3 Using RCU to Protect Read-Mostly Linked Lists
4 =============================================
6 One of the most common uses of RCU is protecting read-mostly linked lists
7 (``struct list_head`` in list.h). One big advantage of this approach is
8 that all of the required memory ordering is provided by the list macros.
9 This document describes several list-based RCU use cases.
12 Example 1: Read-mostly list: Deferred Destruction
13 -------------------------------------------------
15 A widely used usecase for RCU lists in the kernel is lockless iteration over
16 all processes in the system. ``task_struct::tasks`` represents the list node that
17 links all the processes. The list can be traversed in parallel to any list
18 additions or removals.
20 The traversal of the list is done using ``for_each_process()`` which is defined
23 #define next_task(p) \
24 list_entry_rcu((p)->tasks.next, struct task_struct, tasks)
26 #define for_each_process(p) \
27 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
29 The code traversing the list of all processes typically looks like::
33 /* Do something with p */
37 The simplified and heavily inlined code for removing a process from a
40 void release_task(struct task_struct *p)
42 write_lock(&tasklist_lock);
43 list_del_rcu(&p->tasks);
44 write_unlock(&tasklist_lock);
45 call_rcu(&p->rcu, delayed_put_task_struct);
48 When a process exits, ``release_task()`` calls ``list_del_rcu(&p->tasks)``
49 via __exit_signal() and __unhash_process() under ``tasklist_lock``
50 writer lock protection. The list_del_rcu() invocation removes
51 the task from the list of all tasks. The ``tasklist_lock``
52 prevents concurrent list additions/removals from corrupting the
53 list. Readers using ``for_each_process()`` are not protected with the
54 ``tasklist_lock``. To prevent readers from noticing changes in the list
55 pointers, the ``task_struct`` object is freed only after one or more
56 grace periods elapse, with the help of call_rcu(), which is invoked via
57 put_task_struct_rcu_user(). This deferring of destruction ensures that
58 any readers traversing the list will see valid ``p->tasks.next`` pointers
59 and deletion/freeing can happen in parallel with traversal of the list.
60 This pattern is also called an **existence lock**, since RCU refrains
61 from invoking the delayed_put_task_struct() callback function until
62 all existing readers finish, which guarantees that the ``task_struct``
63 object in question will remain in existence until after the completion
64 of all RCU readers that might possibly have a reference to that object.
67 Example 2: Read-Side Action Taken Outside of Lock: No In-Place Updates
68 ----------------------------------------------------------------------
70 Some reader-writer locking use cases compute a value while holding
71 the read-side lock, but continue to use that value after that lock is
72 released. These use cases are often good candidates for conversion
73 to RCU. One prominent example involves network packet routing.
74 Because the packet-routing data tracks the state of equipment outside
75 of the computer, it will at times contain stale data. Therefore, once
76 the route has been computed, there is no need to hold the routing table
77 static during transmission of the packet. After all, you can hold the
78 routing table static all you want, but that won't keep the external
79 Internet from changing, and it is the state of the external Internet
80 that really matters. In addition, routing entries are typically added
81 or deleted, rather than being modified in place. This is a rare example
82 of the finite speed of light and the non-zero size of atoms actually
83 helping make synchronization be lighter weight.
85 A straightforward example of this type of RCU use case may be found in
86 the system-call auditing support. For example, a reader-writer locked
87 implementation of ``audit_filter_task()`` might be as follows::
89 static enum audit_state audit_filter_task(struct task_struct *tsk, char **key)
91 struct audit_entry *e;
92 enum audit_state state;
94 read_lock(&auditsc_lock);
95 /* Note: audit_filter_mutex held by caller. */
96 list_for_each_entry(e, &audit_tsklist, list) {
97 if (audit_filter_rules(tsk, &e->rule, NULL, &state)) {
98 if (state == AUDIT_STATE_RECORD)
99 *key = kstrdup(e->rule.filterkey, GFP_ATOMIC);
100 read_unlock(&auditsc_lock);
104 read_unlock(&auditsc_lock);
105 return AUDIT_BUILD_CONTEXT;
108 Here the list is searched under the lock, but the lock is dropped before
109 the corresponding value is returned. By the time that this value is acted
110 on, the list may well have been modified. This makes sense, since if
111 you are turning auditing off, it is OK to audit a few extra system calls.
113 This means that RCU can be easily applied to the read side, as follows::
115 static enum audit_state audit_filter_task(struct task_struct *tsk, char **key)
117 struct audit_entry *e;
118 enum audit_state state;
121 /* Note: audit_filter_mutex held by caller. */
122 list_for_each_entry_rcu(e, &audit_tsklist, list) {
123 if (audit_filter_rules(tsk, &e->rule, NULL, &state)) {
124 if (state == AUDIT_STATE_RECORD)
125 *key = kstrdup(e->rule.filterkey, GFP_ATOMIC);
131 return AUDIT_BUILD_CONTEXT;
134 The read_lock() and read_unlock() calls have become rcu_read_lock()
135 and rcu_read_unlock(), respectively, and the list_for_each_entry()
136 has become list_for_each_entry_rcu(). The **_rcu()** list-traversal
137 primitives add READ_ONCE() and diagnostic checks for incorrect use
138 outside of an RCU read-side critical section.
140 The changes to the update side are also straightforward. A reader-writer lock
141 might be used as follows for deletion and insertion in these simplified
142 versions of audit_del_rule() and audit_add_rule()::
144 static inline int audit_del_rule(struct audit_rule *rule,
145 struct list_head *list)
147 struct audit_entry *e;
149 write_lock(&auditsc_lock);
150 list_for_each_entry(e, list, list) {
151 if (!audit_compare_rule(rule, &e->rule)) {
153 write_unlock(&auditsc_lock);
157 write_unlock(&auditsc_lock);
158 return -EFAULT; /* No matching rule */
161 static inline int audit_add_rule(struct audit_entry *entry,
162 struct list_head *list)
164 write_lock(&auditsc_lock);
165 if (entry->rule.flags & AUDIT_PREPEND) {
166 entry->rule.flags &= ~AUDIT_PREPEND;
167 list_add(&entry->list, list);
169 list_add_tail(&entry->list, list);
171 write_unlock(&auditsc_lock);
175 Following are the RCU equivalents for these two functions::
177 static inline int audit_del_rule(struct audit_rule *rule,
178 struct list_head *list)
180 struct audit_entry *e;
182 /* No need to use the _rcu iterator here, since this is the only
183 * deletion routine. */
184 list_for_each_entry(e, list, list) {
185 if (!audit_compare_rule(rule, &e->rule)) {
186 list_del_rcu(&e->list);
187 call_rcu(&e->rcu, audit_free_rule);
191 return -EFAULT; /* No matching rule */
194 static inline int audit_add_rule(struct audit_entry *entry,
195 struct list_head *list)
197 if (entry->rule.flags & AUDIT_PREPEND) {
198 entry->rule.flags &= ~AUDIT_PREPEND;
199 list_add_rcu(&entry->list, list);
201 list_add_tail_rcu(&entry->list, list);
206 Normally, the write_lock() and write_unlock() would be replaced by a
207 spin_lock() and a spin_unlock(). But in this case, all callers hold
208 ``audit_filter_mutex``, so no additional locking is required. The
209 auditsc_lock can therefore be eliminated, since use of RCU eliminates the
210 need for writers to exclude readers.
212 The list_del(), list_add(), and list_add_tail() primitives have been
213 replaced by list_del_rcu(), list_add_rcu(), and list_add_tail_rcu().
214 The **_rcu()** list-manipulation primitives add memory barriers that are
215 needed on weakly ordered CPUs. The list_del_rcu() primitive omits the
216 pointer poisoning debug-assist code that would otherwise cause concurrent
217 readers to fail spectacularly.
219 So, when readers can tolerate stale data and when entries are either added or
220 deleted, without in-place modification, it is very easy to use RCU!
223 Example 3: Handling In-Place Updates
224 ------------------------------------
226 The system-call auditing code does not update auditing rules in place. However,
227 if it did, the reader-writer-locked code to do so might look as follows
228 (assuming only ``field_count`` is updated, otherwise, the added fields would
229 need to be filled in)::
231 static inline int audit_upd_rule(struct audit_rule *rule,
232 struct list_head *list,
234 __u32 newfield_count)
236 struct audit_entry *e;
237 struct audit_entry *ne;
239 write_lock(&auditsc_lock);
240 /* Note: audit_filter_mutex held by caller. */
241 list_for_each_entry(e, list, list) {
242 if (!audit_compare_rule(rule, &e->rule)) {
243 e->rule.action = newaction;
244 e->rule.field_count = newfield_count;
245 write_unlock(&auditsc_lock);
249 write_unlock(&auditsc_lock);
250 return -EFAULT; /* No matching rule */
253 The RCU version creates a copy, updates the copy, then replaces the old
254 entry with the newly updated entry. This sequence of actions, allowing
255 concurrent reads while making a copy to perform an update, is what gives
256 RCU (*read-copy update*) its name.
258 The RCU version of audit_upd_rule() is as follows::
260 static inline int audit_upd_rule(struct audit_rule *rule,
261 struct list_head *list,
263 __u32 newfield_count)
265 struct audit_entry *e;
266 struct audit_entry *ne;
268 list_for_each_entry(e, list, list) {
269 if (!audit_compare_rule(rule, &e->rule)) {
270 ne = kmalloc(sizeof(*entry), GFP_ATOMIC);
273 audit_copy_rule(&ne->rule, &e->rule);
274 ne->rule.action = newaction;
275 ne->rule.field_count = newfield_count;
276 list_replace_rcu(&e->list, &ne->list);
277 call_rcu(&e->rcu, audit_free_rule);
281 return -EFAULT; /* No matching rule */
284 Again, this assumes that the caller holds ``audit_filter_mutex``. Normally, the
285 writer lock would become a spinlock in this sort of code.
287 The update_lsm_rule() does something very similar, for those who would
288 prefer to look at real Linux-kernel code.
290 Another use of this pattern can be found in the openswitch driver's *connection
291 tracking table* code in ``ct_limit_set()``. The table holds connection tracking
292 entries and has a limit on the maximum entries. There is one such table
293 per-zone and hence one *limit* per zone. The zones are mapped to their limits
294 through a hashtable using an RCU-managed hlist for the hash chains. When a new
295 limit is set, a new limit object is allocated and ``ct_limit_set()`` is called
296 to replace the old limit object with the new one using list_replace_rcu().
297 The old limit object is then freed after a grace period using kfree_rcu().
300 Example 4: Eliminating Stale Data
301 ---------------------------------
303 The auditing example above tolerates stale data, as do most algorithms
304 that are tracking external state. After all, given there is a delay
305 from the time the external state changes before Linux becomes aware
306 of the change, and so as noted earlier, a small quantity of additional
307 RCU-induced staleness is generally not a problem.
309 However, there are many examples where stale data cannot be tolerated.
310 One example in the Linux kernel is the System V IPC (see the shm_lock()
311 function in ipc/shm.c). This code checks a *deleted* flag under a
312 per-entry spinlock, and, if the *deleted* flag is set, pretends that the
313 entry does not exist. For this to be helpful, the search function must
314 return holding the per-entry spinlock, as shm_lock() does in fact do.
319 For the deleted-flag technique to be helpful, why is it necessary
320 to hold the per-entry lock while returning from the search function?
322 :ref:`Answer to Quick Quiz <quick_quiz_answer>`
324 If the system-call audit module were to ever need to reject stale data, one way
325 to accomplish this would be to add a ``deleted`` flag and a ``lock`` spinlock to the
326 ``audit_entry`` structure, and modify audit_filter_task() as follows::
328 static enum audit_state audit_filter_task(struct task_struct *tsk)
330 struct audit_entry *e;
331 enum audit_state state;
334 list_for_each_entry_rcu(e, &audit_tsklist, list) {
335 if (audit_filter_rules(tsk, &e->rule, NULL, &state)) {
338 spin_unlock(&e->lock);
340 return AUDIT_BUILD_CONTEXT;
343 if (state == AUDIT_STATE_RECORD)
344 *key = kstrdup(e->rule.filterkey, GFP_ATOMIC);
349 return AUDIT_BUILD_CONTEXT;
352 The ``audit_del_rule()`` function would need to set the ``deleted`` flag under the
353 spinlock as follows::
355 static inline int audit_del_rule(struct audit_rule *rule,
356 struct list_head *list)
358 struct audit_entry *e;
360 /* No need to use the _rcu iterator here, since this
361 * is the only deletion routine. */
362 list_for_each_entry(e, list, list) {
363 if (!audit_compare_rule(rule, &e->rule)) {
365 list_del_rcu(&e->list);
367 spin_unlock(&e->lock);
368 call_rcu(&e->rcu, audit_free_rule);
372 return -EFAULT; /* No matching rule */
375 This too assumes that the caller holds ``audit_filter_mutex``.
377 Note that this example assumes that entries are only added and deleted.
378 Additional mechanism is required to deal correctly with the update-in-place
379 performed by audit_upd_rule(). For one thing, audit_upd_rule() would
380 need to hold the locks of both the old ``audit_entry`` and its replacement
381 while executing the list_replace_rcu().
384 Example 5: Skipping Stale Objects
385 ---------------------------------
387 For some use cases, reader performance can be improved by skipping
388 stale objects during read-side list traversal, where stale objects
389 are those that will be removed and destroyed after one or more grace
390 periods. One such example can be found in the timerfd subsystem. When a
391 ``CLOCK_REALTIME`` clock is reprogrammed (for example due to setting
392 of the system time) then all programmed ``timerfds`` that depend on
393 this clock get triggered and processes waiting on them are awakened in
394 advance of their scheduled expiry. To facilitate this, all such timers
395 are added to an RCU-managed ``cancel_list`` when they are setup in
396 ``timerfd_setup_cancel()``::
398 static void timerfd_setup_cancel(struct timerfd_ctx *ctx, int flags)
400 spin_lock(&ctx->cancel_lock);
401 if ((ctx->clockid == CLOCK_REALTIME ||
402 ctx->clockid == CLOCK_REALTIME_ALARM) &&
403 (flags & TFD_TIMER_ABSTIME) && (flags & TFD_TIMER_CANCEL_ON_SET)) {
404 if (!ctx->might_cancel) {
405 ctx->might_cancel = true;
406 spin_lock(&cancel_lock);
407 list_add_rcu(&ctx->clist, &cancel_list);
408 spin_unlock(&cancel_lock);
411 __timerfd_remove_cancel(ctx);
413 spin_unlock(&ctx->cancel_lock);
416 When a timerfd is freed (fd is closed), then the ``might_cancel``
417 flag of the timerfd object is cleared, the object removed from the
418 ``cancel_list`` and destroyed, as shown in this simplified and inlined
419 version of timerfd_release()::
421 int timerfd_release(struct inode *inode, struct file *file)
423 struct timerfd_ctx *ctx = file->private_data;
425 spin_lock(&ctx->cancel_lock);
426 if (ctx->might_cancel) {
427 ctx->might_cancel = false;
428 spin_lock(&cancel_lock);
429 list_del_rcu(&ctx->clist);
430 spin_unlock(&cancel_lock);
432 spin_unlock(&ctx->cancel_lock);
435 alarm_cancel(&ctx->t.alarm);
437 hrtimer_cancel(&ctx->t.tmr);
442 If the ``CLOCK_REALTIME`` clock is set, for example by a time server, the
443 hrtimer framework calls ``timerfd_clock_was_set()`` which walks the
444 ``cancel_list`` and wakes up processes waiting on the timerfd. While iterating
445 the ``cancel_list``, the ``might_cancel`` flag is consulted to skip stale
448 void timerfd_clock_was_set(void)
450 ktime_t moffs = ktime_mono_to_real(0);
451 struct timerfd_ctx *ctx;
455 list_for_each_entry_rcu(ctx, &cancel_list, clist) {
456 if (!ctx->might_cancel)
458 spin_lock_irqsave(&ctx->wqh.lock, flags);
459 if (ctx->moffs != moffs) {
460 ctx->moffs = KTIME_MAX;
462 wake_up_locked_poll(&ctx->wqh, EPOLLIN);
464 spin_unlock_irqrestore(&ctx->wqh.lock, flags);
469 The key point is that because RCU-protected traversal of the
470 ``cancel_list`` happens concurrently with object addition and removal,
471 sometimes the traversal can access an object that has been removed from
472 the list. In this example, a flag is used to skip such objects.
478 Read-mostly list-based data structures that can tolerate stale data are
479 the most amenable to use of RCU. The simplest case is where entries are
480 either added or deleted from the data structure (or atomically modified
481 in place), but non-atomic in-place modifications can be handled by making
482 a copy, updating the copy, then replacing the original with the copy.
483 If stale data cannot be tolerated, then a *deleted* flag may be used
484 in conjunction with a per-entry spinlock in order to allow the search
485 function to reject newly deleted data.
487 .. _quick_quiz_answer:
489 Answer to Quick Quiz:
490 For the deleted-flag technique to be helpful, why is it necessary
491 to hold the per-entry lock while returning from the search function?
493 If the search function drops the per-entry lock before returning,
494 then the caller will be processing stale data in any case. If it
495 is really OK to be processing stale data, then you don't need a
496 *deleted* flag. If processing stale data really is a problem,
497 then you need to hold the per-entry lock across all of the code
498 that uses the value that was returned.
500 :ref:`Back to Quick Quiz <quick_quiz>`