1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
|
From: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Date: Thu, 21 Sep 2017 14:25:13 +0200
Subject: [PATCH] rtmutex: annotate sleeping lock context
Origin: https://www.kernel.org/pub/linux/kernel/projects/rt/5.0/older/patches-5.0.10-rt7.tar.xz
The RCU code complains on schedule() within a rcu_readlock() section.
The valid scenario on -RT is if a sleeping is held. In order to suppress
the warning the mirgrate_disable counter was used to identify the
invocation of schedule() due to lock contention.
Grygorii Strashko report that during CPU hotplug we might see the
warning via
rt_spin_lock() -> migrate_disable() -> pin_current_cpu() -> __read_rt_lock()
because the counter is not yet set.
It is also possible to trigger the warning from cpu_chill()
(seen on a kblockd_mod_delayed_work_on() caller).
To address this RCU warning I annotate the sleeping lock context. The
counter is incremented before migrate_disable() so the warning Grygorii
should not trigger anymore. Additionally I use that counter in
cpu_chill() to avoid the RCU warning from there.
Reported-by: Grygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
include/linux/preempt.h | 9 +++++++++
include/linux/sched.h | 26 ++++++++++++++++++++++++++
kernel/locking/rtmutex.c | 12 ++++++++++--
kernel/locking/rwlock-rt.c | 18 ++++++++++++++----
kernel/rcu/tree_plugin.h | 6 +++++-
kernel/sched/core.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
6 files changed, 109 insertions(+), 7 deletions(-)
--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -208,6 +208,15 @@ extern void migrate_enable(void);
int __migrate_disabled(struct task_struct *p);
+#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT_BASE)
+
+extern void migrate_disable(void);
+extern void migrate_enable(void);
+static inline int __migrate_disabled(struct task_struct *p)
+{
+ return 0;
+}
+
#else
#define migrate_disable() barrier()
#define migrate_enable() barrier()
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -672,6 +672,15 @@ struct task_struct {
# ifdef CONFIG_SCHED_DEBUG
int migrate_disable_atomic;
# endif
+
+#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT_BASE)
+ int migrate_disable;
+# ifdef CONFIG_SCHED_DEBUG
+ int migrate_disable_atomic;
+# endif
+#endif
+#ifdef CONFIG_PREEMPT_RT_FULL
+ int sleeping_lock;
#endif
#ifdef CONFIG_PREEMPT_RCU
@@ -1813,6 +1822,23 @@ static __always_inline bool need_resched
return unlikely(tif_need_resched());
}
+#ifdef CONFIG_PREEMPT_RT_FULL
+static inline void sleeping_lock_inc(void)
+{
+ current->sleeping_lock++;
+}
+
+static inline void sleeping_lock_dec(void)
+{
+ current->sleeping_lock--;
+}
+
+#else
+
+static inline void sleeping_lock_inc(void) { }
+static inline void sleeping_lock_dec(void) { }
+#endif
+
/*
* Wrappers for p->thread_info->cpu access. No-op on UP.
*/
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -1141,6 +1141,7 @@ void __sched rt_spin_lock_slowunlock(str
void __lockfunc rt_spin_lock(spinlock_t *lock)
{
+ sleeping_lock_inc();
migrate_disable();
spin_acquire(&lock->dep_map, 0, 0, _RET_IP_);
rt_spin_lock_fastlock(&lock->lock, rt_spin_lock_slowlock);
@@ -1155,6 +1156,7 @@ void __lockfunc __rt_spin_lock(struct rt
#ifdef CONFIG_DEBUG_LOCK_ALLOC
void __lockfunc rt_spin_lock_nested(spinlock_t *lock, int subclass)
{
+ sleeping_lock_inc();
migrate_disable();
spin_acquire(&lock->dep_map, subclass, 0, _RET_IP_);
rt_spin_lock_fastlock(&lock->lock, rt_spin_lock_slowlock);
@@ -1168,6 +1170,7 @@ void __lockfunc rt_spin_unlock(spinlock_
spin_release(&lock->dep_map, 1, _RET_IP_);
rt_spin_lock_fastunlock(&lock->lock, rt_spin_lock_slowunlock);
migrate_enable();
+ sleeping_lock_dec();
}
EXPORT_SYMBOL(rt_spin_unlock);
@@ -1193,12 +1196,15 @@ int __lockfunc rt_spin_trylock(spinlock_
{
int ret;
+ sleeping_lock_inc();
migrate_disable();
ret = __rt_mutex_trylock(&lock->lock);
- if (ret)
+ if (ret) {
spin_acquire(&lock->dep_map, 0, 1, _RET_IP_);
- else
+ } else {
migrate_enable();
+ sleeping_lock_dec();
+ }
return ret;
}
EXPORT_SYMBOL(rt_spin_trylock);
@@ -1210,6 +1216,7 @@ int __lockfunc rt_spin_trylock_bh(spinlo
local_bh_disable();
ret = __rt_mutex_trylock(&lock->lock);
if (ret) {
+ sleeping_lock_inc();
migrate_disable();
spin_acquire(&lock->dep_map, 0, 1, _RET_IP_);
} else
@@ -1225,6 +1232,7 @@ int __lockfunc rt_spin_trylock_irqsave(s
*flags = 0;
ret = __rt_mutex_trylock(&lock->lock);
if (ret) {
+ sleeping_lock_inc();
migrate_disable();
spin_acquire(&lock->dep_map, 0, 1, _RET_IP_);
}
--- a/kernel/locking/rwlock-rt.c
+++ b/kernel/locking/rwlock-rt.c
@@ -305,12 +305,15 @@ int __lockfunc rt_read_trylock(rwlock_t
{
int ret;
+ sleeping_lock_inc();
migrate_disable();
ret = do_read_rt_trylock(rwlock);
- if (ret)
+ if (ret) {
rwlock_acquire_read(&rwlock->dep_map, 0, 1, _RET_IP_);
- else
+ } else {
migrate_enable();
+ sleeping_lock_dec();
+ }
return ret;
}
EXPORT_SYMBOL(rt_read_trylock);
@@ -319,18 +322,22 @@ int __lockfunc rt_write_trylock(rwlock_t
{
int ret;
+ sleeping_lock_inc();
migrate_disable();
ret = do_write_rt_trylock(rwlock);
- if (ret)
+ if (ret) {
rwlock_acquire(&rwlock->dep_map, 0, 1, _RET_IP_);
- else
+ } else {
migrate_enable();
+ sleeping_lock_dec();
+ }
return ret;
}
EXPORT_SYMBOL(rt_write_trylock);
void __lockfunc rt_read_lock(rwlock_t *rwlock)
{
+ sleeping_lock_inc();
migrate_disable();
rwlock_acquire_read(&rwlock->dep_map, 0, 0, _RET_IP_);
do_read_rt_lock(rwlock);
@@ -339,6 +346,7 @@ EXPORT_SYMBOL(rt_read_lock);
void __lockfunc rt_write_lock(rwlock_t *rwlock)
{
+ sleeping_lock_inc();
migrate_disable();
rwlock_acquire(&rwlock->dep_map, 0, 0, _RET_IP_);
do_write_rt_lock(rwlock);
@@ -350,6 +358,7 @@ void __lockfunc rt_read_unlock(rwlock_t
rwlock_release(&rwlock->dep_map, 1, _RET_IP_);
do_read_rt_unlock(rwlock);
migrate_enable();
+ sleeping_lock_dec();
}
EXPORT_SYMBOL(rt_read_unlock);
@@ -358,6 +367,7 @@ void __lockfunc rt_write_unlock(rwlock_t
rwlock_release(&rwlock->dep_map, 1, _RET_IP_);
do_write_rt_unlock(rwlock);
migrate_enable();
+ sleeping_lock_dec();
}
EXPORT_SYMBOL(rt_write_unlock);
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -330,11 +330,15 @@ void rcu_note_context_switch(bool preemp
struct task_struct *t = current;
struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
struct rcu_node *rnp;
+ int sleeping_l = 0;
barrier(); /* Avoid RCU read-side critical sections leaking down. */
trace_rcu_utilization(TPS("Start context switch"));
lockdep_assert_irqs_disabled();
- WARN_ON_ONCE(!preempt && t->rcu_read_lock_nesting > 0);
+#if defined(CONFIG_PREEMPT_RT_FULL)
+ sleeping_l = t->sleeping_lock;
+#endif
+ WARN_ON_ONCE(!preempt && t->rcu_read_lock_nesting > 0 && !sleeping_l);
if (t->rcu_read_lock_nesting > 0 &&
!t->rcu_read_unlock_special.b.blocked) {
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7321,4 +7321,49 @@ void migrate_enable(void)
preempt_enable();
}
EXPORT_SYMBOL(migrate_enable);
+
+#elif !defined(CONFIG_SMP) && defined(CONFIG_PREEMPT_RT_BASE)
+void migrate_disable(void)
+{
+ struct task_struct *p = current;
+
+ if (in_atomic() || irqs_disabled()) {
+#ifdef CONFIG_SCHED_DEBUG
+ p->migrate_disable_atomic++;
+#endif
+ return;
+ }
+#ifdef CONFIG_SCHED_DEBUG
+ if (unlikely(p->migrate_disable_atomic)) {
+ tracing_off();
+ WARN_ON_ONCE(1);
+ }
+#endif
+
+ p->migrate_disable++;
+}
+EXPORT_SYMBOL(migrate_disable);
+
+void migrate_enable(void)
+{
+ struct task_struct *p = current;
+
+ if (in_atomic() || irqs_disabled()) {
+#ifdef CONFIG_SCHED_DEBUG
+ p->migrate_disable_atomic--;
+#endif
+ return;
+ }
+
+#ifdef CONFIG_SCHED_DEBUG
+ if (unlikely(p->migrate_disable_atomic)) {
+ tracing_off();
+ WARN_ON_ONCE(1);
+ }
+#endif
+
+ WARN_ON_ONCE(p->migrate_disable <= 0);
+ p->migrate_disable--;
+}
+EXPORT_SYMBOL(migrate_enable);
#endif
|