLE_DEATH_LIST)) |
0771dfefc [PATCH] lightweig... |
3453 |
return; |
ca16d5bee futex: Prevent ro... |
3454 |
} |
9f96cb1e8 robust futex thre... |
3455 |
if (rc) |
0771dfefc [PATCH] lightweig... |
3456 |
return; |
9f96cb1e8 robust futex thre... |
3457 3458 |
entry = next_entry; pi = next_pi; |
0771dfefc [PATCH] lightweig... |
3459 3460 3461 3462 3463 3464 3465 3466 |
/* * Avoid excessively long or circular lists: */ if (!--limit) break; cond_resched(); } |
9f96cb1e8 robust futex thre... |
3467 |
|
ca16d5bee futex: Prevent ro... |
3468 |
if (pending) { |
9f96cb1e8 robust futex thre... |
3469 |
handle_futex_death((void __user *)pending + futex_offset, |
ca16d5bee futex: Prevent ro... |
3470 3471 |
curr, pip, HANDLE_DEATH_PENDING); } |
0771dfefc [PATCH] lightweig... |
3472 |
} |
af8cbda2c futex: Provide st... |
3473 |
static void futex_cleanup(struct task_struct *tsk) |
ba31c1a48 futex: Move futex... |
3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 |
{ 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); } |
18f694385 futex: Mark the b... |
3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 |
/** * 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) { |
3f186d974 futex: Add mutex ... |
3509 3510 3511 |
/* 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); |
18f694385 futex: Mark the b... |
3512 3513 |
tsk->futex_state = FUTEX_STATE_DEAD; } |
af8cbda2c futex: Provide st... |
3514 |
static void futex_cleanup_begin(struct task_struct *tsk) |
150d71584 futex: Split fute... |
3515 |
{ |
18f694385 futex: Mark the b... |
3516 |
/* |
3f186d974 futex: Add mutex ... |
3517 3518 3519 3520 3521 3522 3523 3524 |
* 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); /* |
4a8e991b9 futex: Sanitize e... |
3525 3526 3527 3528 3529 3530 3531 3532 3533 |
* 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(). |
18f694385 futex: Mark the b... |
3534 3535 |
*/ raw_spin_lock_irq(&tsk->pi_lock); |
4a8e991b9 futex: Sanitize e... |
3536 |
tsk->futex_state = FUTEX_STATE_EXITING; |
18f694385 futex: Mark the b... |
3537 |
raw_spin_unlock_irq(&tsk->pi_lock); |
af8cbda2c futex: Provide st... |
3538 |
} |
18f694385 futex: Mark the b... |
3539 |
|
af8cbda2c futex: Provide st... |
3540 3541 3542 3543 3544 3545 3546 |
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; |
3f186d974 futex: Add mutex ... |
3547 3548 3549 3550 3551 |
/* * Drop the exit protection. This unblocks waiters which observed * FUTEX_STATE_EXITING to reevaluate the state. */ mutex_unlock(&tsk->futex_exit_mutex); |
af8cbda2c futex: Provide st... |
3552 |
} |
18f694385 futex: Mark the b... |
3553 |
|
af8cbda2c futex: Provide st... |
3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 |
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); |
150d71584 futex: Split fute... |
3577 |
} |
c19384b5b Make futex_wait()... |
3578 |
long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, |
e2970f2fb [PATCH] pi-futex:... |
3579 |
u32 __user *uaddr2, u32 val2, u32 val3) |
1da177e4c Linux-2.6.12-rc2 |
3580 |
{ |
81b40539e futex: Simplify r... |
3581 |
int cmd = op & FUTEX_CMD_MASK; |
b41277dc7 futex: Replace fs... |
3582 |
unsigned int flags = 0; |
34f01cc1f FUTEX: new PRIVAT... |
3583 3584 |
if (!(op & FUTEX_PRIVATE_FLAG)) |
b41277dc7 futex: Replace fs... |
3585 |
flags |= FLAGS_SHARED; |
1da177e4c Linux-2.6.12-rc2 |
3586 |
|
b41277dc7 futex: Replace fs... |
3587 3588 |
if (op & FUTEX_CLOCK_REALTIME) { flags |= FLAGS_CLOCKRT; |
337f13046 futex: Allow FUTE... |
3589 3590 |
if (cmd != FUTEX_WAIT && cmd != FUTEX_WAIT_BITSET && \ cmd != FUTEX_WAIT_REQUEUE_PI) |
b41277dc7 futex: Replace fs... |
3591 3592 |
return -ENOSYS; } |
1da177e4c Linux-2.6.12-rc2 |
3593 |
|
34f01cc1f FUTEX: new PRIVAT... |
3594 |
switch (cmd) { |
59263b513 futex: Cover all ... |
3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 |
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 |
3605 |
case FUTEX_WAIT: |
cd689985c futex: Add bitset... |
3606 |
val3 = FUTEX_BITSET_MATCH_ANY; |
405fa8ac8 futex: Convert to... |
3607 |
fallthrough; |
cd689985c futex: Add bitset... |
3608 |
case FUTEX_WAIT_BITSET: |
81b40539e futex: Simplify r... |
3609 |
return futex_wait(uaddr, flags, val, timeout, val3); |
1da177e4c Linux-2.6.12-rc2 |
3610 |
case FUTEX_WAKE: |
cd689985c futex: Add bitset... |
3611 |
val3 = FUTEX_BITSET_MATCH_ANY; |
405fa8ac8 futex: Convert to... |
3612 |
fallthrough; |
cd689985c futex: Add bitset... |
3613 |
case FUTEX_WAKE_BITSET: |
81b40539e futex: Simplify r... |
3614 |
return futex_wake(uaddr, flags, val, val3); |
1da177e4c Linux-2.6.12-rc2 |
3615 |
case FUTEX_REQUEUE: |
81b40539e futex: Simplify r... |
3616 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, NULL, 0); |
1da177e4c Linux-2.6.12-rc2 |
3617 |
case FUTEX_CMP_REQUEUE: |
81b40539e futex: Simplify r... |
3618 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 0); |
4732efbeb [PATCH] FUTEX_WAK... |
3619 |
case FUTEX_WAKE_OP: |
81b40539e futex: Simplify r... |
3620 |
return futex_wake_op(uaddr, flags, uaddr2, val, val2, val3); |
c87e2837b [PATCH] pi-futex:... |
3621 |
case FUTEX_LOCK_PI: |
996636dda futex: Fix argume... |
3622 |
return futex_lock_pi(uaddr, flags, timeout, 0); |
c87e2837b [PATCH] pi-futex:... |
3623 |
case FUTEX_UNLOCK_PI: |
81b40539e futex: Simplify r... |
3624 |
return futex_unlock_pi(uaddr, flags); |
c87e2837b [PATCH] pi-futex:... |
3625 |
case FUTEX_TRYLOCK_PI: |
996636dda futex: Fix argume... |
3626 |
return futex_lock_pi(uaddr, flags, NULL, 1); |
52400ba94 futex: add requeu... |
3627 3628 |
case FUTEX_WAIT_REQUEUE_PI: val3 = FUTEX_BITSET_MATCH_ANY; |
81b40539e futex: Simplify r... |
3629 3630 |
return futex_wait_requeue_pi(uaddr, flags, val, timeout, val3, uaddr2); |
52400ba94 futex: add requeu... |
3631 |
case FUTEX_CMP_REQUEUE_PI: |
81b40539e futex: Simplify r... |
3632 |
return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 1); |
1da177e4c Linux-2.6.12-rc2 |
3633 |
} |
81b40539e futex: Simplify r... |
3634 |
return -ENOSYS; |
1da177e4c Linux-2.6.12-rc2 |
3635 |
} |
17da2bd90 [CVE-2009-0029] S... |
3636 |
SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, |
bec2f7cbb y2038: futex: Add... |
3637 |
struct __kernel_timespec __user *, utime, u32 __user *, uaddr2, |
17da2bd90 [CVE-2009-0029] S... |
3638 |
u32, val3) |
1da177e4c Linux-2.6.12-rc2 |
3639 |
{ |
bec2f7cbb y2038: futex: Add... |
3640 |
struct timespec64 ts; |
c19384b5b Make futex_wait()... |
3641 |
ktime_t t, *tp = NULL; |
e2970f2fb [PATCH] pi-futex:... |
3642 |
u32 val2 = 0; |
34f01cc1f FUTEX: new PRIVAT... |
3643 |
int cmd = op & FUTEX_CMD_MASK; |
1da177e4c Linux-2.6.12-rc2 |
3644 |
|
cd689985c futex: Add bitset... |
3645 |
if (utime && (cmd == FUTEX_WAIT || cmd == FUTEX_LOCK_PI || |
52400ba94 futex: add requeu... |
3646 3647 |
cmd == FUTEX_WAIT_BITSET || cmd == FUTEX_WAIT_REQUEUE_PI)) { |
ab51fbab3 futex: Fault/erro... |
3648 3649 |
if (unlikely(should_fail_futex(!(op & FUTEX_PRIVATE_FLAG)))) return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3650 |
if (get_timespec64(&ts, utime)) |
1da177e4c Linux-2.6.12-rc2 |
3651 |
return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3652 |
if (!timespec64_valid(&ts)) |
9741ef964 [PATCH] futex: ch... |
3653 |
return -EINVAL; |
c19384b5b Make futex_wait()... |
3654 |
|
bec2f7cbb y2038: futex: Add... |
3655 |
t = timespec64_to_ktime(ts); |
34f01cc1f FUTEX: new PRIVAT... |
3656 |
if (cmd == FUTEX_WAIT) |
5a7780e72 hrtimer: check re... |
3657 |
t = ktime_add_safe(ktime_get(), t); |
c2f7d08cc futex: Adjust abs... |
3658 3659 |
else if (!(op & FUTEX_CLOCK_REALTIME)) t = timens_ktime_to_host(CLOCK_MONOTONIC, t); |
c19384b5b Make futex_wait()... |
3660 |
tp = &t; |
1da177e4c Linux-2.6.12-rc2 |
3661 3662 |
} /* |
52400ba94 futex: add requeu... |
3663 |
* requeue parameter in 'utime' if cmd == FUTEX_*_REQUEUE_*. |
f54f09861 futex: pass nr_wa... |
3664 |
* number of waiters to wake in 'utime' if cmd == FUTEX_WAKE_OP. |
1da177e4c Linux-2.6.12-rc2 |
3665 |
*/ |
f54f09861 futex: pass nr_wa... |
3666 |
if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE || |
ba9c22f2c futex: remove FUT... |
3667 |
cmd == FUTEX_CMP_REQUEUE_PI || cmd == FUTEX_WAKE_OP) |
e2970f2fb [PATCH] pi-futex:... |
3668 |
val2 = (u32) (unsigned long) utime; |
1da177e4c Linux-2.6.12-rc2 |
3669 |
|
c19384b5b Make futex_wait()... |
3670 |
return do_futex(uaddr, op, val, tp, uaddr2, val2, val3); |
1da177e4c Linux-2.6.12-rc2 |
3671 |
} |
04e7712f4 y2038: futex: Mov... |
3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 |
#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. */ |
ba31c1a48 futex: Move futex... |
3704 |
static void compat_exit_robust_list(struct task_struct *curr) |
04e7712f4 y2038: futex: Mov... |
3705 3706 3707 3708 |
{ 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; |
3f649ab72 treewide: Remove ... |
3709 |
unsigned int next_pi; |
04e7712f4 y2038: futex: Mov... |
3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 |
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); |
ca16d5bee futex: Prevent ro... |
3750 3751 |
if (handle_futex_death(uaddr, curr, pi, HANDLE_DEATH_LIST)) |
04e7712f4 y2038: futex: Mov... |
3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 |
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); |
ca16d5bee futex: Prevent ro... |
3769 |
handle_futex_death(uaddr, curr, pip, HANDLE_DEATH_PENDING); |
04e7712f4 y2038: futex: Mov... |
3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 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 3824 3825 |
} } 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... |
3826 |
#endif /* CONFIG_COMPAT */ |
04e7712f4 y2038: futex: Mov... |
3827 |
|
bec2f7cbb y2038: futex: Add... |
3828 |
#ifdef CONFIG_COMPAT_32BIT_TIME |
8dabe7245 y2038: syscalls: ... |
3829 |
SYSCALL_DEFINE6(futex_time32, u32 __user *, uaddr, int, op, u32, val, |
04e7712f4 y2038: futex: Mov... |
3830 3831 3832 |
struct old_timespec32 __user *, utime, u32 __user *, uaddr2, u32, val3) { |
bec2f7cbb y2038: futex: Add... |
3833 |
struct timespec64 ts; |
04e7712f4 y2038: futex: Mov... |
3834 3835 3836 3837 3838 3839 3840 |
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... |
3841 |
if (get_old_timespec32(&ts, utime)) |
04e7712f4 y2038: futex: Mov... |
3842 |
return -EFAULT; |
bec2f7cbb y2038: futex: Add... |
3843 |
if (!timespec64_valid(&ts)) |
04e7712f4 y2038: futex: Mov... |
3844 |
return -EINVAL; |
bec2f7cbb y2038: futex: Add... |
3845 |
t = timespec64_to_ktime(ts); |
04e7712f4 y2038: futex: Mov... |
3846 3847 |
if (cmd == FUTEX_WAIT) t = ktime_add_safe(ktime_get(), t); |
c2f7d08cc futex: Adjust abs... |
3848 3849 |
else if (!(op & FUTEX_CLOCK_REALTIME)) t = timens_ktime_to_host(CLOCK_MONOTONIC, t); |
04e7712f4 y2038: futex: Mov... |
3850 3851 3852 3853 3854 3855 3856 3857 |
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... |
3858 |
#endif /* CONFIG_COMPAT_32BIT_TIME */ |
04e7712f4 y2038: futex: Mov... |
3859 |
|
03b8c7b62 futex: Allow arch... |
3860 |
static void __init futex_detect_cmpxchg(void) |
1da177e4c Linux-2.6.12-rc2 |
3861 |
{ |
03b8c7b62 futex: Allow arch... |
3862 |
#ifndef CONFIG_HAVE_FUTEX_CMPXCHG |
a0c1e9073 futex: runtime en... |
3863 |
u32 curval; |
03b8c7b62 futex: Allow arch... |
3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 |
/* * 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... |
3882 |
unsigned int futex_shift; |
a52b89ebb futexes: Increase... |
3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 |
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... |
3894 3895 3896 |
&futex_shift, NULL, futex_hashsize, futex_hashsize); futex_hashsize = 1UL << futex_shift; |
03b8c7b62 futex: Allow arch... |
3897 3898 |
futex_detect_cmpxchg(); |
a0c1e9073 futex: runtime en... |
3899 |
|
a52b89ebb futexes: Increase... |
3900 |
for (i = 0; i < futex_hashsize; i++) { |
11d4616bd futex: revert bac... |
3901 |
atomic_set(&futex_queues[i].waiters, 0); |
732375c6a plist: Remove the... |
3902 |
plist_head_init(&futex_queues[i].chain); |
3e4ab747e futex: fix init o... |
3903 3904 |
spin_lock_init(&futex_queues[i].lock); } |
1da177e4c Linux-2.6.12-rc2 |
3905 3906 |
return 0; } |
25f71d1c3 futex: Move futex... |
3907 |
core_initcall(futex_init); |